port to 64 bit Intel Mac OS X version

14 files changed, 27172 insertions, 0 deletions

diff --git a/macho64/aap.s b/macho64/aap.s
new file mode 100644
index 0000000..0db8c1c
--- /dev/null
+++ b/macho64/aap.s
@@ -0,0 +1,2418 @@
+
+	.intel_syntax noprefix
+
+/*	d0 = rax */
+/*	d1 = rbx */
+/*	d2 = r10 */
+/*	d3 = r11 */
+/*	d7 = r15 */
+/*	a0 = rcx */
+/*	a1 = rdx */
+/*	a2 = r8 */
+/*	a3 = r9 */
+/*	a4 = rsi */
+/*	a6 = rdi */
+
+/*	d0l = eax */
+/*	a2l = r8d */
+/*	a3l = r9d */
+
+/*	d1w = bx */
+
+/*	qw = qword ptr */
+
+	.text
+
+	.globl	ap_2
+	.globl	ap_3
+	.globl	ap_4
+	.globl	ap_5
+	.globl	ap_6
+	.globl	ap_7
+	.globl	ap_8
+	.globl	ap_9
+	.globl	ap_10
+	.globl	ap_11
+	.globl	ap_12
+	.globl	ap_13
+	.globl	ap_14
+	.globl	ap_15
+	.globl	ap_16
+	.globl	ap_17
+	.globl	ap_18
+	.globl	ap_19
+	.globl	ap_20
+	.globl	ap_21
+	.globl	ap_22
+	.globl	ap_23
+	.globl	ap_24
+	.globl	ap_25
+	.globl	ap_26
+	.globl	ap_27
+	.globl	ap_28
+	.globl	ap_29
+	.globl	ap_30
+	.globl	ap_31
+	.globl	ap_32
+
+	.globl	add_empty_node_2
+	.globl	add_empty_node_3
+	.globl	add_empty_node_4
+	.globl	add_empty_node_5
+	.globl	add_empty_node_6
+	.globl	add_empty_node_7
+	.globl	add_empty_node_8
+	.globl	add_empty_node_9
+	.globl	add_empty_node_10
+	.globl	add_empty_node_11
+	.globl	add_empty_node_12
+	.globl	add_empty_node_13
+	.globl	add_empty_node_14
+	.globl	add_empty_node_15
+	.globl	add_empty_node_16
+	.globl	add_empty_node_17
+	.globl	add_empty_node_18
+	.globl	add_empty_node_19
+	.globl	add_empty_node_20
+	.globl	add_empty_node_21
+	.globl	add_empty_node_22
+	.globl	add_empty_node_23
+	.globl	add_empty_node_24
+	.globl	add_empty_node_25
+	.globl	add_empty_node_26
+	.globl	add_empty_node_27
+	.globl	add_empty_node_28
+	.globl	add_empty_node_29
+	.globl	add_empty_node_30
+	.globl	add_empty_node_31
+	.globl	add_empty_node_32
+
+	.globl	yet_args_needed_5
+	.globl	yet_args_needed_6
+	.globl	yet_args_needed_7
+	.globl	yet_args_needed_8
+	.globl	yet_args_needed_9
+	.globl	yet_args_needed_10
+	.globl	yet_args_needed_11
+	.globl	yet_args_needed_12
+	.globl	yet_args_needed_13
+	.globl	yet_args_needed_14
+	.globl	yet_args_needed_15
+	.globl	yet_args_needed_16
+	.globl	yet_args_needed_17
+	.globl	yet_args_needed_18
+	.globl	yet_args_needed_19
+	.globl	yet_args_needed_20
+	.globl	yet_args_needed_21
+	.globl	yet_args_needed_22
+	.globl	yet_args_needed_23
+	.globl	yet_args_needed_24
+	.globl	yet_args_needed_25
+	.globl	yet_args_needed_26
+	.globl	yet_args_needed_27
+	.globl	yet_args_needed_28
+	.globl	yet_args_needed_29
+	.globl	yet_args_needed_30
+	.globl	yet_args_needed_31
+
+ap_32:
+	mov	r9,qword ptr [r8]
+	mov	rbx,32*16
+	cmp	word ptr [r9],bx
+	je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap32
+
+ap_31:
+	mov	r9,qword ptr [r8]
+	mov	rbx,31*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap31
+
+ap_30:
+	mov	r9,qword ptr [r8]
+	mov	rbx,30*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap30
+
+ap_29:
+	mov	r9,qword ptr [r8]
+	mov	rbx,29*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap29
+
+ap_28:
+	mov	r9,qword ptr [r8]
+	mov	rbx,28*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap28
+
+ap_27:
+	mov	r9,qword ptr [r8]
+	mov	rbx,27*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap27
+
+ap_26:
+	mov	r9,qword ptr [r8]
+	mov	rbx,26*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap26
+
+ap_25:
+	mov	r9,qword ptr [r8]
+	mov	rbx,25*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap25
+
+ap_24:
+	mov	r9,qword ptr [r8]
+	mov	rbx,24*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap24
+
+ap_23:
+	mov	r9,qword ptr [r8]
+	mov	rbx,23*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap23
+
+ap_22:
+	mov	r9,qword ptr [r8]
+	mov	rbx,22*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap22
+
+ap_21:
+	mov	r9,qword ptr [r8]
+	mov	rbx,21*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap21
+
+ap_20:
+	mov	r9,qword ptr [r8]
+	mov	rbx,20*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap20
+
+ap_19:
+	mov	r9,qword ptr [r8]
+	mov	rbx,19*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap19
+
+ap_18:
+	mov	r9,qword ptr [r8]
+	mov	rbx,18*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap18
+
+ap_17:
+	mov	r9,qword ptr [r8]
+	mov	rbx,17*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap17
+
+ap_16:
+	mov	r9,qword ptr [r8]
+	mov	rbx,16*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap16
+
+ap_15:
+	mov	r9,qword ptr [r8]
+	mov	rbx,15*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap15
+
+ap_14:
+	mov	r9,qword ptr [r8]
+	mov	rbx,14*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap14
+
+ap_13:
+	mov	r9,qword ptr [r8]
+	mov	rbx,13*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap13
+
+ap_12:
+	mov	r9,qword ptr [r8]
+	mov	rbx,12*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap12
+
+ap_11:
+	mov	r9,qword ptr [r8]
+	mov	rbx,11*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap11
+
+ap_10:
+	mov	r9,qword ptr [r8]
+	mov	rbx,10*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap10
+
+ap_9:
+	mov	r9,qword ptr [r8]
+	mov	rbx,9*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap9
+
+ap_8:
+	mov	r9,qword ptr [r8]
+	mov	rbx,8*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap8
+
+ap_7:
+	mov	r9,qword ptr [r8]
+	mov	rbx,7*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap7
+
+ap_6:
+	mov	r9,qword ptr [r8]
+	mov	rbx,6*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap6
+
+ap_5:
+	mov	r9,qword ptr [r8]
+	mov	rbx,5*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap5
+
+ap_4:
+	mov	r9,qword ptr [r8]
+	mov	rbx,4*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap4
+
+ap_3:
+	mov	r9,qword ptr [r8]
+	mov	rbx,3*16
+	cmp	word ptr [r9],bx
+	att_je	fast_ap
+
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap3
+
+ap_2:
+	mov	r9,qword ptr [r8]
+	mov	rbx,2*16
+	cmp	word ptr [r9],bx
+	jne	no_fast_ap2_
+
+fast_ap_2_2_:
+	movzx	rax,word ptr -2[r9]
+	add	rbx,r9
+	mov	r9,qword ptr -10[rbx]
+ .if PROFILE
+	sub	r9,24
+ .else
+	sub	r9,12
+ .endif
+	cmp	rax,1
+	jb	repl_args_0_2
+	je	repl_args_1
+
+	cmp	rax,3
+	jb	repl_args_2
+
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr 8[rsi],rdx
+	lea	rsi,16[rsi]
+	mov	rdx,qword ptr 16[r8]
+
+	jmp	fast_ap_
+
+no_fast_ap2_:
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	mov	rdx,r8
+	mov	r8,qword ptr 6[r9]
+	add	rsi,8
+	jmp	no_fast_ap2
+
+fast_ap_2_2:
+	mov	r8,rdx
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+	att_jmp	fast_ap_2_2_
+
+fast_ap_2:
+	mov	r8,rdx
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+fast_ap:
+	movzx	rax,word ptr -2[r9]
+	add	rbx,r9
+	mov	r9,qword ptr -10[rbx]
+ .if PROFILE
+	sub	r9,24
+ .else
+	sub	r9,12
+ .endif
+	cmp	rax,1
+	jb	repl_args_0
+	att_je	repl_args_1
+
+	cmp	rax,3
+	att_jb	repl_args_2
+
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr 8[rsi],rdx
+	lea	rsi,16[rsi]
+	mov	rdx,qword ptr 16[r8]
+
+fast_ap_:
+	mov	r8,qword ptr 8[r8]
+	je	repl_args_3
+
+	cmp	rax,5
+	jb	repl_args_4
+	je	repl_args_5
+
+	cmp	rax,7
+	jb	repl_args_6
+
+repl_args_7_:
+	mov	rbp,qword ptr -16[rdx+rax*8]
+	mov	qword ptr [rsi],rbp
+	sub	rax,1
+	add	rsi,8
+	cmp	rax,6
+	att_jne	repl_args_7_
+
+repl_args_6:
+	mov	rax,qword ptr 32[rdx]
+	mov	qword ptr [rsi],rax
+	mov	rax,qword ptr 24[rdx]
+	mov	qword ptr 8[rsi],rax
+	mov	rax,qword ptr 16[rdx]
+	mov	qword ptr 16[rsi],rax
+	mov	rcx,qword ptr 8[rdx]
+	mov	rdx,qword ptr [rdx]
+	add	rsi,24
+	jmp	r9
+
+repl_args_0:
+	mov	r8,rdx
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+repl_args_0_2:
+	jmp	r9
+
+repl_args_1:
+	mov	r8,qword ptr 8[r8]
+	jmp	r9
+
+repl_args_2:
+	mov	qword ptr [rsi],rcx
+	mov	rcx,rdx
+	add	rsi,8
+	mov	rdx,qword ptr 16[r8]
+	mov	r8,qword ptr 8[r8]
+	jmp	r9
+
+repl_args_3:
+	mov	rcx,qword ptr 8[rdx]
+	mov	rdx,qword ptr [rdx]
+	jmp	r9
+
+repl_args_4:
+	mov	rax,qword ptr 16[rdx]
+	mov	qword ptr [rsi],rax
+	mov	rcx,qword ptr 8[rdx]
+	mov	rdx,qword ptr [rdx]
+	add	rsi,8
+	jmp	r9
+
+repl_args_5:
+	mov	rax,qword ptr 24[rdx]
+	mov	qword ptr [rsi],rax
+	mov	rax,qword ptr 16[rdx]
+	mov	qword ptr 8[rsi],rax
+	mov	rcx,qword ptr 8[rdx]
+	mov	rdx,qword ptr [rdx]
+	add	rsi,16
+	jmp	r9
+
+
+no_fast_ap32:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,31*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap31:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,30*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap30:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,29*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap29:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,28*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap28:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,27*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap27:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,26*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap26:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,25*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap25:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,24*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap24:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,23*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap23:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,22*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap22:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,21*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap21:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,20*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap20:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,19*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap19:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,18*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap18:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,17*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap17:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,16*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap16:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,15*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap15:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,14*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap14:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,13*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap13:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,12*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap12:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,11*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap11:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,10*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap10:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,9*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap9:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,8*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap8:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,7*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap7:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,6*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap6:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,5*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap5:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,4*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap4:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,3*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap3:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+
+	mov	rbx,2*16
+ 	cmp	word ptr [r9],bx
+	att_je	fast_ap_2
+
+	mov	r8,qword ptr 6[r9]
+no_fast_ap2:
+	call	r8
+	mov	r9,qword ptr [rcx]
+	mov	rdx,rcx
+	mov	rcx,qword ptr -8[rsi]
+	sub	rsi,8
+	jmp	qword ptr 6[r9]
+
+
+add_empty_node_2:
+	sub	r15,3
+	jb	add_empty_node_2_gc
+add_empty_node_2_gc_:
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	mov	r8,rdx
+	mov	rdx,rcx
+	mov	rcx,rdi
+	add	rdi,24
+	ret
+add_empty_node_2_gc:
+	att_call	collect_2
+	att_jmp	add_empty_node_2_gc_
+
+add_empty_node_3:
+	sub	r15,3
+	jb	add_empty_node_3_gc
+add_empty_node_3_gc_:
+	mov	qword ptr [rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_3_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_3_gc_
+
+add_empty_node_4:
+	sub	r15,3
+	jb	add_empty_node_4_gc
+add_empty_node_4_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	qword ptr -8[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_4_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_4_gc_
+
+add_empty_node_5:
+	sub	r15,3
+	jb	add_empty_node_5_gc
+add_empty_node_5_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	qword ptr -16[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_5_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_5_gc_
+
+add_empty_node_6:
+	sub	r15,3
+	jb	add_empty_node_6_gc
+add_empty_node_6_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	qword ptr -24[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_6_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_6_gc_
+
+add_empty_node_7:
+	sub	r15,3
+	jb	add_empty_node_7_gc
+add_empty_node_7_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	r9,qword ptr -32[rsi]
+	mov	qword ptr -24[rsi],r9
+	mov	qword ptr -32[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_7_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_7_gc_
+
+add_empty_node_8:
+	sub	r15,3
+	jb	add_empty_node_8_gc
+add_empty_node_8_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	r9,qword ptr -32[rsi]
+	mov	qword ptr -24[rsi],r9
+	mov	r9,qword ptr -40[rsi]
+	mov	qword ptr -32[rsi],r9
+	mov	qword ptr -40[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_8_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_8_gc_
+
+add_empty_node_9:
+	sub	r15,3
+	jb	add_empty_node_9_gc
+add_empty_node_9_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	r9,qword ptr -32[rsi]
+	mov	qword ptr -24[rsi],r9
+	mov	r9,qword ptr -40[rsi]
+	mov	qword ptr -32[rsi],r9
+	mov	r9,qword ptr -48[rsi]
+	mov	qword ptr -40[rsi],r9
+	mov	qword ptr -48[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_9_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_9_gc_
+
+add_empty_node_10:
+	sub	r15,3
+	jb	add_empty_node_10_gc
+add_empty_node_10_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	r9,qword ptr -32[rsi]
+	mov	qword ptr -24[rsi],r9
+	mov	r9,qword ptr -40[rsi]
+	mov	qword ptr -32[rsi],r9
+	mov	r9,qword ptr -48[rsi]
+	mov	qword ptr -40[rsi],r9
+	mov	r9,qword ptr -56[rsi]
+	mov	qword ptr -48[rsi],r9
+	mov	qword ptr -56[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_10_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_10_gc_
+
+add_empty_node_11:
+	sub	r15,3
+	jb	add_empty_node_11_gc
+add_empty_node_11_gc_:
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+	mov	r9,qword ptr -32[rsi]
+	mov	qword ptr -24[rsi],r9
+	mov	r9,qword ptr -40[rsi]
+	mov	qword ptr -32[rsi],r9
+	mov	r9,qword ptr -48[rsi]
+	mov	qword ptr -40[rsi],r9
+	mov	r9,qword ptr -56[rsi]
+	mov	qword ptr -48[rsi],r9
+	mov	r9,qword ptr -64[rsi]
+	mov	qword ptr -56[rsi],r9
+	mov	qword ptr -64[rsi],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_11_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_11_gc_
+
+add_empty_node_32:
+	mov	rbx,7
+	att_jmp	add_empty_node_12_
+add_empty_node_28:
+	mov	rbx,6
+	att_jmp	add_empty_node_12_
+add_empty_node_24:
+	mov	rbx,5
+	att_jmp	add_empty_node_12_
+add_empty_node_20:
+	mov	rbx,4
+	att_jmp	add_empty_node_12_
+add_empty_node_16:
+	mov	rbx,3
+	att_jmp	add_empty_node_12_
+add_empty_node_12:
+	mov	rbx,2
+add_empty_node_12_:
+	sub	r15,3
+	jb	add_empty_node_12_gc
+add_empty_node_12_gc_:
+	mov	rax,rsi
+add_empty_node_12_lp:
+	mov	r9,qword ptr -8[rax]
+	mov	qword ptr [rax],r9
+	mov	r9,qword ptr -16[rax]
+	mov	qword ptr -8[rax],r9
+	mov	r9,qword ptr -24[rax]
+	mov	qword ptr -16[rax],r9
+	mov	r9,qword ptr -32[rax]
+	mov	qword ptr -24[rax],r9
+	sub	rax,32
+	sub	rbx,1
+	att_jne	add_empty_node_12_lp
+	mov	qword ptr [rax],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_12_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_12_gc_
+
+add_empty_node_29:
+	mov	rbx,6
+	jmp	add_empty_node_13_
+add_empty_node_25:
+	mov	rbx,5
+	att_jmp	add_empty_node_13_
+add_empty_node_21:
+	mov	rbx,4
+	att_jmp	add_empty_node_13_
+add_empty_node_17:
+	mov	rbx,3
+	att_jmp	add_empty_node_13_
+add_empty_node_13:
+	mov	rbx,2
+add_empty_node_13_:
+	sub	r15,3
+	jb	add_empty_node_13_gc
+add_empty_node_13_gc_:
+	mov	rax,rsi
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+add_empty_node_13_lp:
+	mov	r9,qword ptr -16[rax]
+	mov	qword ptr -8[rax],r9
+	mov	r9,qword ptr -24[rax]
+	mov	qword ptr -16[rax],r9
+	mov	r9,qword ptr -32[rax]
+	mov	qword ptr -24[rax],r9
+	mov	r9,qword ptr -40[rax]
+	mov	qword ptr -32[rax],r9
+	sub	rax,32
+	sub	rbx,1
+	att_jne	add_empty_node_13_lp
+	mov	qword ptr -8[rax],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_13_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_13_gc_
+
+add_empty_node_30:
+	mov	rbx,6
+	att_jmp	add_empty_node_14_
+add_empty_node_26:
+	mov	rbx,5
+	att_jmp	add_empty_node_14_
+add_empty_node_22:
+	mov	rbx,4
+	att_jmp	add_empty_node_14_
+add_empty_node_18:
+	mov	rbx,3
+	att_jmp	add_empty_node_14_
+add_empty_node_14:
+	mov	rbx,2
+add_empty_node_14_:
+	sub	r15,3
+	jb	add_empty_node_14_gc
+add_empty_node_14_gc_:
+	mov	rax,rsi
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	rsi,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+add_empty_node_14_lp:
+	mov	r9,qword ptr -24[rax]
+	mov	qword ptr -16[rax],r9
+	mov	r9,qword ptr -32[rax]
+	mov	qword ptr -24[rax],r9
+	mov	r9,qword ptr -40[rax]
+	mov	qword ptr -32[rax],r9
+	mov	r9,qword ptr -48[rax]
+	mov	qword ptr -40[rax],r9
+	sub	rax,32
+	sub	rbx,1
+	att_jne	add_empty_node_14_lp
+	mov	qword ptr -16[rax],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_14_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_14_gc_
+
+add_empty_node_31:
+	mov	rbx,6
+	att_jmp	add_empty_node_15_
+add_empty_node_27:
+	mov	rbx,5
+	att_jmp	add_empty_node_15_
+add_empty_node_23:
+	mov	rbx,4
+	att_jmp	add_empty_node_15_
+add_empty_node_19:
+	mov	rbx,3
+	att_jmp	add_empty_node_15_
+add_empty_node_15:
+	mov	rbx,2
+add_empty_node_15_:
+	sub	r15,3
+	jb	add_empty_node_15_gc
+add_empty_node_15_gc_:
+	mov	rax,r9
+	mov	r9,qword ptr -8[rsi]
+	mov	qword ptr [rsi],r9
+	mov	r9,qword ptr -16[rsi]
+	mov	qword ptr -8[rsi],r9
+	mov	r9,qword ptr -24[rsi]
+	mov	qword ptr -16[rsi],r9
+add_empty_node_15_lp:
+	mov	r9,qword ptr -32[rax]
+	mov	qword ptr -24[rax],r9
+	mov	r9,qword ptr -40[rax]
+	mov	qword ptr -32[rax],r9
+	mov	r9,qword ptr -48[rax]
+	mov	qword ptr -40[rax],r9
+	mov	r9,qword ptr -56[rax]
+	mov	qword ptr -48[rax],r9
+	sub	rax,32
+	sub	rbx,1
+	att_jne	add_empty_node_15_lp
+	mov	qword ptr -24[rax],rdi
+	lea	rbp,__cycle__in__spine[rip]
+	mov	qword ptr [rdi],rbp
+	add	rsi,8
+	add	rdi,24
+	ret
+add_empty_node_15_gc:
+	att_call	collect_3
+	att_jmp	add_empty_node_15_gc_
+
+
+yet_args_needed_0:
+	sub	r15,2
+	jb	yet_args_needed_0_gc
+yet_args_needed_0_gc_r:
+	mov	qword ptr 8[rdi],rcx
+	mov	rax,qword ptr [rdx]
+	mov	rcx,rdi
+	add	rax,16
+	mov	qword ptr [rdi],rax
+	add	rdi,16
+	ret
+
+yet_args_needed_0_gc:
+	att_call	collect_2
+	att_jmp	yet_args_needed_0_gc_r
+
+
+	.align	2
+	sub	r15,3
+	jae short build_node_2_gc_r
+	jmp short build_node_2_gc
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+ .endif
+yet_args_needed_1:
+	sub	r15,3
+	jb	yet_args_needed_1_gc
+yet_args_needed_1_gc_r:
+	mov	qword ptr 16[rdi],rcx
+	mov	rax,qword ptr [rdx]
+	mov	rcx,rdi
+	add	rax,16
+	mov	qword ptr [rdi],rax
+	mov	rbx,qword ptr 8[rdx]
+	mov	qword ptr 8[rdi],rbx
+	add	rdi,24
+	ret
+
+yet_args_needed_1_gc:
+	att_call	collect_2
+	att_jmp	yet_args_needed_1_gc_r
+
+build_node_2_gc_r:
+	mov	qword ptr [rdi],rbx
+	mov	qword ptr 8[rdi],rdx
+	mov	qword ptr 16[rdi],rcx
+	mov	rcx,rdi
+	add	rdi,24
+	ret
+
+build_node_2_gc:
+	att_call	collect_2
+	att_jmp	build_node_2_gc_r
+
+
+	.align	2
+	sub	r15,5
+	jae short build_node_3_gc_r
+	jmp short build_node_3_gc
+	nop
+	.align	2
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_2:
+	sub	r15,5
+	jb	gc_22
+gc_r_22:
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr 8[rdi],rcx
+	add	rax,16
+	mov	r10,qword ptr 8[rdx]
+	mov	qword ptr 16[rdi],rax
+	lea	rcx,16[rdi]
+	mov	rbp,qword ptr 16[rdx]
+	mov	qword ptr 24[rdi],r10
+	mov	qword ptr [rdi],rbp
+	mov	qword ptr 32[rdi],rdi
+	add	rdi,40
+	ret
+
+gc_22:	att_call	collect_2
+	att_jmp	gc_r_22
+
+build_node_3_gc_r:
+	mov	qword ptr [rdi],rbx
+	lea	rbp,24[rdi]
+	mov	qword ptr 8[rdi],r8
+	mov	qword ptr 16[rdi],rbp
+	mov	qword ptr 24[rdi],rdx
+	mov	qword ptr 32[rdi],rcx
+	mov	rcx,rdi
+	add	rdi,40
+	ret
+
+build_node_3_gc:
+	att_call	collect_2
+	att_jmp	build_node_3_gc_r
+
+
+	.align	2
+	sub	r15,6
+	jae short build_node_4_gc_r
+	jmp short build_node_4_gc
+	nop
+	.align	2
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_3:
+	sub	r15,6
+	jb	gc_23
+gc_r_23:
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr 16[rdi],rcx
+	add	rax,16
+	mov	r10,qword ptr 8[rdx]
+	mov	qword ptr 24[rdi],rax
+	mov	rdx,qword ptr 16[rdx]
+	mov	qword ptr 32[rdi],r10
+	mov	rbp,qword ptr [rdx]
+	mov	qword ptr 40[rdi],rdi
+	mov	qword ptr [rdi],rbp
+	mov	rbp,qword ptr 8[rdx]
+	lea	rcx,24[rdi]
+	mov	qword ptr 8[rdi],rbp
+	add	rdi,48
+	ret
+
+gc_23:	att_call	collect_2
+	att_jmp	gc_r_23
+
+build_node_4_gc_r:
+	mov	qword ptr [rdi],rbx
+	lea	rbp,24[rdi]
+	mov	qword ptr 8[rdi],r8
+	mov	qword ptr 16[rdi],rbp
+	mov	qword ptr 24[rdi],rdx
+	mov	qword ptr 32[rdi],rcx
+	mov	rcx,rdi
+	mov	rbp,qword ptr -8[rsi]
+	mov	qword ptr 40[rdi],rbp
+	sub	rsi,8
+	add	rdi,48
+	ret
+
+build_node_4_gc:
+	att_call	collect_2
+	att_jmp	build_node_4_gc_r
+
+
+	.align	2
+	sub	r15,7
+	jae short build_node_5_gc_r
+	jmp 	build_node_5_gc
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_4:
+	sub	r15,7
+	jb	gc_24
+gc_r_24:
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr 24[rdi],rcx
+	add	rax,16
+	mov	r10,qword ptr 8[rdx]
+	mov	qword ptr 32[rdi],rax
+	mov	rdx,qword ptr 16[rdx]
+	mov	qword ptr 40[rdi],r10
+	mov	rbp,qword ptr [rdx]
+	mov	48[rdi],rdi
+	mov	qword ptr [rdi],rbp
+	mov	rbp,qword ptr 8[rdx]
+	lea	rcx,32[rdi]
+	mov	qword ptr 8[rdi],rbp
+	mov	rbp,qword ptr 16[rdx]
+	mov	qword ptr 16[rdi],rbp
+	add	rdi,56
+	ret
+
+gc_24:	att_call	collect_2
+	att_jmp	gc_r_24
+
+build_node_5_gc_r:
+	mov	qword ptr [rdi],rbx
+	lea	rbp,24[rdi]
+	mov	qword ptr 8[rdi],r8
+	mov	qword ptr 16[rdi],rbp
+	mov	qword ptr 24[rdi],rdx
+	mov	qword ptr 32[rdi],rcx
+	mov	rcx,rdi
+	mov	rbp,qword ptr -8[rsi]
+	mov	qword ptr 40[rdi],rbp
+	mov	rbp,qword ptr -16[rsi]
+	mov	qword ptr 48[rdi],rbp
+	sub	rsi,16
+	add	rdi,56
+	ret
+
+build_node_5_gc:
+	att_call	collect_2
+	att_jmp	build_node_5_gc_r
+
+
+	.align	2
+	mov	eax,8
+	jmp	build_node_
+	nop
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_5:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,8
+	jmp	yet_args_needed_
+
+
+	.align	2
+	mov	eax,9
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_6:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,9
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,10
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_7:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,10
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,11
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_8:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,11
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,12
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_9:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,12
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,13
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_10:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,13
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,14
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2
+ .if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+ .endif
+yet_args_needed_11:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,14
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,15
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_12:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,15
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,16
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_13:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,16
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,17
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_14:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,17
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,18
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_15:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,18
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,19
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_16:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,19
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,20
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_17:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,20
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,21
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_18:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,21
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,22
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_19:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,22
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,23
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_20:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,23
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,24
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_21:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,24
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,25
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_22:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,25
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,26
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_23:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,26
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,27
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_24:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,27
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,28
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_25:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,28
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,29
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_26:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,29
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,30
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_27:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,30
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,31
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_28:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,31
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,32
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_29:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,32
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,33
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2	
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_30:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,33
+	att_jmp	yet_args_needed_
+
+	.align	2
+	mov	eax,34
+	att_jmp	build_node_
+	nop
+	nop
+	.align	2
+.if PROFILE
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop
+	nop	
+	nop
+	nop
+	nop
+	nop	
+.endif
+yet_args_needed_31:
+	mov	rbx,qword ptr [rdx]
+	mov	rax,34
+	att_jmp	yet_args_needed_
+
+yet_args_needed:
+	mov	rbx,qword ptr [rdx]
+	movzx	rax,word ptr -2[rbx]
+	add	rax,3
+yet_args_needed_:
+	sub	r15,rax
+	jb	yet_args_needed_gc
+yet_args_needed_gc_r:
+	mov	r11,qword ptr 8[rdx]
+	sub	rax,3+1+4
+	mov	rdx,qword ptr 16[rdx]
+	mov	r10,rdi
+	mov	rbp,qword ptr [rdx]
+	mov	qword ptr [rdi],rbp
+	mov	rbp,qword ptr 8[rdx]
+	mov	qword ptr 8[rdi],rbp
+	mov	rbp,qword ptr 16[rdx]
+	mov	qword ptr 16[rdi],rbp
+	add	rdx,24
+	add	rdi,24
+
+yet_args_needed_cp_a:
+	mov	rbp,qword ptr [rdx]
+	add	rdx,8
+	mov	qword ptr [rdi],rbp
+	add	rdi,8
+	sub	rax,1
+	att_jge	yet_args_needed_cp_a
+
+	mov	qword ptr [rdi],rcx
+	add	rbx,16
+	mov	qword ptr 8[rdi],rbx
+	lea	rcx,8[rdi]
+	mov	qword ptr 16[rdi],r11
+	mov	qword ptr 24[rdi],r10
+	add	rdi,32
+	ret
+
+yet_args_needed_gc:
+	att_call	collect_2
+	att_jmp	yet_args_needed_gc_r
+
+build_node_:
+	sub	r15,rax
+	jb	build_node_gc
+build_node_gc_r:
+	mov	qword ptr [rdi],rbx
+	lea	rbp,24[rdi]
+	mov	qword ptr 8[rdi],r8
+	mov	qword ptr 16[rdi],rbp
+	mov	qword ptr 24[rdi],rdx
+	mov	qword ptr 32[rdi],rcx
+	mov	rcx,rdi
+	mov	r8,qword ptr -8[rsi]
+	mov	qword ptr 40[rdi],r8
+	mov	r8,qword ptr -16[rsi]
+	sub	rsi,16
+	mov	qword ptr 48[rdi],r8
+	add	rdi,56
+
+	sub	rax,5+2
+build_node_cp_a:
+	mov	r8,qword ptr -8[rsi]
+	sub	rsi,8
+	mov	qword ptr [rdi],r8
+	add	rdi,8
+	sub	rax,1
+	att_jne	build_node_cp_a
+
+	ret
+
+build_node_gc:
+	att_call	collect_3
+	att_jmp	build_node_gc_r
+	
diff --git a/macho64/acompact.s b/macho64/acompact.s
new file mode 100644
index 0000000..44251d4
--- /dev/null
+++ b/macho64/acompact.s
@@ -0,0 +1,1184 @@
+
+/* mark used nodes and pointers in argument parts and link backward pointers */
+
+	mov	rax,qword ptr heap_size_65[rip]
+	shl	rax,6
+	mov	qword ptr heap_size_64_65[rip],rax 
+
+	lea	rax,(-16000)[rsp]
+	mov	qword ptr end_stack[rip],rax
+
+	mov	rax,qword ptr caf_list[rip]
+
+	test	qword ptr _flags[rip],4096
+	jne	pmarkr
+
+	test	rax,rax 
+	je	end_mark_cafs
+
+mark_cafs_lp:
+	push	(-8)[rax]
+
+	lea	rsi,8[rax]
+	mov	rax,qword ptr [rax]
+	lea	rcx,[rsi+rax*8]
+
+	mov	qword ptr end_vector[rip],rcx
+
+	call	rmark_stack_nodes
+	
+	pop	rax
+	test	rax,rax 
+	att_jne	mark_cafs_lp
+
+end_mark_cafs:
+	mov	rsi,qword ptr stack_p[rip]
+
+	mov	rcx,qword ptr stack_top[rip]
+	mov	qword ptr end_vector[rip],rcx 
+
+	att_call	rmark_stack_nodes
+
+	att_call	add_mark_compact_garbage_collect_time
+	
+	jmp	compact_heap
+
+pmarkr:
+	test	rax,rax 
+	je	end_rmarkp_cafs
+
+rmarkp_cafs_lp:
+	push	(-8)[rax]
+
+	lea	rsi,8[rax]
+	mov	rax,qword ptr [rax]
+	lea	rcx,[rsi+rax*8]
+
+	mov	qword ptr end_vector[rip],rcx
+
+	call	rmarkp_stack_nodes
+	
+	pop	rax
+	test	rax,rax 
+	att_jne	rmarkp_cafs_lp
+
+end_rmarkp_cafs:
+	mov	rsi,qword ptr stack_p[rip]
+
+	mov	rcx,qword ptr stack_top[rip]
+	mov	qword ptr end_vector[rip],rcx 
+
+	att_call	rmarkp_stack_nodes
+
+	att_call	add_mark_compact_garbage_collect_time
+	
+	att_jmp	compact_heap
+
+	.include	"acompact_rmark.s"
+
+	.include	"acompact_rmark_prefetch.s"
+
+	.include "acompact_rmarkr.s"
+
+/* compact the heap */
+
+compact_heap:
+
+	lea	rcx,finalizer_list[rip]
+	lea	rdx,free_finalizer_list[rip]
+
+	mov	rbp,qword ptr [rcx]
+determine_free_finalizers_after_compact1:
+	lea	r9,__Nil-8[rip]
+	cmp	rbp,r9
+	je	end_finalizers_after_compact1
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rbp 
+	mov	rbx,rax 
+	and	rax,31*9
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rax]
+	test	esi,dword ptr [rdi+rbx*4]
+	je	finalizer_not_used_after_compact1
+
+	mov	rax,qword ptr [rbp]
+	mov	rsi,rbp 
+	jmp	finalizer_find_descriptor
+
+finalizer_find_descriptor_lp:
+	and	rax,-4
+	mov	rsi,rax 
+	mov	rax,qword ptr [rax]
+finalizer_find_descriptor:
+	test	rax,1
+	att_jne	finalizer_find_descriptor_lp
+
+	lea	r9,e____system__kFinalizerGCTemp+2[rip]
+	mov	qword ptr [rsi],r9
+
+	cmp	rbp,rcx 
+	ja	finalizer_no_reverse
+
+	mov	rax,qword ptr [rbp]
+	lea	rsi,1[rcx]
+	mov	qword ptr [rbp],rsi 
+	mov	qword ptr [rcx],rax 
+
+finalizer_no_reverse:
+	lea	rcx,8[rbp]
+	mov	rbp,qword ptr 8[rbp]
+	att_jmp	determine_free_finalizers_after_compact1
+
+finalizer_not_used_after_compact1:
+	lea	r9,e____system__kFinalizerGCTemp+2[rip]
+	mov	qword ptr [rbp],r9
+
+	mov	qword ptr [rdx],rbp 
+	lea	rdx,8[rbp]
+
+	mov	rbp ,qword ptr 8[rbp]
+	mov	qword ptr [rcx],rbp 
+
+	att_jmp	determine_free_finalizers_after_compact1
+
+end_finalizers_after_compact1:
+	mov	qword ptr [rdx],rbp 
+
+	mov	rcx,qword ptr finalizer_list[rip]
+	lea	r9,__Nil-8[rip]
+	cmp	rcx,r9
+	je	finalizer_list_empty
+	test	rcx,3
+	jne	finalizer_list_already_reversed
+	mov	rax ,qword ptr [rcx]
+	lea	r9,finalizer_list+1[rip]
+	mov	qword ptr [rcx],r9
+	mov	qword ptr finalizer_list[rip],rax 
+finalizer_list_already_reversed:
+finalizer_list_empty:
+
+	lea	rbp,free_finalizer_list[rip]
+	lea	r9,__Nil-8[rip]
+	cmp	qword ptr [rbp],r9
+	je	free_finalizer_list_empty
+
+	lea	r9,free_finalizer_list+8[rip]
+	mov	qword ptr end_vector[rip],r9
+
+	test	qword ptr _flags[rip],4096
+	je	no_pmarkr
+	att_call	rmarkp_stack_nodes
+	att_jmp	free_finalizer_list_empty
+no_pmarkr:
+	att_call	rmark_stack_nodes
+
+free_finalizer_list_empty:
+
+	mov	rax,qword ptr heap_size_65[rip]
+	mov	rbx,rax 
+	shl	rbx,6
+
+	add	rbx,qword ptr heap_p3[rip]
+
+	mov	qword ptr end_heap_p3[rip],rbx
+
+	add	rax,3
+	shr	rax,2
+	mov	r12,rax
+	
+	mov	r8,qword ptr heap_vector[rip]
+
+	lea	rbx,4[r8]
+	neg	rbx 
+	mov	qword ptr neg_heap_vector_plus_4[rip],rbx
+
+	mov	rdi,qword ptr heap_p3[rip]
+	xor	rsi,rsi 
+	jmp	skip_zeros
+
+/* %rax ,%rcx ,%rbp : free */
+find_non_zero_long:
+skip_zeros:
+	sub	r12,1
+	jc	end_move
+	mov	esi,dword ptr [r8]
+	add	r8,4
+	test	rsi,rsi 
+	att_je	skip_zeros
+/* %rbp : free */
+end_skip_zeros:
+	mov	rbp,qword ptr neg_heap_vector_plus_4[rip]
+
+	add	rbp,r8
+
+	shl	rbp,6
+	add	rbp,qword ptr heap_p3[rip]
+
+bsf_and_copy_nodes:
+	movzx	rax,sil
+	lea	r9,first_one_bit_table[rip]
+	test	rax,rax
+	jne	found_bit1
+	movzx	rcx,si
+	shr	rcx,8
+	jne	found_bit2
+	mov	rax,rsi 
+	and	rax,0x0ff0000
+	jne	found_bit3
+	mov	rcx,rsi
+	shr	rcx,24
+	movzx	rcx,byte ptr [r9+rcx*1]
+	add	rcx,24
+	jmp	copy_nodes
+
+found_bit3:
+	shr	rax,16
+	movzx	rcx,byte ptr [r9+rax*1]
+	add	rcx,16
+	att_jmp	copy_nodes
+
+found_bit2:
+	movzx	rcx,byte ptr [r9+rcx*1]
+	add	rcx,8
+	att_jmp	copy_nodes
+
+found_bit1:
+	movzx	rcx,byte ptr [r9+rax*1]
+
+copy_nodes:
+	mov	rax,qword ptr [rbp+rcx*8]
+	shr	esi,1
+	lea	rbp,8[rbp+rcx*8]
+	shr	esi,cl
+	mov	rcx,rbp
+
+	dec	rax
+
+	test	rax,2
+	je	begin_update_list_2
+
+move_argument_part:
+	mov	rbx,qword ptr (-18)[rax]
+	sub	rax,2
+
+	test	rbx,1
+	je	end_list_2
+find_descriptor_2:
+	and	rbx,-4
+	mov	rbx,qword ptr [rbx]
+	test	rbx,1
+	att_jne	find_descriptor_2
+
+end_list_2:
+	mov	rdx,rbx 
+	movzx	rbx,word ptr (-2)[rbx]
+	cmp	rbx,256
+	jb	no_record_arguments
+
+	movzx	rdx,word ptr (-2+2)[rdx]
+	sub	rdx,2
+	jae	copy_record_arguments_aa
+
+	sub	rbx,256+3
+
+copy_record_arguments_all_b:
+	push	rbx 
+	mov	rbx,qword ptr heap_vector[rip]
+
+update_up_list_1r:
+	mov	rdx,rax 
+	add	rax,qword ptr neg_heap_p3[rip]
+
+	push	rcx 
+
+	mov	rcx,rax 
+
+	shr	rax,8
+	and	rcx,31*8
+
+	lea	r9,bit_set_table2[rip]
+	mov	ecx,dword ptr [r9+rcx*1]
+	mov	eax,dword ptr [rbx+rax*4]
+
+	and	rax,rcx 
+
+	pop	rcx 
+	je	copy_argument_part_1r
+
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr [rdx],rdi 
+	sub	rax,3
+	att_jmp	update_up_list_1r
+
+copy_argument_part_1r:
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr [rdx],rdi 
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	shr	rax,3
+
+	mov	rbx,rax
+	and	rbx,31
+	cmp	rbx,1
+	jae	bit_in_this_word
+
+	dec	r12
+	mov	esi,dword ptr [r8]
+	add	r8,4
+
+	mov	rbp,qword ptr neg_heap_vector_plus_4[rip]
+	add	rbp,r8
+	shl	rbp,6
+	add	rbp,qword ptr heap_p3[rip]
+
+bit_in_this_word:
+	shr	esi,1
+	add	rbp,8
+
+	pop	rbx 
+
+copy_b_record_argument_part_arguments:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	copy_b_record_argument_part_arguments
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+copy_record_arguments_aa:
+	sub	rbx,256+2
+	sub	rbx,rdx 
+	
+	push	rbx 
+	push	rdx 
+
+update_up_list_2r:
+	mov	rdx,rax 
+	mov	rax,qword ptr [rdx]
+	mov	rbx,3
+	and	rbx,rax 
+	sub	rbx,3
+	jne	copy_argument_part_2r
+
+	mov	qword ptr [rdx],rdi 
+	sub	rax,3
+	att_jmp	update_up_list_2r
+
+copy_argument_part_2r:
+	mov	qword ptr [rdx],rdi 
+	cmp	rax,rcx 
+	jb	copy_record_argument_2
+
+	cmp	rax,qword ptr end_heap_p3[rip]
+	att_jae	copy_record_argument_2
+
+	mov	rdx,rax 
+	mov	rax,qword ptr [rdx]
+	lea	rbx,1[rdi]
+	mov	qword ptr [rdx],rbx 
+copy_record_argument_2:
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+
+	pop	rbx 
+	sub	rbx,1
+	jc	no_pointers_in_record
+
+copy_record_pointers:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jb	copy_record_pointers_2
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jae	copy_record_pointers_2
+
+	mov	rax,qword ptr [rdx]
+	inc	rdi 
+	mov	qword ptr [rdx],rdi 
+	dec	rdi 
+	mov	rdx,rax 
+copy_record_pointers_2:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	copy_record_pointers
+
+no_pointers_in_record:
+	pop	rbx 
+	
+	sub	rbx,1
+	jc	no_non_pointers_in_record
+
+copy_non_pointers_in_record:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	copy_non_pointers_in_record
+
+no_non_pointers_in_record:
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+no_record_arguments:
+	sub	rbx,3
+update_up_list_2:
+	mov	rdx,rax
+	mov	rax,qword ptr [rax]
+	inc	rax
+	mov	qword ptr [rdx],rdi
+	test	al,3
+	jne	copy_argument_part_2
+
+	sub	rax,4
+	att_jmp	update_up_list_2
+
+copy_argument_part_2:
+	dec	rax 
+	cmp	rax,rcx 
+	jc	copy_arguments_1
+
+	cmp	rax,qword ptr end_heap_p3[rip]
+	att_jnc	copy_arguments_1
+
+	mov	rdx,rax
+	mov	rax,qword ptr [rax]
+	inc	rdi 
+	mov	qword ptr [rdx],rdi 
+	dec	rdi 
+copy_arguments_1:
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+
+copy_argument_part_arguments:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	copy_arguments_2
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	copy_arguments_2
+
+	mov	rax,qword ptr [rdx]
+	inc	rdi 
+	mov	qword ptr [rdx],rdi 
+	dec	rdi
+	mov	rdx,rax 
+copy_arguments_2:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	copy_argument_part_arguments
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+update_list_2_:
+	dec	rax 
+	mov	qword ptr [rdx],rdi 
+begin_update_list_2:
+	mov	rdx,rax 
+	mov	rax,qword ptr [rax]
+update_list__2:
+	test	rax,1
+	jz	end_update_list_2
+	test	rax,2
+	att_jz	update_list_2_
+	lea	rdx,(-3)[rax]
+	mov	rax,qword ptr (-3)[rax]
+	att_jmp	update_list__2
+
+end_update_list_2:
+	mov	qword ptr [rdx],rdi 
+
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+
+	test	al,2
+	je	move_lazy_node
+
+	movzx	rbx,word ptr (-2)[rax]
+	test	rbx,rbx 
+	je	move_hnf_0
+
+	cmp	rbx,256
+	jae	move_record
+
+	sub	rbx,2
+	jc	move_hnf_1
+	je	move_hnf_2
+
+move_hnf_3:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_hnf_3_1
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_hnf_3_1
+
+	lea	rax,1[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx 
+move_hnf_3_1:
+	mov	qword ptr [rdi],rdx 
+	
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_hnf_3_2
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_hnf_3_2
+
+	lea	rax,(8+2+1)[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx
+move_hnf_3_2:
+	mov	qword ptr 8[rdi],rdx 
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_hnf_2:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_hnf_2_1
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_hnf_2_1
+
+	lea	rax,1[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx 
+move_hnf_2_1:
+	mov	qword ptr [rdi],rdx 
+	
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_hnf_2_2
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_hnf_2_2
+
+	lea	rax,(8+1)[rdi]
+	mov	rbx ,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx
+move_hnf_2_2:
+	mov	qword ptr 8[rdi],rdx 
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_hnf_1:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_hnf_1_
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_hnf_1_
+
+	lea	rax,1[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx
+move_hnf_1_:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_record:
+	sub	rbx,258
+	jb	move_record_1
+	je	move_record_2
+
+move_record_3:
+	movzx	rbx,word ptr (-2+2)[rax]
+	sub	rbx,1
+	att_ja	move_hnf_3
+
+	mov	rdx,qword ptr [rcx]
+	lea	rcx,8[rcx]
+	jb	move_record_3_1b
+
+move_record_3_1a:
+	cmp	rdx,rcx 
+	att_jb	move_record_3_1b
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jae	move_record_3_1b
+
+	lea	rax,1[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx 
+move_record_3_1b:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jb	move_record_3_2
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jae	move_record_3_2
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+
+	push	rbp 
+
+	add	rax,rdx 
+
+	mov	rbx,qword ptr heap_vector[rip]
+	add	rax,8
+	mov	rbp,rax 
+	and	rbp,31*8
+	shr	rax,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	test	ebp,dword ptr [rbx+rax*4]
+	je	not_linked_record_argument_part_3_b
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdi 
+
+	mov	rbp,rax 
+	and	rbp,31*8
+	shr	rax,8
+	mov	ebp,dword ptr [r9+rbp]
+	or	dword ptr [rbx+rax*4],ebp
+	pop	rbp
+
+	jmp	linked_record_argument_part_3_b
+
+not_linked_record_argument_part_3_b:
+	or	dword ptr [rbx+rax*4],ebp 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdi 
+
+	mov	rbp,rax 
+	and	rbp,31*8
+	shr	rax,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	and	dword ptr [rbx+rax*4],ebp 
+	pop	rbp 
+
+linked_record_argument_part_3_b:
+	mov	rbx,qword ptr [rdx]
+	lea	rax,(2+1)[rdi]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx 
+move_record_3_2:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx
+	shr	rbx,3
+	dec	rbx
+	and	rbx,31
+	cmp	rbx,2
+	jb	bit_in_next_word
+
+	shr	esi,2
+	add	rbp,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+bit_in_next_word:
+	dec	r12
+	mov	esi,dword ptr [r8]
+	add	r8,4
+
+	lea	r9,bit_clear_table[rip]
+	and	esi,dword ptr [r9+rbx*4]
+
+	test	rsi,rsi
+	att_je	skip_zeros
+	att_jmp	end_skip_zeros
+
+move_record_2:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	move_hnf_2
+	jb	move_real_or_file
+
+move_record_2_ab:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jb	move_record_2_1
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jae	move_record_2_1
+
+	lea	rax,1[rdi]
+	mov	rbx ,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx
+move_record_2_1:
+	mov	qword ptr [rdi],rdx 
+	mov	rbx,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr 8[rdi],rbx 
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_record_1:
+	movzx	rbx,word ptr (-2+2)[rax]
+	test	rbx,rbx 
+	att_jne	move_hnf_1
+	jmp	move_int_bool_or_char
+
+move_real_or_file:
+	mov	rax ,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+move_int_bool_or_char:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+copy_normal_hnf_0:
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jb	move_real_file_string_or_array
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+	att_jbe	move_int_bool_or_char
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_real_file_string_or_array:
+	lea	r9,__STRING__+2[rip]
+	cmp	rax,r9
+	att_ja	move_real_or_file
+	jne	move_array
+
+	mov	rax,qword ptr [rcx]
+	add	rax,7
+	shr	rax,3
+
+cp_s_arg_lp3:
+	mov	rbx,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rbx 
+	add	rdi,8
+	sub	rax,1
+	att_jnc	cp_s_arg_lp3
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_array:
+	test	rsi,rsi
+	push	rcx
+	jne	bsf_and_end_array_bit
+
+skip_zeros_a:
+	sub	r12,1
+	mov	esi,dword ptr [r8]
+	add	r8,4
+	test	rsi,rsi
+	att_je	skip_zeros_a
+
+	mov	rbp,qword ptr neg_heap_vector_plus_4[rip]
+	add	rbp,r8
+
+	shl	rbp,6
+
+	add	rbp,qword ptr heap_p3[rip]
+
+bsf_and_end_array_bit:
+	mov	rax,rsi
+	lea	r9,first_one_bit_table[rip]
+	mov	rdx,rsi
+	and	rax,0x0ff
+	jne	a_found_bit1
+	and	rdx,0x0ff00
+	jne	a_found_bit2
+	mov	rax,rsi
+	mov	rdx,rsi
+	and	rax,0x0ff0000
+	jne	a_found_bit3
+	shr	rdx,24
+	movzx	rcx,byte ptr [r9+rdx*1]
+	add	rcx,24
+	jmp	end_array_bit
+a_found_bit3:
+	shr	rax,16
+	movzx	rcx,byte ptr [r9+rax*1]
+	add	rcx,16
+	att_jmp	end_array_bit
+a_found_bit2:
+	shr	rdx,8
+	movzx	rcx,byte ptr [r9+rdx*1]
+	add	rcx,8
+	att_jmp	end_array_bit
+a_found_bit1:
+	movzx	rcx,byte ptr [r9+rax*1]
+
+end_array_bit:
+	lea	rbx,[rbp+rcx*8]
+	shr	esi,1
+	lea	rbp,8[rbp+rcx*8]
+	shr	esi,cl
+	pop	rcx
+
+	cmp	rcx,rbx
+	jne	move_a_array
+
+move_b_array:
+	mov	rdx,qword ptr [rcx]
+	mov	qword ptr [rdi],rdx
+	mov	rbx,qword ptr 8[rcx]
+	add	rcx,8
+
+	movzx	rax,word ptr (-2)[rbx]
+	add	rdi,8
+	test	rax,rax 
+	je	move_strict_basic_array
+
+	sub	rax,256
+	imul	rdx,rax 
+	mov	rax,rdx
+	att_jmp	cp_s_arg_lp3
+
+move_strict_basic_array:
+	mov	rax,rdx
+	lea	r9,dINT+2[rip]
+	cmp	rbx,r9
+	att_jle	cp_s_arg_lp3
+	lea	r9,BOOL+2[rip]
+	cmp	rbx,r9
+	je	move_bool_array
+
+move_int32_or_real32_array:
+	add	rax,1
+	shr	rax,1
+	att_jmp	cp_s_arg_lp3
+
+move_bool_array:
+	add	rax,7
+	shr	rax,3
+	att_jmp	cp_s_arg_lp3
+
+move_a_array:
+	mov	rdx,rbx
+	sub	rbx,rcx
+	shr	rbx,3
+
+	push	rsi
+	sub	rbx,1
+	jb	end_array
+	mov	rsi,qword ptr [rcx]
+
+	mov	rax,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rsi
+
+	mov	qword ptr [rdi],rax 
+
+	mov	rax,qword ptr [rdx]
+
+	mov	rsi,qword ptr 8[rcx]
+	add	rcx,16
+
+	mov	qword ptr [rdx],rsi 
+
+	mov	qword ptr 8[rdi],rax 
+	add	rdi,16
+
+	test	rax,rax 
+	je	st_move_array_lp
+
+	movzx	rsi,word ptr (-2+2)[rax]
+	movzx	rax,word ptr (-2)[rax]
+	sub	rax,256
+	cmp	rax,rsi 
+	att_je	st_move_array_lp
+
+move_array_ab:
+	push	rcx 
+
+	mov	rdx,qword ptr (-16)[rdi]
+	mov	rbx,rsi 
+	imul	rdx,rax 
+	shl	rdx,3
+
+	sub	rax,rbx 
+	add	rdx,rcx 
+	att_call	reorder
+
+	pop	rcx 
+	sub	rbx,1
+	sub	rax,1
+
+	push	rbx 
+	push	rax 
+	push	(-16)[rdi]
+	jmp	st_move_array_lp_ab
+
+move_array_ab_lp1:
+	mov	rax,qword ptr 16[rsp]
+move_array_ab_a_elements:
+	mov	rbx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rbx,rcx 
+	jb	move_array_element_ab
+
+	cmp	rbx,qword ptr end_heap_p3[rip]
+	att_jnc	move_array_element_ab
+
+	mov	rdx,rbx 
+	mov	rbx,qword ptr [rdx]
+	inc	rdi
+	mov	qword ptr [rdx],rdi 
+	dec	rdi
+move_array_element_ab:
+	mov	qword ptr [rdi],rbx 
+	add	rdi,8
+	sub	rax,1
+	att_jnc	move_array_ab_a_elements
+
+	mov	rax,qword ptr 8[rsp]
+move_array_ab_b_elements:
+	mov	rbx,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rbx 
+	add	rdi,8
+	sub	rax,1
+	att_jnc	move_array_ab_b_elements
+
+st_move_array_lp_ab:
+	sub	qword ptr [rsp],1
+	att_jnc	move_array_ab_lp1
+
+	add	rsp,24
+	att_jmp	end_array	
+
+move_array_lp1:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+	add	rdi,8
+	cmp	rax,rcx 
+	jb	move_array_element
+
+	cmp	rax,qword ptr end_heap_p3[rip]
+	att_jnc	move_array_element
+
+	mov	rsi,qword ptr [rax]
+	mov	rdx,rax 
+	mov	qword ptr (-8)[rdi],rsi 
+	lea	rax,(-8+1)[rdi]
+	mov	qword ptr [rdx],rax 
+
+	sub	rbx,1
+	att_jnc	move_array_lp1
+
+	att_jmp	end_array
+
+move_array_element:
+	mov	qword ptr (-8)[rdi],rax 
+st_move_array_lp:
+	sub	rbx,1
+	att_jnc	move_array_lp1
+
+end_array:
+	pop	rsi
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_lazy_node:
+	mov	rdx,rax 
+	movsxd	rbx,dword ptr (-4)[rdx]
+	test	rbx,rbx 
+	je	move_lazy_node_0
+
+	sub	rbx,1
+	jle	move_lazy_node_1
+
+	cmp	rbx,256
+	jge	move_closure_with_unboxed_arguments
+
+move_lazy_node_arguments:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_lazy_node_arguments_
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_lazy_node_arguments_
+
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr [rdi],rax 
+	lea	rax,1[rdi]
+	add	rdi,8
+	mov	qword ptr [rdx],rax 
+	sub	rbx,1
+	att_jnc	move_lazy_node_arguments
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_lazy_node_arguments_:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	move_lazy_node_arguments
+	
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_lazy_node_1:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_lazy_node_1_
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_lazy_node_1_
+
+	lea	rax,1[rdi]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+	mov	rdx,rbx 
+move_lazy_node_1_:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_lazy_node_0:
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_closure_with_unboxed_arguments:
+	je	move_closure_with_unboxed_arguments_1
+	add	rbx,1
+	mov	rax,rbx 
+	and	rbx,255
+	shr	rax,8
+	sub	rbx,rax 
+	je	move_non_pointers_of_closure
+
+	push	rax 
+
+move_closure_with_unboxed_arguments_lp:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	cmp	rdx,rcx 
+	jc	move_closure_with_unboxed_arguments_
+
+	cmp	rdx,qword ptr end_heap_p3[rip]
+	att_jnc	move_closure_with_unboxed_arguments_
+
+	mov	rax,qword ptr [rdx]
+	mov	qword ptr [rdi],rax 
+	lea	rax,1[rdi]
+	add	rdi,8
+	mov	qword ptr [rdx],rax 
+	sub	rbx,1
+	att_jne	move_closure_with_unboxed_arguments_lp
+
+	pop	rax 
+	att_jmp	move_non_pointers_of_closure
+
+move_closure_with_unboxed_arguments_:
+	mov	qword ptr [rdi],rdx 
+	add	rdi,8
+	sub	rbx,1
+	att_jne	move_closure_with_unboxed_arguments_lp
+
+	pop	rax 
+
+move_non_pointers_of_closure:
+	mov	rbx,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rbx 
+	add	rdi,8
+	sub	rax,1
+	att_jne	move_non_pointers_of_closure
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+move_closure_with_unboxed_arguments_1:
+	mov	rax,qword ptr [rcx]
+	mov	qword ptr [rdi],rax 
+	add	rdi,16
+
+	test	rsi,rsi
+	att_jne	bsf_and_copy_nodes
+	att_jmp	find_non_zero_long
+
+end_move:
+
+	mov	rcx,qword ptr finalizer_list[rip]
+
+restore_finalizer_descriptors:
+	lea	r9,__Nil-8[rip]
+	cmp	rcx,r9
+	je	end_restore_finalizer_descriptors
+
+	lea	r9,e____system__kFinalizer+2[rip]
+	mov	qword ptr [rcx],r9
+	mov	rcx,qword ptr 8[rcx]
+	att_jmp	restore_finalizer_descriptors
+
+end_restore_finalizer_descriptors:
+
diff --git a/macho64/acompact_rmark.s b/macho64/acompact_rmark.s
new file mode 100644
index 0000000..359fe1b
--- /dev/null
+++ b/macho64/acompact_rmark.s
@@ -0,0 +1,953 @@
+
+rmark_stack_nodes1:
+	mov	rbx,qword ptr [rcx]
+	lea	rax,1[rsi]
+	mov	qword ptr [rsi],rbx
+	mov	qword ptr [rcx],rax
+
+rmark_next_stack_node:
+	add	rsi,8
+	cmp	rsi,qword ptr end_vector[rip]
+	je	end_rmark_nodes
+
+rmark_stack_nodes:
+	mov	rcx,qword ptr [rsi]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmark_next_stack_node
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rbx*4]
+	test	rbp,rax 
+	att_jne	rmark_stack_nodes1
+
+	or	rbp,rax
+	mov	dword ptr [rdi+rbx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+	call	rmark_stack_node
+
+	add	rsi,8
+	cmp	rsi,qword ptr end_vector[rip]
+	att_jne	rmark_stack_nodes
+	ret
+
+rmark_stack_node:
+	sub	rsp,16
+	mov	qword ptr [rsi],rax 
+	lea	rbp,1[rsi]
+	mov	qword ptr 8[rsp],rsi 
+	mov	rbx,-1
+	mov	qword ptr [rsp],0
+	mov	qword ptr [rcx],rbp 
+	jmp	rmark_no_reverse
+
+rmark_node_d1:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	jnc	rmark_next_node
+
+	jmp	rmark_node_
+
+rmark_hnf_2:
+	lea	rbx,8[rcx]
+	mov	rax,qword ptr 8[rcx]
+	sub	rsp,16
+
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+
+	mov	qword ptr 8[rsp],rbx 
+	mov	qword ptr [rsp],rax 
+
+	cmp	rsp,qword ptr end_stack[rip]
+	jb	rmark_using_reversal
+
+rmark_node:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmark_next_node
+
+	mov	rbx,rsi 
+
+rmark_node_:
+	mov	rdx,rax 
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rdx*4]
+	test	rbp,rax 
+	jne	rmark_reverse_and_mark_next_node
+	
+	or	rbp,rax 
+	mov	dword ptr [rdi+rdx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+rmark_arguments:
+	cmp	rcx,rbx 
+	att_ja	rmark_no_reverse
+
+	lea	rbp,1[rsi]
+	mov	qword ptr [rsi],rax 
+	mov	qword ptr [rcx],rbp 
+
+rmark_no_reverse:
+	test	al,2
+	je	rmark_lazy_node
+
+	movzx	rbp,word ptr (-2)[rax]
+	test	rbp,rbp 
+	je	rmark_hnf_0
+
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	rmark_record
+
+	sub	rbp,2
+	att_je	rmark_hnf_2
+	jc	rmark_hnf_1
+
+rmark_hnf_3:
+	mov	rdx,qword ptr 8[rcx]
+rmark_hnf_3_:
+	cmp	rsp,qword ptr end_stack[rip]
+	jb	rmark_using_reversal_
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,[rdi+rbx*4]
+	jne	rmark_shared_argument_part
+
+	or	dword ptr [rdi+rbx*4],eax 
+
+rmark_no_shared_argument_part:
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	lea	rsi,8[rcx]
+	mov	rcx,qword ptr [rcx]
+	lea	rdx,[rdx+rbp*8]
+	mov	qword ptr [rsp],rcx 
+
+rmark_push_hnf_args:
+	mov	rbx,qword ptr [rdx]
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rdx 
+	sub	rdx,8
+	mov	qword ptr [rsp],rbx 
+
+	sub	rbp,1
+	att_jg	rmark_push_hnf_args
+
+	mov	rcx,qword ptr [rdx]
+
+	cmp	rdx,rsi 
+	ja	rmark_no_reverse_argument_pointer
+
+	lea	rbp,3[rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr [rdx],rbp 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmark_next_node
+
+	mov	rbx,rdx 
+	att_jmp	rmark_node_
+
+rmark_no_reverse_argument_pointer:
+	mov	rsi,rdx 
+	att_jmp	rmark_node
+
+rmark_shared_argument_part:
+	cmp	rdx,rcx 
+	att_ja	rmark_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr [rdx],rax 
+	mov	qword ptr 8[rcx],rbx 
+	att_jmp	rmark_hnf_1
+
+rmark_record:
+	sub	rbp,258
+	je	rmark_record_2
+	jb	rmark_record_1
+
+rmark_record_3:
+	movzx	rbp,word ptr (-2+2)[rax]
+	mov	rdx,qword ptr (16-8)[rcx]
+	sub	rbp,1
+	jb	rmark_record_3_bb
+	je	rmark_record_3_ab
+	sub	rbp,1
+	je	rmark_record_3_aab
+	att_jmp	rmark_hnf_3_
+
+rmark_record_3_bb:
+	sub	rcx,8
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+	
+	cmp	rdx,rcx 
+	att_ja	rmark_next_node
+
+	add	eax,eax 
+	jne	rmark_bit_in_same_word1
+	inc	rbp
+	mov	rax,1
+rmark_bit_in_same_word1:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmark_not_yet_linked_bb
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	add	rax,16
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmark_next_node
+
+rmark_not_yet_linked_bb:
+	or	dword ptr [rdi+rbp*4],eax 
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmark_next_node
+
+rmark_record_3_ab:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	cmp	rdx,rcx 
+	att_ja	rmark_hnf_1
+
+	add	eax,eax 
+	jne	rmark_bit_in_same_word2
+	inc	rbp
+	mov	rax,1
+rmark_bit_in_same_word2:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmark_not_yet_linked_ab
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	add	rax,8
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmark_hnf_1
+
+rmark_not_yet_linked_ab:
+	or	dword ptr [rdi+rbp*4],eax 
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmark_hnf_1
+
+rmark_record_3_aab:
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_using_reversal_
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbp*4]
+	att_jne	rmark_shared_argument_part
+	or	dword ptr [rdi+rbp*4],eax 
+
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	lea	rsi,8[rcx]
+	mov	rcx,qword ptr [rcx]
+	mov	qword ptr [rsp],rcx 
+
+	mov	rcx,qword ptr [rdx]
+
+	cmp	rdx,rsi 
+	att_ja	rmark_no_reverse_argument_pointer
+
+	lea	rbp,3[rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr [rdx],rbp 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmark_next_node
+
+	mov	rbx,rdx 
+	att_jmp	rmark_node_
+
+rmark_record_2:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	rmark_hnf_2
+	att_je	rmark_hnf_1
+	att_jmp	rmark_next_node
+
+rmark_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	rmark_hnf_1
+	att_jmp	rmark_next_node
+
+rmark_lazy_node_1:
+/* selectors: */
+	jne	rmark_selector_node_1
+
+rmark_hnf_1:
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+	att_jmp	rmark_node
+
+/* selectors */
+rmark_indirection_node:
+	mov	rdx,qword ptr neg_heap_p3[rip]
+	sub	rcx,8
+	add	rdx,rcx 
+
+	mov	rbp,rdx 
+	and	rbp,31*8
+	shr	rdx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	and	dword ptr [rdi+rdx*4],ebp 
+
+	mov	rdx,rcx
+	cmp	rcx,rbx 
+	mov	rcx,qword ptr 8[rcx]
+	mov	qword ptr [rsi],rcx 
+	att_ja	rmark_node_d1
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmark_node_d1
+
+rmark_selector_node_1:
+	add	rbp,3
+	att_je	rmark_indirection_node
+
+	mov	rdx,qword ptr [rcx]
+	mov	qword ptr pointer_compare_address[rip],rbx
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rdx 
+	shr	rbx,3
+
+	add	rbp,1
+	jle	rmark_record_selector_node_1
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmark_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmark_hnf_1
+
+	cmp	word ptr (-2)[rbx],2
+	jbe	rmark_small_tuple_or_record
+
+rmark_large_tuple_or_record:
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	shr	rbx,3
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmark_hnf_1
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	movsxd	r11,dword ptr (-8)[rax]
+	add	rax,r11
+
+	mov	r11,rbx 
+	and	r11,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	r11d,dword ptr [r9+r11]
+	and	dword ptr [rdi+rbx*4],r11d
+
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbx,qword ptr pointer_compare_address[rip]
+
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rcx],r9
+
+	cmp	rax,16
+	jl	rmark_tuple_or_record_selector_node_2
+
+	mov	rdx,rcx
+	je	rmark_tuple_selector_node_2
+
+	mov	rcx,qword ptr (-24)[r10+rax]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmark_node_d1
+
+rmark_tuple_selector_node_2:
+	mov	rcx,qword ptr [r10]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmark_node_d1
+
+rmark_record_selector_node_1:
+	je	rmark_strict_record_selector_node_1
+
+	mov	rbp,rbx
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp
+	att_jne	rmark_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmark_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	att_jbe	rmark_small_tuple_or_record
+
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	shr	rbx,3
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmark_hnf_1
+
+rmark_small_tuple_or_record:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	movsxd	r11,dword ptr(-8)[rax]
+	add	rax,r11
+
+	mov	r11,rbx
+	and	r11,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	r11d,dword ptr [r9+r11]
+	and	dword ptr [rdi+rbx*4],r11d 
+
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbx,qword ptr pointer_compare_address[rip]
+
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rcx],r9
+
+	cmp	rax,16
+	att_jle	rmark_tuple_or_record_selector_node_2
+	mov	rdx,r10
+	sub	rax,24
+rmark_tuple_or_record_selector_node_2:
+	mov	rbp,rcx
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rbp],rcx
+	mov	rdx,rbp
+	att_jmp	rmark_node_d1
+
+rmark_strict_record_selector_node_1:
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmark_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmark_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	jbe	rmark_select_from_small_record
+
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	mov	rbp,rbx 
+
+	shr	rbx,8
+	and	rbp,31*8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp
+	att_jne	rmark_hnf_1
+
+rmark_select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rbx,rax
+	sub	rcx,8
+
+	cmp	rcx,qword ptr pointer_compare_address[rip]
+	ja	rmark_selector_pointer_not_reversed
+
+	movzx	eax,word ptr (4-8)[rbx]
+	cmp	rax,16
+	jle	rmark_strict_record_selector_node_2
+	mov	rax,qword ptr (-24)[r10+rax]
+	jmp	rmark_strict_record_selector_node_3
+rmark_strict_record_selector_node_2:
+	mov	rax,qword ptr [rdx+rax]
+rmark_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rax
+
+	movzx	eax,word ptr (6-8)[rbx]
+	test	rax,rax
+	je	rmark_strict_record_selector_node_5
+	cmp	rax,16
+	jle	rmark_strict_record_selector_node_4
+	mov	rdx,r10
+	sub	rax,24
+rmark_strict_record_selector_node_4:
+	mov	rax,qword ptr [rdx+rax]
+	mov	qword ptr 16[rcx],rax
+rmark_strict_record_selector_node_5:
+
+	mov	rax,qword ptr ((-8)-8)[rbx]
+
+	add	rsi,1
+	mov	qword ptr [rcx],rsi 
+	mov	qword ptr (-1)[rsi],rax 
+	att_jmp	rmark_next_node
+
+rmark_selector_pointer_not_reversed:
+	movzx	eax,word ptr (4-8)[rbx]
+	cmp	rax,16
+	jle	rmark_strict_record_selector_node_6
+	mov	rax,qword ptr (-24)[r10+rax]
+	jmp	rmark_strict_record_selector_node_7
+rmark_strict_record_selector_node_6:
+	mov	rax,qword ptr [rdx+rax]
+rmark_strict_record_selector_node_7:
+	mov	qword ptr 8[rcx],rax
+
+	movzx	eax,word ptr (6-8)[rbx]
+	test	rax,rax
+	je	rmark_strict_record_selector_node_9
+	cmp	rax,16
+	jle	rmark_strict_record_selector_node_8
+	mov	rdx,r10
+	sub	rax,24
+rmark_strict_record_selector_node_8:
+	mov	rax,qword ptr [rdx+rax]
+	mov	qword ptr 16[rcx],rax
+rmark_strict_record_selector_node_9:
+
+	mov	rax,qword ptr ((-8)-8)[rbx]
+	mov	qword ptr [rcx],rax
+	att_jmp	rmark_next_node
+
+rmark_reverse_and_mark_next_node:
+	cmp	rcx,rbx 
+	att_ja	rmark_next_node
+
+	mov	rax,qword ptr [rcx]
+	mov	qword ptr [rsi],rax 
+	add	rsi,1
+	mov	qword ptr [rcx],rsi 
+
+/* %rbp ,%rbx : free */
+
+rmark_next_node:
+	mov	rcx,qword ptr [rsp]
+	mov	rsi,qword ptr 8[rsp]
+	add	rsp,16
+
+	cmp	rcx,1
+	att_ja	rmark_node
+
+rmark_next_node_:
+end_rmark_nodes:
+	ret
+
+rmark_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp
+	att_je	rmark_next_node
+
+	add	rcx,8
+
+	sub	rbp,1
+	att_jle	rmark_lazy_node_1
+
+	cmp	rbp,255
+	jge	rmark_closure_with_unboxed_arguments
+
+rmark_closure_with_unboxed_arguments_:
+	lea	rcx,[rcx+rbp*8]
+
+rmark_push_lazy_args:
+	mov	rbx,qword ptr [rcx]
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	sub	rcx,8
+	mov	qword ptr [rsp],rbx 
+	sub	rbp,1
+	att_jg	rmark_push_lazy_args
+
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jae	rmark_node
+
+	att_jmp	rmark_using_reversal
+
+rmark_closure_with_unboxed_arguments:
+/* (a_size+b_size)+(b_size<<8) */
+/*	addl	$1,%rbp  */
+	mov	rax,rbp 
+	and	rbp,255
+	shr	rax,8
+	sub	rbp,rax 
+/*	subl	$1,%rbp  */
+	att_jg	rmark_closure_with_unboxed_arguments_
+	att_je	rmark_hnf_1
+	att_jmp	rmark_next_node
+
+rmark_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	je	rmark_int_3
+
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+ 	je	rmark_char_3
+
+	jb	rmark_no_normal_hnf_0
+
+	mov	rbp,qword ptr neg_heap_p3[rip]
+	add	rbp,rcx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx 
+
+	lea	rdx,((-8)-2)[rax]
+	mov	qword ptr [rsi],rdx 
+	cmp	rcx,rbx 
+	att_ja	rmark_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmark_next_node
+
+rmark_int_3:
+	mov	rbp,qword ptr 8[rcx]
+	cmp	rbp,33
+	att_jnc	rmark_next_node
+
+	shl	rbp,4
+	lea	rdx,small_integers[rip]
+	add	rdx,rbp
+	mov	rbp,qword ptr neg_heap_p3[rip]
+	mov	qword ptr [rsi],rdx 
+	add	rbp,rcx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx 
+
+	cmp	rcx,rbx
+	att_ja	rmark_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmark_next_node
+
+rmark_char_3:
+	movzx	rdx,byte ptr 8[rcx]
+	mov	rbp,qword ptr neg_heap_p3[rip]
+
+	shl	rdx,4
+	add	rbp,rcx 
+	lea	r9,static_characters[rip]
+	add	rdx,r9
+	mov	qword ptr [rsi],rdx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx
+
+	cmp	rcx,rbx
+	att_ja	rmark_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmark_next_node
+
+rmark_no_normal_hnf_0:
+	lea	r9,__ARRAY__+2[rip]
+	cmp	rax,r9
+	att_jne	rmark_next_node
+
+	mov	rax,qword ptr 16[rcx]
+	test	rax,rax 
+	je	rmark_lazy_array
+
+	movzx	rdx,word ptr (-2+2)[rax]
+	test	rdx,rdx
+	je	rmark_b_array
+
+	movzx	rax,word ptr (-2)[rax]
+	test	rax,rax 
+	att_je	rmark_b_array
+
+	cmp	rsp,qword ptr end_stack[rip]
+	jb	rmark_array_using_reversal
+
+	sub	rax,256
+	cmp	rdx,rax 
+	mov	rbx,rdx
+	je	rmark_a_record_array
+
+rmark_ab_record_array:
+	mov	rdx,qword ptr 8[rcx]
+	add	rcx,16
+	push	rcx
+
+	imul	rdx,rax 
+	shl	rdx,3
+
+	sub	rax,rbx 
+	add	rcx,8
+	add	rdx,rcx 
+	att_call	reorder
+	
+	pop	rcx 
+	mov	rax,rbx
+	imul	rax,qword ptr (-8)[rcx]
+	jmp	rmark_lr_array
+
+rmark_b_array:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	add	rax,8
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	att_jmp	rmark_next_node
+
+rmark_a_record_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+	cmp	rbx,2
+	att_jb	rmark_lr_array
+
+	imul	rax,rbx 
+	att_jmp	rmark_lr_array
+
+rmark_lazy_array:
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_array_using_reversal
+
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+
+rmark_lr_array:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx 
+	shr	rbx,3
+	add	rbx,rax 
+
+	mov	rdx,rbx 
+	and	rbx,31
+	shr	rdx,5
+	lea	r9,bit_set_table[rip]
+	mov	ebx,dword ptr [r9+rbx*4]
+	or	dword ptr [rdi+rdx*4],ebx 
+
+	cmp	rax,1
+	jbe	rmark_array_length_0_1
+
+	mov	rdx,rcx 
+	lea	rcx,[rcx+rax*8]
+
+	mov	rax,qword ptr [rcx]
+
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+
+	mov	qword ptr [rcx],rbx 
+
+	mov	rax,qword ptr (-8)[rcx]
+	sub	rcx,8
+
+	mov	rbx,qword ptr (-8)[rdx]
+
+	sub	rdx,8
+	mov	qword ptr [rcx],rbx 
+
+	mov	qword ptr [rdx],rax 
+
+	push	rcx
+	mov	rsi,rdx 
+	jmp	rmark_array_nodes
+
+rmark_array_nodes1:
+	cmp	rcx,rsi 
+	ja	rmark_next_array_node
+
+	mov	rbx,qword ptr [rcx]
+	lea	rax,1[rsi]
+	mov	qword ptr [rsi],rbx 
+	mov	qword ptr [rcx],rax 
+
+rmark_next_array_node:
+	add	rsi,8
+	cmp	rsi,qword ptr [rsp]
+	je	end_rmark_array_node
+
+rmark_array_nodes:
+	mov	rcx,qword ptr [rsi]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmark_next_array_node
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rbx*4]
+	test	rbp,rax 
+	att_jne	rmark_array_nodes1
+	
+	or	rbp,rax 
+	mov	dword ptr [rdi+rbx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+	call	rmark_array_node
+
+	add	rsi,8
+	cmp	rsi,qword ptr [rsp]
+	att_jne	rmark_array_nodes
+
+end_rmark_array_node:
+	add	rsp,8
+	att_jmp	rmark_next_node
+
+rmark_array_node:
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rsi 
+	mov	rbx,rsi
+	mov	qword ptr [rsp],1
+	att_jmp	rmark_arguments
+
+rmark_array_length_0_1:
+	lea	rcx,-16[rcx]
+	att_jb	rmark_next_node
+
+	mov	rbx,qword ptr 24[rcx]
+	mov	rbp,qword ptr 16[rcx]
+	mov	qword ptr 24[rcx],rbp
+	mov	rbp,qword ptr 8[rcx] 
+	mov	qword ptr 16[rcx],rbp
+	mov	qword ptr 8[rcx],rbx
+	add	rcx,8
+	att_jmp	rmark_hnf_1
+
+
+	.data
+pointer_compare_address:
+	.quad	0
+
+	.text
diff --git a/macho64/acompact_rmark_prefetch.s b/macho64/acompact_rmark_prefetch.s
new file mode 100644
index 0000000..4ca3ec5
--- /dev/null
+++ b/macho64/acompact_rmark_prefetch.s
@@ -0,0 +1,1181 @@
+
+
+	.data
+rmarkp_n_queue_items_16:
+	.quad	0
+rmarkp_queue_first:
+	.quad	0
+rmarkp_queue:
+	.quad	0,0,0,0,0,0,0,0
+	.quad	0,0,0,0,0,0,0,0
+	.quad	0,0,0,0,0,0,0,0
+	.quad	0,0,0,0,0,0,0,0
+
+	.text
+
+rmarkp_stack_nodes1:
+	mov	rbx,qword ptr [rcx]
+	lea	rax,1[rsi]
+	mov	qword ptr [rsi],rbx
+	mov	qword ptr [rcx],rax
+
+rmarkp_next_stack_node:
+	add	rsi,8
+	cmp	rsi,qword ptr end_vector[rip]
+	je	end_rmarkp_nodes
+
+rmarkp_stack_nodes:
+	mov	rcx,qword ptr [rsi]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_stack_node
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rbx*4]
+	test	rbp,rax 
+	att_jne	rmarkp_stack_nodes1
+
+	or	rbp,rax
+	mov	dword ptr [rdi+rbx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+	call	rmarkp_stack_node
+
+	add	rsi,8
+	cmp	rsi,qword ptr end_vector[rip]
+	att_jne	rmarkp_stack_nodes
+	ret
+
+rmarkp_stack_node:
+	sub	rsp,16
+	mov	qword ptr [rsi],rax 
+	lea	rbp,1[rsi]
+	mov	qword ptr 8[rsp],rsi 
+	mov	rbx,-1
+	mov	qword ptr [rsp],0
+	mov	qword ptr [rcx],rbp 
+	jmp	rmarkp_no_reverse
+
+rmarkp_node_d1:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	jnc	rmarkp_next_node
+
+	jmp	rmarkp_node_
+
+rmarkp_hnf_2:
+	lea	rbx,8[rcx]
+	mov	rax,qword ptr 8[rcx]
+	sub	rsp,16
+
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+
+	mov	qword ptr 8[rsp],rbx 
+	mov	qword ptr [rsp],rax 
+
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_using_reversal
+
+rmarkp_node:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_node
+
+	mov	rbx,rsi 
+
+rmarkp_node_:
+
+
+
+	mov	rdx,rax 
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rdx*4]
+	jne	rmarkp_reverse_and_mark_next_node
+
+	mov	rbp,qword ptr rmarkp_queue_first[rip]
+	mov	rdx,qword ptr rmarkp_n_queue_items_16[rip]
+
+	prefetch	[rcx]
+	lea	r9,rmarkp_queue[rip]
+	mov	qword ptr [r9+rbp],rcx
+	mov	qword ptr 8[r9+rbp],rsi
+	mov	qword ptr 16[r9+rbp],rbx
+	lea	rbx,[rbp+rdx]
+	add	rbp,32
+
+	and	rbp,7*32
+	and	rbx,7*32
+
+	mov	qword ptr rmarkp_queue_first[rip],rbp
+
+	cmp	rdx,-4*32
+	je	rmarkp_last_item_in_queue
+
+rmarkp_add_items:
+	mov	rcx,[rsp]
+	cmp	rcx,1
+	jbe	rmarkp_add_stacked_item
+
+	mov	rsi,8[rsp]
+	add	rsp,16
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_add_items
+
+	mov	rdx,rax
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rdx*4]
+	test	rbp,rax
+	je	rmarkp_add_item
+
+	cmp	rcx,rsi
+	att_ja	rmarkp_add_items
+
+	mov	rax,[rcx]
+	mov	[rsi],rax
+	add	rsi,1
+	mov	[rcx],rsi
+	att_jmp	rmarkp_add_items
+
+rmarkp_add_stacked_item:
+	att_je	rmarkp_last_item_in_queue
+rmarkp_add_items2:
+	mov	rsi,8[rsp]
+	add	rsi,8
+	cmp	rsi,qword ptr end_vector[rip]
+	att_je	rmarkp_last_item_in_queue
+
+	mov	rcx,[rsi]
+	mov	8[rsp],rsi
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_add_items2
+
+	mov	rdx,rax
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rdx*4]
+	test	rbp,rax
+	je	rmarkp_add_item2
+
+	mov	rax,[rcx]
+	mov	[rsi],rax
+	add	rsi,1
+	mov	[rcx],rsi
+	att_jmp	rmarkp_add_items2
+
+rmarkp_add_item2:
+	prefetch	[rcx]
+	mov	rbp,qword ptr rmarkp_queue_first[rip]
+	mov	rdx,qword ptr rmarkp_n_queue_items_16[rip]
+
+	lea	r9,rmarkp_queue[rip]
+	mov	qword ptr [r9+rbp],rcx
+	mov	qword ptr 8[r9+rbp],rsi
+	mov	qword ptr 16[r9+rbp],-1
+	add	rbp,32
+	and	rbp,7*32
+
+	sub	rdx,32
+
+	mov	qword ptr rmarkp_queue_first[rip],rbp
+	mov	qword ptr rmarkp_n_queue_items_16[rip],rdx
+
+	cmp	rdx,-4*32
+	att_jne	rmarkp_add_items2
+	att_jmp	rmarkp_last_item_in_queue
+
+rmarkp_add_items3:
+	mov	rsi,8[rsp]
+	add	rsi,8
+	cmp	rsi,24[rsp]
+	att_je	rmarkp_last_item_in_queue
+
+	mov	rcx,[rsi]
+	mov	8[rsp],rsi
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_add_items3
+
+	mov	rdx,rax
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,[rdi+rdx*4]
+	test	rbp,rax
+	je	rmarkp_add_item3
+
+	cmp	rcx,rsi
+	att_ja	rmarkp_add_items3
+
+	mov	rax,[rcx]
+	mov	[rsi],rax
+	add	rsi,1
+	mov	[rcx],rsi
+	att_jmp	rmarkp_add_items3
+
+rmarkp_add_item3:
+	prefetch	[rcx]
+	mov	rbp,qword ptr rmarkp_queue_first[rip]
+	mov	rdx,qword ptr rmarkp_n_queue_items_16[rip]
+
+	lea	r9,rmarkp_queue[rip]
+	mov	qword ptr 8[r9+rbp],rsi
+	mov	qword ptr 16[r9+rbp],rsi
+	add	rbp,32
+	and	rbp,7*32
+
+	sub	rdx,32
+
+	mov	qword ptr rmarkp_queue_first[rip],rbp
+	mov	qword ptr rmarkp_n_queue_items_16[rip],rdx
+
+	cmp	rdx,-4*32
+	att_jne	rmarkp_add_items3
+	att_jmp	rmarkp_last_item_in_queue
+
+rmarkp_add_item:
+	prefetch	[rcx]
+	mov	rbp,qword ptr rmarkp_queue_first[rip]
+	mov	rdx,qword ptr rmarkp_n_queue_items_16[rip]
+
+	lea	r9,rmarkp_queue[rip]
+	mov	qword ptr [r9+rbp],rcx
+	mov	qword ptr 8[r9+rbp],rsi
+	mov	qword ptr 16[r9+rbp],rsi
+	add	rbp,32
+	and	rbp,7*32
+
+	sub	rdx,32
+
+	mov	qword ptr rmarkp_queue_first[rip],rbp
+	mov	qword ptr rmarkp_n_queue_items_16[rip],rdx
+
+	cmp	rdx,-4*32
+	att_jne	rmarkp_add_items
+
+rmarkp_last_item_in_queue:
+	lea	r9,rmarkp_queue[rip]
+	mov	rcx,qword ptr [r9+rbx]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+
+	mov	rsi,qword ptr 8[r9+rbx]
+	mov	rbx,qword ptr 16[r9+rbx]
+
+	add	rax,rcx
+
+rmarkp_node_no_prefetch:
+
+
+
+	mov	rdx,rax 
+	and	rax,31*8
+	shr	rdx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rdx*4]
+	test	rbp,rax 
+	att_jne	rmarkp_reverse_and_mark_next_node
+	
+	or	rbp,rax 
+	mov	dword ptr [rdi+rdx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+rmarkp_arguments:
+	cmp	rcx,rbx 
+	att_ja	rmarkp_no_reverse
+
+	lea	rbp,1[rsi]
+	mov	qword ptr [rsi],rax 
+	mov	qword ptr [rcx],rbp 
+
+rmarkp_no_reverse:
+	test	al,2
+	je	rmarkp_lazy_node
+
+	movzx	rbp,word ptr (-2)[rax]
+	test	rbp,rbp 
+	je	rmarkp_hnf_0
+
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	rmarkp_record
+
+	sub	rbp,2
+	att_je	rmarkp_hnf_2
+	jc	rmarkp_hnf_1
+
+rmarkp_hnf_3:
+	mov	rdx,qword ptr 8[rcx]
+rmarkp_hnf_3_:
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_using_reversal_
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,[rdi+rbx*4]
+	jne	rmarkp_shared_argument_part
+
+	or	dword ptr [rdi+rbx*4],eax 
+
+rmarkp_no_shared_argument_part:
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	lea	rsi,8[rcx]
+	mov	rcx,qword ptr [rcx]
+	lea	rdx,[rdx+rbp*8]
+	mov	qword ptr [rsp],rcx 
+
+rmarkp_push_hnf_args:
+	mov	rbx,qword ptr [rdx]
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rdx 
+	sub	rdx,8
+	mov	qword ptr [rsp],rbx 
+
+	sub	rbp,1
+	att_jg	rmarkp_push_hnf_args
+
+	mov	rcx,qword ptr [rdx]
+
+	cmp	rdx,rsi 
+	ja	rmarkp_no_reverse_argument_pointer
+
+	lea	rbp,3[rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr [rdx],rbp 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_node
+
+	mov	rbx,rdx 
+	att_jmp	rmarkp_node_
+
+rmarkp_no_reverse_argument_pointer:
+	mov	rsi,rdx 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_node
+	mov	rbx,rsi
+	att_jmp	rmarkp_node_no_prefetch
+
+rmarkp_shared_argument_part:
+	cmp	rdx,rcx 
+	att_ja	rmarkp_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr [rdx],rax 
+	mov	qword ptr 8[rcx],rbx 
+	att_jmp	rmarkp_hnf_1
+
+rmarkp_record:
+	sub	rbp,258
+	je	rmarkp_record_2
+	jb	rmarkp_record_1
+
+rmarkp_record_3:
+	movzx	rbp,word ptr (-2+2)[rax]
+	mov	rdx,qword ptr (16-8)[rcx]
+	sub	rbp,1
+	jb	rmarkp_record_3_bb
+	je	rmarkp_record_3_ab
+	sub	rbp,1
+	je	rmarkp_record_3_aab
+	att_jmp	rmarkp_hnf_3_
+
+rmarkp_record_3_bb:
+	sub	rcx,8
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+	
+	cmp	rdx,rcx 
+	att_ja	rmarkp_next_node
+
+	add	eax,eax 
+	jne	rmarkp_bit_in_same_word1
+	inc	rbp
+	mov	rax,1
+rmarkp_bit_in_same_word1:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmarkp_not_yet_linked_bb
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	add	rax,16
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_not_yet_linked_bb:
+	or	dword ptr [rdi+rbp*4],eax 
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_record_3_ab:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	cmp	rdx,rcx 
+	att_ja	rmarkp_hnf_1
+
+	add	eax,eax 
+	jne	rmarkp_bit_in_same_word2
+	inc	rbp
+	mov	rax,1
+rmarkp_bit_in_same_word2:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmarkp_not_yet_linked_ab
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	add	rax,8
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax 
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkp_hnf_1
+
+rmarkp_not_yet_linked_ab:
+	or	dword ptr [rdi+rbp*4],eax 
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkp_hnf_1
+
+rmarkp_record_3_aab:
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_using_reversal_
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbp*4]
+	att_jne	rmarkp_shared_argument_part
+	or	dword ptr [rdi+rbp*4],eax 
+
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	lea	rsi,8[rcx]
+	mov	rcx,qword ptr [rcx]
+	mov	qword ptr [rsp],rcx 
+
+	mov	rcx,qword ptr [rdx]
+
+	cmp	rdx,rsi 
+	att_ja	rmarkp_no_reverse_argument_pointer
+
+	lea	rbp,3[rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr [rdx],rbp 
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_node
+
+	mov	rbx,rdx 
+	att_jmp	rmarkp_node_
+
+rmarkp_record_2:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	rmarkp_hnf_2
+	att_je	rmarkp_hnf_1
+	att_jmp	rmarkp_next_node
+
+rmarkp_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	rmarkp_hnf_1
+	att_jmp	rmarkp_next_node
+
+rmarkp_lazy_node_1:
+/* selectors: */
+	jne	rmarkp_selector_node_1
+
+rmarkp_hnf_1:
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+	att_jmp	rmarkp_node
+
+/* selectors */
+rmarkp_indirection_node:
+	mov	rdx,qword ptr neg_heap_p3[rip]
+	sub	rcx,8
+	add	rdx,rcx 
+
+	mov	rbp,rdx 
+	and	rbp,31*8
+	shr	rdx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	and	dword ptr [rdi+rdx*4],ebp 
+
+	mov	rdx,rcx
+	cmp	rcx,rbx 
+	mov	rcx,qword ptr 8[rcx]
+	mov	qword ptr [rsi],rcx 
+	att_ja	rmarkp_node_d1
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkp_node_d1
+
+rmarkp_selector_node_1:
+	add	rbp,3
+	att_je	rmarkp_indirection_node
+
+	mov	rdx,qword ptr [rcx]
+	mov	qword ptr pointer_compare_address[rip],rbx
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rdx 
+	shr	rbx,3
+
+	add	rbp,1
+	jle	rmarkp_record_selector_node_1
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmarkp_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkp_hnf_1
+
+	cmp	word ptr (-2)[rbx],2
+	jbe	rmarkp_small_tuple_or_record
+
+rmarkp_large_tuple_or_record:
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	shr	rbx,3
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmarkp_hnf_1
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	movsxd	r11,dword ptr(-8)[rax]
+	add	rax,r11
+
+	mov	r11,rbx 
+	and	r11,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	r11d,dword ptr [r9+r11]
+	and	dword ptr [rdi+rbx*4],r11d
+
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbx,qword ptr pointer_compare_address[rip]
+
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rcx],r9
+
+	cmp	rax,16
+	jl	rmarkp_tuple_or_record_selector_node_2
+
+	mov	rdx,rcx
+	je	rmarkp_tuple_selector_node_2
+
+	mov	rcx,qword ptr (-24)[r10+rax]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmarkp_node_d1
+
+rmarkp_tuple_selector_node_2:
+	mov	rcx,qword ptr [r10]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmarkp_node_d1
+
+rmarkp_record_selector_node_1:
+	je	rmarkp_strict_record_selector_node_1
+
+	mov	rbp,rbx
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp
+	att_jne	rmarkp_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkp_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	att_jbe	rmarkp_small_tuple_or_record
+
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	shr	rbx,3
+
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmarkp_hnf_1
+
+rmarkp_small_tuple_or_record:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	movsxd	r11,dword ptr(-8)[rax]
+	add	rax,r11
+
+	mov	r11,rbx
+	and	r11,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	r11d,dword ptr [r9+r11]
+	and	dword ptr [rdi+rbx*4],r11d 
+
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbx,qword ptr pointer_compare_address[rip]
+
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rcx],r9
+
+	cmp	rax,16
+	att_jle	rmarkp_tuple_or_record_selector_node_2
+	mov	rdx,r10
+	sub	rax,24
+rmarkp_tuple_or_record_selector_node_2:
+	mov	rbp,rcx
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rbp],rcx
+	mov	rdx,rbp
+	att_jmp	rmarkp_node_d1
+
+rmarkp_strict_record_selector_node_1:
+	mov	rbp,rbx 
+	shr	rbx,5
+	and	rbp,31
+	lea	r9,bit_set_table[rip]
+	mov	ebp,dword ptr [r9+rbp*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp 
+	att_jne	rmarkp_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkp_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	jbe	rmarkp_select_from_small_record
+
+	mov	r10,qword ptr 16[rdx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,r10
+	mov	rbp,rbx 
+
+	shr	rbx,8
+	and	rbp,31*8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rbp
+	att_jne	rmarkp_hnf_1
+
+rmarkp_select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rbx,rax
+	sub	rcx,8
+
+	cmp	rcx,qword ptr pointer_compare_address[rip]
+	ja	rmarkp_selector_pointer_not_reversed
+
+	movzx	eax,word ptr (4-8)[rbx]
+	cmp	rax,16
+	jle	rmarkp_strict_record_selector_node_2
+	mov	rax,qword ptr (-24)[r10+rax]
+	jmp	rmarkp_strict_record_selector_node_3
+rmarkp_strict_record_selector_node_2:
+	mov	rax,qword ptr [rdx+rax]
+rmarkp_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rax
+
+	movzx	eax,word ptr (6-8)[rbx]
+	test	rax,rax
+	je	rmarkp_strict_record_selector_node_5
+	cmp	rax,16
+	jle	rmarkp_strict_record_selector_node_4
+	mov	rdx,r10
+	sub	rax,24
+rmarkp_strict_record_selector_node_4:
+	mov	rax,qword ptr [rdx+rax]
+	mov	qword ptr 16[rcx],rax
+rmarkp_strict_record_selector_node_5:
+
+	mov	rax,qword ptr ((-8)-8)[rbx]
+	add	rsi,1
+	mov	qword ptr [rcx],rsi 
+	mov	qword ptr (-1)[rsi],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_selector_pointer_not_reversed:
+	movzx	eax,word ptr (4-8)[rbx]
+	cmp	rax,16
+	jle	rmarkp_strict_record_selector_node_6
+	mov	rax,qword ptr (-24)[r10+rax]
+	jmp	rmarkp_strict_record_selector_node_7
+rmarkp_strict_record_selector_node_6:
+	mov	rax,qword ptr [rdx+rax]
+rmarkp_strict_record_selector_node_7:
+	mov	qword ptr 8[rcx],rax
+
+	movzx	eax,word ptr (6-8)[rbx]
+	test	rax,rax
+	je	rmarkp_strict_record_selector_node_9
+	cmp	rax,16
+	jle	rmarkp_strict_record_selector_node_8
+	mov	rdx,r10
+	sub	rax,24
+rmarkp_strict_record_selector_node_8:
+	mov	rax,qword ptr [rdx+rax]
+	mov	qword ptr 16[rcx],rax
+rmarkp_strict_record_selector_node_9:
+
+	mov	rax,qword ptr ((-8)-8)[rbx]
+	mov	qword ptr [rcx],rax
+	att_jmp	rmarkp_next_node
+
+rmarkp_reverse_and_mark_next_node:
+	cmp	rcx,rbx 
+	att_ja	rmarkp_next_node
+
+	mov	rax,qword ptr [rcx]
+	mov	qword ptr [rsi],rax 
+	add	rsi,1
+	mov	qword ptr [rcx],rsi 
+
+/* %rbp ,%rbx : free */
+
+rmarkp_next_node:
+	mov	rcx,qword ptr [rsp]
+	mov	rsi,qword ptr 8[rsp]
+	add	rsp,16
+
+	cmp	rcx,1
+	att_ja	rmarkp_node
+
+rmarkp_next_node_:
+	mov	rdx,qword ptr rmarkp_n_queue_items_16[rip]
+	test	rdx,rdx
+	att_je	end_rmarkp_nodes
+
+	sub	rsp,16
+
+	mov	rbp,qword ptr rmarkp_queue_first[rip]
+
+	lea	rbx,[rbp+rdx]
+	add	rdx,32
+
+	and	rbx,7*32
+
+	mov	qword ptr rmarkp_n_queue_items_16[rip],rdx
+	att_jmp	rmarkp_last_item_in_queue
+
+end_rmarkp_nodes:
+	ret
+
+rmarkp_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp
+	att_je	rmarkp_next_node
+
+	add	rcx,8
+
+	sub	rbp,1
+	att_jle	rmarkp_lazy_node_1
+
+	cmp	rbp,255
+	jge	rmarkp_closure_with_unboxed_arguments
+
+rmarkp_closure_with_unboxed_arguments_:
+	lea	rcx,[rcx+rbp*8]
+
+rmarkp_push_lazy_args:
+	mov	rbx,qword ptr [rcx]
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rcx 
+	sub	rcx,8
+	mov	qword ptr [rsp],rbx 
+	sub	rbp,1
+	att_jg	rmarkp_push_lazy_args
+
+	mov	rsi,rcx 
+	mov	rcx,qword ptr [rcx]
+
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jae	rmarkp_node
+
+	att_jmp	rmark_using_reversal
+
+rmarkp_closure_with_unboxed_arguments:
+/* (a_size+b_size)+(b_size<<8) */
+/*	addl	$1,%rbp  */
+	mov	rax,rbp 
+	and	rbp,255
+	shr	rax,8
+	sub	rbp,rax 
+/*	subl	$1,%rbp  */
+	att_jg	rmarkp_closure_with_unboxed_arguments_
+	att_je	rmarkp_hnf_1
+	att_jmp	rmarkp_next_node
+
+rmarkp_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	je	rmarkp_int_3
+
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+ 	je	rmarkp_char_3
+
+	jb	rmarkp_no_normal_hnf_0
+
+	mov	rbp,qword ptr neg_heap_p3[rip]
+	add	rbp,rcx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx 
+
+	lea	rdx,((-8)-2)[rax]
+	mov	qword ptr [rsi],rdx 
+	cmp	rcx,rbx 
+	att_ja	rmarkp_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_int_3:
+	mov	rbp,qword ptr 8[rcx]
+	cmp	rbp,33
+	att_jnc	rmarkp_next_node
+
+	shl	rbp,4
+	lea	rdx,small_integers[rip]
+	add	rdx,rbp
+	mov	rbp,qword ptr neg_heap_p3[rip]
+	mov	qword ptr [rsi],rdx 
+	add	rbp,rcx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx 
+
+	cmp	rcx,rbx
+	att_ja	rmarkp_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_char_3:
+	movzx	rdx,byte ptr 8[rcx]
+	mov	rbp,qword ptr neg_heap_p3[rip]
+
+	shl	rdx,4
+	add	rbp,rcx 
+	lea	r9,static_characters[rip]
+	add	rdx,r9
+	mov	qword ptr [rsi],rdx 
+
+	mov	rdx,rbp 
+	and	rdx,31*8
+	shr	rbp,8
+	lea	r9,bit_clear_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	and	dword ptr [rdi+rbp*4],edx
+
+	cmp	rcx,rbx
+	att_ja	rmarkp_next_node
+	mov	qword ptr [rcx],rax 
+	att_jmp	rmarkp_next_node
+
+rmarkp_no_normal_hnf_0:
+	lea	r9,__ARRAY__+2[rip]
+	cmp	rax,r9
+	att_jne	rmarkp_next_node
+
+	mov	rax,qword ptr 16[rcx]
+	test	rax,rax 
+	je	rmarkp_lazy_array
+
+	movzx	rdx,word ptr (-2+2)[rax]
+	test	rdx,rdx
+	je	rmarkp_b_array
+
+	movzx	rax,word ptr (-2)[rax]
+	test	rax,rax 
+	att_je	rmarkp_b_array
+
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_array_using_reversal
+
+	sub	rax,256
+	cmp	rdx,rax 
+	mov	rbx,rdx
+	je	rmarkp_a_record_array
+
+rmarkp_ab_record_array:
+	mov	rdx,qword ptr 8[rcx]
+	add	rcx,16
+	push	rcx
+
+	imul	rdx,rax 
+	shl	rdx,3
+
+	sub	rax,rbx 
+	add	rcx,8
+	add	rdx,rcx 
+	att_call	reorder
+	
+	pop	rcx 
+	mov	rax,rbx
+	imul	rax,qword ptr (-8)[rcx]
+	jmp	rmarkp_lr_array
+
+rmarkp_b_array:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	add	rax,8
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	att_jmp	rmarkp_next_node
+
+rmarkp_a_record_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+	cmp	rbx,2
+	att_jb	rmarkp_lr_array
+
+	imul	rax,rbx 
+	att_jmp	rmarkp_lr_array
+
+rmarkp_lazy_array:
+	cmp	rsp,qword ptr end_stack[rip]
+	att_jb	rmark_array_using_reversal
+
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+
+rmarkp_lr_array:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx 
+	shr	rbx,3
+	add	rbx,rax 
+
+	mov	rdx,rbx 
+	and	rbx,31
+	shr	rdx,5
+	lea	r9,bit_set_table[rip]
+	mov	ebx,dword ptr [r9+rbx*4]
+	or	dword ptr [rdi+rdx*4],ebx 
+
+	cmp	rax,1
+	jbe	rmarkp_array_length_0_1
+
+	mov	rdx,rcx 
+	lea	rcx,[rcx+rax*8]
+
+	mov	rax,qword ptr [rcx]
+
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax 
+
+	mov	qword ptr [rcx],rbx 
+
+	mov	rax,qword ptr (-8)[rcx]
+	sub	rcx,8
+
+	mov	rbx,qword ptr (-8)[rdx]
+
+	sub	rdx,8
+	mov	qword ptr [rcx],rbx 
+
+	mov	qword ptr [rdx],rax 
+
+	push	rcx
+	mov	rsi,rdx 
+	jmp	rmarkp_array_nodes
+
+rmarkp_array_nodes1:
+	cmp	rcx,rsi 
+	ja	rmarkp_next_array_node
+
+	mov	rbx,qword ptr [rcx]
+	lea	rax,1[rsi]
+	mov	qword ptr [rsi],rbx 
+	mov	qword ptr [rcx],rax 
+
+rmarkp_next_array_node:
+	add	rsi,8
+	cmp	rsi,qword ptr [rsp]
+	je	end_rmarkp_array_node
+
+rmarkp_array_nodes:
+	mov	rcx,qword ptr [rsi]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx 
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	att_jnc	rmarkp_next_array_node
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rbx*4]
+	test	rbp,rax 
+	att_jne	rmarkp_array_nodes1
+	
+	or	rbp,rax 
+	mov	dword ptr [rdi+rbx*4],ebp 
+
+	mov	rax,qword ptr [rcx]
+	call	rmarkp_array_node
+
+	add	rsi,8
+	cmp	rsi,qword ptr [rsp]
+	att_jne	rmarkp_array_nodes
+
+end_rmarkp_array_node:
+	add	rsp,8
+	att_jmp	rmarkp_next_node
+
+rmarkp_array_node:
+	sub	rsp,16
+	mov	qword ptr 8[rsp],rsi 
+	mov	rbx,rsi
+	mov	qword ptr [rsp],1
+	att_jmp	rmarkp_arguments
+
+rmarkp_array_length_0_1:
+	lea	rcx,-16[rcx]
+	att_jb	rmarkp_next_node
+
+	mov	rbx,qword ptr 24[rcx]
+	mov	rbp,qword ptr 16[rcx]
+	mov	qword ptr 24[rcx],rbp
+	mov	rbp,qword ptr 8[rcx] 
+	mov	qword ptr 16[rcx],rbp
+	mov	qword ptr 8[rcx],rbx
+	add	rcx,8
+	att_jmp	rmarkp_hnf_1
diff --git a/macho64/acompact_rmarkr.s b/macho64/acompact_rmarkr.s
new file mode 100644
index 0000000..852978f
--- /dev/null
+++ b/macho64/acompact_rmarkr.s
@@ -0,0 +1,994 @@
+
+rmark_using_reversal:
+	push	rsi
+	push	rsi
+	mov	rsi,1
+	jmp	rmarkr_node
+
+rmark_using_reversal_:
+	sub	rcx,8
+	push	rbx
+	push	rsi
+	cmp	rcx,rbx
+	ja	rmark_no_undo_reverse_1
+	mov	qword ptr [rsi],rcx
+	mov	qword ptr [rcx],rax
+rmark_no_undo_reverse_1:
+	mov	rsi,1
+	jmp	rmarkr_arguments
+
+rmark_array_using_reversal:
+	push	rbx
+	push	rsi
+	cmp	rcx,rbx
+	ja	rmark_no_undo_reverse_2
+	mov	qword ptr [rsi],rcx
+	lea	r9,__ARRAY__+2[rip]
+	mov	qword ptr [rcx],r9
+rmark_no_undo_reverse_2:
+	mov	rsi,1
+	att_jmp	rmarkr_arguments
+
+rmarkr_hnf_2:
+	or	qword ptr [rcx],2
+	mov	rbp,qword ptr 8[rcx]
+	mov	qword ptr 8[rcx],rsi
+	lea	rsi,8[rcx]
+	mov	rcx,rbp
+
+rmarkr_node:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+
+	cmp	rax,qword ptr heap_size_64_65[rip]
+	jnc	rmarkr_next_node_after_static
+
+	mov	rbx,rax
+	and	rax,31*8
+	shr	rbx,8
+
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebp,dword ptr [rdi+rbx*4]
+
+	test	rbp,rax
+	jne	rmarkr_next_node
+
+	or	rbp,rax
+	mov	dword ptr [rdi+rbx*4],ebp
+
+rmarkr_arguments:
+	mov	rax,qword ptr [rcx]
+	test	al,2
+	je	rmarkr_lazy_node
+
+	movzx	rbp,word ptr (-2)[rax]
+	test	rbp,rbp
+	je	rmarkr_hnf_0
+
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	rmarkr_record
+
+	sub	rbp,2
+	att_je	rmarkr_hnf_2
+	jc	rmarkr_hnf_1
+
+rmarkr_hnf_3:
+	mov	rdx,qword ptr 8[rcx]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx
+
+	mov	rbx,rax
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,[rdi+rbx*4]
+	jne	rmarkr_shared_argument_part
+
+	or	dword ptr [rdi+rbx*4],eax
+
+rmarkr_no_shared_argument_part:
+	or	qword ptr [rcx],2
+	mov	qword ptr 8[rcx],rsi
+	add	rcx,8
+
+	or	qword ptr [rdx],1
+	lea	rdx,[rdx+rbp*8]
+
+	mov	rbp,qword ptr [rdx]
+	mov	qword ptr [rdx],rcx
+	mov	rsi,rdx
+	mov	rcx,rbp
+	att_jmp	rmarkr_node
+
+rmarkr_shared_argument_part:
+	cmp	rdx,rcx
+	att_ja	rmarkr_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr [rdx],rax
+	mov	qword ptr 8[rcx],rbx
+	att_jmp	rmarkr_hnf_1
+
+rmarkr_record:
+	sub	rbp,258
+	je	rmarkr_record_2
+	jb	rmarkr_record_1
+
+rmarkr_record_3:
+	movzx	rbp,word ptr (-2+2)[rax]
+	sub	rbp,1
+	jb	rmarkr_record_3_bb
+	je	rmarkr_record_3_ab
+	dec	rbp
+	je	rmarkr_record_3_aab
+	att_jmp	rmarkr_hnf_3
+
+rmarkr_record_3_bb:
+	mov	rdx,qword ptr (16-8)[rcx]
+	sub	rcx,8
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx
+
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	cmp	rdx,rcx
+	att_ja	rmarkr_next_node
+
+	add	eax,eax
+	jne	rmarkr_bit_in_same_word1
+	inc	rbp
+	mov	rax,1
+rmarkr_bit_in_same_word1:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmarkr_not_yet_linked_bb
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+
+	add	rax,2*8
+
+	mov	rbp,rax 
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkr_next_node
+
+rmarkr_not_yet_linked_bb:
+	or	dword ptr [rdi+rbp*4],eax
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(16+2+1)[rcx]
+	mov	qword ptr 16[rcx],rbp 
+	mov	qword ptr [rdx],rax 
+	att_jmp	rmarkr_next_node
+
+rmarkr_record_3_ab:
+	mov	rdx,qword ptr 8[rcx]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx
+
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	cmp	rdx,rcx
+	att_ja	rmarkr_hnf_1
+
+	add	eax,eax 
+	jne	rmarkr_bit_in_same_word2
+	inc	rbp
+	mov	rax,1
+rmarkr_bit_in_same_word2:
+	test	eax,dword ptr [rdi+rbp*4]
+	je	rmarkr_not_yet_linked_ab
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+
+	add	rax,8
+
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp
+	mov	qword ptr [rdx],rax
+	att_jmp	rmarkr_hnf_1
+
+rmarkr_not_yet_linked_ab:
+	or	dword ptr [rdi+rbp*4],eax
+	mov	rbp,qword ptr [rdx]
+	lea	rax,(8+2+1)[rcx]
+	mov	qword ptr 8[rcx],rbp 
+	mov	qword ptr [rdx],rax
+	att_jmp	rmarkr_hnf_1
+
+rmarkr_record_3_aab:
+	mov	rdx,qword ptr 8[rcx]
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdx 
+
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbp*4]
+	att_jne	rmarkr_shared_argument_part
+	or	dword ptr [rdi+rbp*4],eax
+
+	add	qword ptr [rcx],2
+	mov	qword ptr 8[rcx],rsi
+	add	rcx,8
+
+	mov	rsi,qword ptr [rdx]
+	mov	qword ptr [rdx],rcx 
+	mov	rcx,rsi
+	lea	rsi,1[rdx]
+	att_jmp	rmarkr_node
+
+rmarkr_record_2:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	rmarkr_hnf_2
+	att_je	rmarkr_hnf_1
+	sub	rcx,8
+	att_jmp	rmarkr_next_node
+
+rmarkr_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	rmarkr_hnf_1
+	sub	rcx,8
+	att_jmp	rmarkr_next_node
+
+rmarkr_lazy_node_1:
+	jne	rmarkr_selector_node_1
+
+rmarkr_hnf_1:
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi 
+
+	lea	rsi,2[rcx]
+	mov	rcx,rbp 
+	att_jmp	rmarkr_node
+
+rmarkr_indirection_node:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	mov	rax,rbx
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	and	dword ptr [rdi+rbx*4],eax
+
+	mov	rcx,qword ptr [rcx]
+	att_jmp	rmarkr_node
+
+rmarkr_selector_node_1:
+	add	rbp,3
+	att_je	rmarkr_indirection_node
+
+	mov	rdx,qword ptr [rcx]
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rdx
+	shr	rbx,3
+
+	add	rbp,1
+	jle	rmarkr_record_selector_node_1
+
+	push	rax
+	mov	rax,rbx
+
+	shr	rbx,5
+	and	rax,31
+
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax 
+
+	pop	rax
+	att_jne	rmarkr_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkr_hnf_1
+
+	cmp	word ptr (-2)[rbx],2
+	jbe	rmarkr_small_tuple_or_record
+
+rmarkr_large_tuple_or_record:
+	mov	rbx,qword ptr 16[rdx]
+	add	rbx,qword ptr neg_heap_p3[rip]
+	shr	rbx,3
+
+	push	rax
+	mov	rax,rbx
+
+	shr	rbx,5
+	and	rax,31
+
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax
+
+	pop	rax
+	att_jne	rmarkr_hnf_1
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	push	rcx
+
+	movsxd	rcx,dword ptr (-8)[rax]
+	add	rax,rcx
+
+	mov	rcx,rbx
+	and	rcx,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ecx,dword ptr [r9+rcx]
+	and	dword ptr [rdi+rbx*4],ecx
+
+	movzx	eax,word ptr (4-8)[rax]
+	cmp	rax,16
+	jl	rmarkr_tuple_or_record_selector_node_2
+	mov	rdx,qword ptr 16[rdx]
+	je	rmarkr_tuple_selector_node_2
+	mov	rcx,qword ptr (-24)[rdx+rax]
+	pop	rdx
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rdx],r9
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmarkr_node
+
+rmarkr_tuple_selector_node_2:
+	mov	rcx,qword ptr [rdx]
+	pop	rdx
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rdx],r9
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmarkr_node
+
+rmarkr_record_selector_node_1:
+	je	rmarkr_strict_record_selector_node_1
+
+	push	rax
+	mov	rax,rbx
+
+	shr	rbx,5
+	and	rax,31
+
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax 
+
+	pop	rax
+	att_jne	rmarkr_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkr_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	att_jbe	rmarkr_small_tuple_or_record
+
+	mov	rbx,qword ptr 16[rdx]
+	add	rbx,qword ptr neg_heap_p3[rip]
+	shr	rbx,3
+
+	push	rax
+	mov	rax,rbx
+	shr	rbx,5
+	and	rax,31
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax
+	pop	rax
+	att_jne	rmarkr_hnf_1
+
+rmarkr_small_tuple_or_record:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	lea	rbx,(-8)[rcx+rbx]
+
+	push	rcx
+
+	movsxd	rcx,dword ptr(-8)[rax]
+	add	rax,rcx
+
+	mov	rcx,rbx
+	and	rcx,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ecx,dword ptr [r9+rcx]
+	and	dword ptr [rdi+rbx*4],ecx 
+
+	movzx	eax,word ptr (4-8)[rax]
+	cmp	rax,16
+	att_jle	rmarkr_tuple_or_record_selector_node_2
+	mov	rdx,qword ptr 16[rdx]
+	sub	rax,24
+rmarkr_tuple_or_record_selector_node_2:
+	mov	rcx,qword ptr [rdx+rax]
+	pop	rdx
+	lea	r9,__indirection[rip]
+	mov	qword ptr (-8)[rdx],r9
+	mov	qword ptr [rdx],rcx
+	att_jmp	rmarkr_node
+
+rmarkr_strict_record_selector_node_1:
+	push	rax
+	mov	rax,rbx
+
+	shr	rbx,5
+	and	rax,31
+
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax 
+
+	pop	rax 
+	att_jne	rmarkr_hnf_1
+
+	mov	rbx,qword ptr [rdx]
+	test	bl,2
+	att_je	rmarkr_hnf_1
+
+	cmp	word ptr (-2)[rbx],258
+	jbe	rmarkr_select_from_small_record
+
+	mov	rbx,qword ptr 16[rdx]
+	add	rbx,qword ptr neg_heap_p3[rip]
+
+	push	rax
+	mov	rax,rbx
+
+	shr	rbx,8
+	and	rax,31*8
+
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	mov	ebx,dword ptr [rdi+rbx*4]
+	and	rbx,rax
+
+	pop	rax
+	att_jne	rmarkr_hnf_1
+
+rmarkr_select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rax,rbx
+	sub	rcx,8
+
+	movzx	ebx,word ptr (4-8)[rax]
+	cmp	rbx,16
+	jle	rmarkr_strict_record_selector_node_2
+	add	rbx,qword ptr 16[rdx]
+	mov	rbx,qword ptr (-24)[rbx]
+	jmp	rmarkr_strict_record_selector_node_3
+rmarkr_strict_record_selector_node_2:
+	mov	rbx,qword ptr [rdx+rbx]
+rmarkr_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rbx
+
+	movzx	ebx,word ptr (6-8)[rax]
+	test	rbx,rbx
+	je	rmarkr_strict_record_selector_node_5
+	cmp	rbx,16
+	jle	rmarkr_strict_record_selector_node_4
+	mov	rdx,qword ptr 16[rdx]
+	sub	rbx,24
+rmarkr_strict_record_selector_node_4:
+	mov	rbx,qword ptr [rdx+rbx]
+	mov	qword ptr 16[rcx],rbx
+rmarkr_strict_record_selector_node_5:
+
+	mov	rax,qword ptr ((-8)-8)[rbx]
+	mov	qword ptr [rcx],rax
+	att_jmp	rmarkr_next_node
+
+/* a2,d1: free */
+
+rmarkr_next_node:
+	test	rsi,3
+	jne	rmarkr_parent
+
+	mov	rbp,qword ptr (-8)[rsi]
+	mov	rbx,3
+	
+	and	rbx,rbp
+	sub	rsi,8
+
+	cmp	rbx,3
+	je	rmarkr_argument_part_cycle1
+
+	mov	rdx,qword ptr 8[rsi]
+	mov	qword ptr [rsi],rdx
+
+rmarkr_c_argument_part_cycle1:
+	cmp	rcx,rsi 
+	ja	rmarkr_no_reverse_1
+
+	mov	rdx,qword ptr [rcx]
+	lea	rax,(8+1)[rsi]
+	mov	qword ptr 8[rsi],rdx
+	mov	qword ptr [rcx],rax
+
+	or	rsi,rbx
+	mov	rcx,rbp
+	xor	rcx,rbx
+	att_jmp	rmarkr_node
+
+rmarkr_no_reverse_1:
+	mov	qword ptr 8[rsi],rcx
+	mov	rcx,rbp
+	or	rsi,rbx
+	xor	rcx,rbx
+	att_jmp	rmarkr_node
+
+rmarkr_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp
+	att_je	rmarkr_next_node
+
+	add	rcx,8
+
+	sub	rbp,1
+	att_jle	rmarkr_lazy_node_1
+
+	cmp	rbp,255
+	jge	rmarkr_closure_with_unboxed_arguments
+
+rmarkr_closure_with_unboxed_arguments_:
+	or	qword ptr [rcx],2
+	lea	rcx,[rcx+rbp*8]
+
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi
+	mov	rsi,rcx
+	mov	rcx,rbp
+	att_jmp	rmarkr_node
+
+rmarkr_closure_with_unboxed_arguments:
+/* (a_size+b_size)+(b_size<<8) */
+/*	add	rbp,1 */
+	mov	rax,rbp
+	and	rbp,255
+	shr	rax,8
+	sub	rbp,rax
+/*	sub	rbp,1 */
+	att_jg	rmarkr_closure_with_unboxed_arguments_
+	att_je	rmarkr_hnf_1
+	sub	rcx,8
+	att_jmp	rmarkr_next_node
+
+rmarkr_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	je	rmarkr_int_3
+
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+ 	je	rmarkr_char_3
+
+	jb	rmarkr_no_normal_hnf_0
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx
+
+	mov	rcx,rbx
+	and	rcx,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ecx,dword ptr [r9+rcx]
+	and	dword ptr [rdi+rbx*4],ecx 
+
+	lea	rcx,((-8)-2)[rax]
+	att_jmp	rmarkr_next_node_after_static
+
+rmarkr_int_3:
+	mov	rbp,qword ptr 8[rcx]
+	cmp	rbp,33
+	att_jnc	rmarkr_next_node
+
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx
+
+	mov	rcx,rbx
+	and	rcx,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ecx,dword ptr [r9+rcx]
+	shl	rbp,4
+	and	dword ptr [rdi+rbx*4],ecx 
+
+	lea	rcx,small_integers[rip]
+	add	rcx,rbp
+	att_jmp	rmarkr_next_node_after_static
+
+rmarkr_char_3:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+
+	movzx	rax,byte ptr 8[rcx]
+	add	rbx,rcx
+
+	mov	rbp,rbx
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r9,bit_clear_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	and	dword ptr [rdi+rbx*4],ebp
+
+	shl	rax,4
+	lea	rcx,static_characters[rip]
+	add	rcx,rax
+	att_jmp	rmarkr_next_node_after_static
+
+rmarkr_no_normal_hnf_0:
+	lea	r9,__ARRAY__+2[rip]
+	cmp	rax,r9
+	att_jne	rmarkr_next_node
+
+	mov	rax,qword ptr 16[rcx]
+	test	rax,rax
+	je	rmarkr_lazy_array
+
+	movzx	rbx,word ptr (-2+2)[rax]
+	test	rbx,rbx
+	je	rmarkr_b_array
+
+	movzx	rax,word ptr (-2)[rax]
+	test	rax,rax
+	att_je	rmarkr_b_array
+
+	sub	rax,256
+	cmp	rbx,rax
+	je	rmarkr_a_record_array
+
+rmarkr_ab_record_array:
+	mov	rdx,qword ptr 8[rcx]
+	add	rcx,16
+	push	rcx
+
+	imul	rdx,rax 
+	shl	rdx,3
+
+	sub	rax,rbx
+	add	rcx,8
+	add	rdx,rcx
+	att_call	reorder
+	
+	pop	rcx
+	mov	rax,rbx
+	imul	rax,qword ptr (-8)[rcx]
+	jmp	rmarkr_lr_array
+
+rmarkr_b_array:
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rcx
+	add	rax,8
+	mov	rbp,rax
+	and	rax,31*8
+	shr	rbp,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	or	dword ptr [rdi+rbp*4],eax
+	att_jmp	rmarkr_next_node
+
+rmarkr_a_record_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+	cmp	rbx,2
+	att_jb	rmarkr_lr_array
+
+	imul	rax,rbx
+	att_jmp	rmarkr_lr_array
+
+rmarkr_lazy_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+
+rmarkr_lr_array:
+	mov	rbx,qword ptr neg_heap_p3[rip]
+	add	rbx,rcx 
+	shr	rbx,3
+	add	rbx,rax 
+
+	mov	rdx,rbx
+	and	rbx,31
+	shr	rdx,5
+	lea	r9,bit_set_table[rip]
+	mov	ebx,dword ptr [r9+rbx*4]
+	or	dword ptr [rdi+rdx*4],ebx
+
+	cmp	rax,1
+	jbe	rmarkr_array_length_0_1
+
+	mov	rdx,rcx 
+	lea	rcx,[rcx+rax*8]
+
+	mov	rax,qword ptr [rcx]
+	mov	rbx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax
+	mov	qword ptr [rcx],rbx 
+	
+	mov	rax,qword ptr (-8)[rcx]
+	sub	rcx,8
+	add	rax,2
+	mov	rbx,qword ptr (-8)[rdx]
+	sub	rdx,8
+	mov	qword ptr [rcx],rbx
+	mov	qword ptr [rdx],rax
+
+	mov	rax,qword ptr (-8)[rcx]
+	sub	rcx,8
+	mov	qword ptr [rcx],rsi
+	mov	rsi,rcx
+	mov	rcx,rax
+	att_jmp	rmarkr_node
+
+rmarkr_array_length_0_1:
+	lea	rcx,-16[rcx]
+	att_jb	rmarkr_next_node
+
+	mov	rbx,qword ptr 24[rcx]
+	mov	rbp,qword ptr 16[rcx]
+	mov	qword ptr 24[rcx],rbp
+	mov	rbp,qword ptr 8[rcx]
+	mov	qword ptr 16[rcx],rbp
+	mov	qword ptr 8[rcx],rbx
+	add	rcx,8
+	att_jmp	rmarkr_hnf_1
+
+/* a2: free */
+
+rmarkr_parent:
+	mov	rbx,rsi
+	and	rbx,3
+
+	and	rsi,-4
+	je	end_rmarkr
+
+	sub	rbx,1
+	je	rmarkr_argument_part_parent
+
+	mov	rbp,qword ptr [rsi]
+
+	cmp	rcx,rsi
+	ja	rmarkr_no_reverse_2
+
+	mov	rdx,rcx
+	lea	rax,1[rsi]
+	mov	rcx,qword ptr [rdx]
+	mov	qword ptr [rdx],rax
+
+rmarkr_no_reverse_2:
+	mov	qword ptr [rsi],rcx 
+	lea	rcx,(-8)[rsi]
+	mov	rsi,rbp
+	att_jmp	rmarkr_next_node
+
+rmarkr_argument_part_parent:
+	mov	rbp,qword ptr [rsi]
+
+	mov	rdx,rsi
+	mov	rsi,rcx
+	mov	rcx,rdx
+
+rmarkr_skip_upward_pointers:
+	mov	rax,rbp
+	and	rax,3
+	cmp	rax,3
+	jne	rmarkr_no_upward_pointer
+
+	lea	rdx,(-3)[rbp]
+	mov	rbp,qword ptr (-3)[rbp]
+	att_jmp	rmarkr_skip_upward_pointers
+
+rmarkr_no_upward_pointer:
+	cmp	rsi,rcx
+	ja	rmarkr_no_reverse_3
+
+	mov	rbx,rsi
+	mov	rsi,qword ptr [rsi]
+	lea	rax,1[rcx]
+	mov	qword ptr [rbx],rax
+
+rmarkr_no_reverse_3:
+	mov	qword ptr [rdx],rsi
+	lea	rsi,(-8)[rbp]
+
+	and	rsi,-4
+
+	mov	rdx,rsi
+	mov	rbx,3
+
+	mov	rbp,qword ptr [rsi]
+
+	and	rbx,rbp
+	mov	rax,qword ptr 8[rdx]
+
+	or	rsi,rbx
+	mov	qword ptr [rdx],rax
+
+	cmp	rcx,rdx
+	ja	rmarkr_no_reverse_4
+
+	mov	rax,qword ptr [rcx]
+	mov	qword ptr 8[rdx],rax
+	lea	rax,(8+2+1)[rdx]
+	mov	qword ptr [rcx],rax
+	mov	rcx,rbp
+	and	rcx,-4
+	att_jmp	rmarkr_node
+
+rmarkr_no_reverse_4:
+	mov	qword ptr 8[rdx],rcx
+	mov	rcx,rbp
+	and	rcx,-4
+	att_jmp	rmarkr_node
+
+rmarkr_argument_part_cycle1:
+	mov	rax,qword ptr 8[rsi]
+	push	rdx
+
+rmarkr_skip_pointer_list1:
+	mov	rdx,rbp
+	and	rdx,-4
+	mov	rbp,qword ptr [rdx]
+	mov	rbx,3
+	and	rbx,rbp
+	cmp	rbx,3
+	att_je	rmarkr_skip_pointer_list1
+
+	mov	qword ptr [rdx],rax
+	pop	rdx
+	att_jmp	rmarkr_c_argument_part_cycle1
+
+rmarkr_next_node_after_static:
+	test	rsi,3
+	jne	rmarkr_parent_after_static
+
+	mov	rbp,qword ptr (-8)[rsi]
+	mov	rbx,3
+	
+	and	rbx,rbp
+	sub	rsi,8
+
+	cmp	rbx,3
+	je	rmarkr_argument_part_cycle2
+	
+	mov	rax,qword ptr 8[rsi]
+	mov	qword ptr [rsi],rax
+
+rmarkr_c_argument_part_cycle2:
+	mov	qword ptr 8[rsi],rcx
+	mov	rcx,rbp
+	or	rsi,rbx
+	xor	rcx,rbx
+	att_jmp	rmarkr_node
+
+rmarkr_parent_after_static:
+	mov	rbx,rsi
+	and	rbx,3
+
+	and	rsi,-4
+	je	end_rmarkr_after_static
+
+	sub	rbx,1
+	je	rmarkr_argument_part_parent_after_static
+
+	mov	rbp,qword ptr [rsi]
+	mov	qword ptr [rsi],rcx
+	lea	rcx,(-8)[rsi]
+	mov	rsi,rbp
+	att_jmp	rmarkr_next_node
+	
+rmarkr_argument_part_parent_after_static:
+	mov	rbp,qword ptr [rsi]
+
+	mov	rdx,rsi
+	mov	rsi,rcx
+	mov	rcx,rdx
+
+/*	movl	rbp,qword ptr [rdx] */
+rmarkr_skip_upward_pointers_2:
+	mov	rax,rbp
+	and	rax,3
+	cmp	rax,3
+	att_jne	rmarkr_no_reverse_3
+
+	lea	rdx,(-3)[rbp]
+	mov	rbp,qword ptr (-3)[rbp]
+	att_jmp	rmarkr_skip_upward_pointers_2
+
+rmarkr_argument_part_cycle2:
+	mov	rax,qword ptr 8[rsi]
+	push	rdx
+
+rmarkr_skip_pointer_list2:
+	mov	rdx,rbp 
+	and	rdx,-4
+	mov	rbp,qword ptr [rdx]
+	mov	rbx,3
+	and	rbx,rbp
+	cmp	rbx,3
+	att_je	rmarkr_skip_pointer_list2
+
+	mov	qword ptr [rdx],rax
+	pop	rdx
+	att_jmp	rmarkr_c_argument_part_cycle2
+
+end_rmarkr_after_static:
+	mov	rsi,qword ptr [rsp]
+	add	rsp,16
+	mov	qword ptr [rsi],rcx
+	jmp	rmarkr_next_stack_node
+
+end_rmarkr:
+	pop	rsi
+	pop	rbx
+
+	cmp	rcx,rbx
+	ja	rmarkr_no_reverse_5
+
+	mov	rdx,rcx
+	lea	rax,1[rsi]
+	mov	rcx,qword ptr [rcx]
+	mov	qword ptr [rdx],rax
+
+rmarkr_no_reverse_5:
+	mov	qword ptr [rsi],rcx
+
+rmarkr_next_stack_node:
+	cmp	rsp,qword ptr end_stack[rip]
+	jae	rmarkr_end
+
+	mov	rcx,qword ptr [rsp]
+	mov	rsi,qword ptr 8[rsp]
+	add	rsp,16
+
+	cmp	rcx,1
+	att_ja	rmark_using_reversal
+
+	test	qword ptr _flags[rip],4096
+	att_je	rmark_next_node_
+	att_jmp	rmarkp_next_node_
+
+rmarkr_end:
+	test	qword ptr _flags[rip],4096
+	att_je	rmark_next_node
+	att_jmp	rmarkp_next_node
diff --git a/macho64/acopy.s b/macho64/acopy.s
new file mode 100644
index 0000000..b4b68bc
--- /dev/null
+++ b/macho64/acopy.s
@@ -0,0 +1,1267 @@
+
+COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP = 1
+
+	push	rsi
+
+	mov	rdi,heap_p2[rip]
+
+	mov	rax,heap_size_257[rip]
+	shl	rax,7
+	mov	semi_space_size[rip],rax 
+	lea	rsi,[rdi+rax]
+
+	mov	qword ptr (heap2_begin_and_end+8)[rip],rsi 
+
+	mov	rax,qword ptr caf_list[rip]
+	test	rax,rax 
+	je	end_copy_cafs
+
+copy_cafs_lp:
+	push	(-8)[rax]
+	
+	lea	rbp,8[rax]
+	mov	rbx,qword ptr [rax]
+	sub	rbx,1
+	call	copy_lp2
+	
+	pop	rax 
+	test	rax,rax 
+	att_jne	copy_cafs_lp
+
+end_copy_cafs:
+	mov	rbx,qword ptr [rsp]
+	mov	rbp,stack_p[rip]
+	sub	rbx,rbp 
+	shr	rbx,3
+
+	sub	rbx,1
+	jb	end_copy0
+	att_call	copy_lp2
+end_copy0:
+	mov	rbp,heap_p2[rip]
+
+	jmp	copy_lp1
+/* */
+/*	Copy all referenced nodes to the other semi space */
+/* */
+
+in_hnf_1_2:
+	dec	rbx 
+copy_lp2_lp1:
+	att_call	copy_lp2
+copy_lp1:
+	cmp	rbp,rdi 
+	jae	end_copy1
+
+	mov	rax,[rbp]
+	add	rbp,8
+	test	al,2
+	je	not_in_hnf_1
+in_hnf_1:
+	movzx	rbx,word ptr (-2)[rax]
+
+	test	rbx,rbx 
+	je	copy_array_21
+
+	cmp	rbx,2
+	att_jbe	in_hnf_1_2
+
+	cmp	rbx,256
+	jae	copy_record_21
+	
+	mov	rax,8[rbp]
+
+	test	al,1
+	jne	node_without_arguments_part
+
+	push	rbx 
+	xor	rbx,rbx 
+	
+	att_call	copy_lp2
+
+	pop	rbx 
+	add	rbp,8
+
+	sub	rbx,2
+	att_jmp	copy_lp2_lp1
+
+node_without_arguments_part:
+	dec	rax 
+	xor	rbx,rbx 
+
+	mov	8[rbp],rax 
+	att_call	copy_lp2
+	
+	add	rbp,8
+	att_jmp	copy_lp1
+
+copy_record_21:
+	sub	rbx,258
+	ja	copy_record_arguments_3
+
+	movzx	rbx,word ptr (-2+2)[rax]
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	att_jb	in_hnf_1_2
+
+	sub	rbx,1
+	att_ja	copy_lp2_lp1
+	jmp	copy_node_arity1
+ .else
+	jb	copy_record_arguments_1
+
+	sub	rbx,1
+	ja	copy_lp2_lp1
+	je	copy_node_arity1
+	add	rbp,16
+	jmp	copy_lp1
+
+copy_record_arguments_1:
+	dec	rbx
+	jmp	copy_lp2_lp1
+	je	copy_lp2_lp1
+	add	rbp,8
+	jmp	copy_lp1
+ .endif
+
+copy_record_arguments_3:
+	test	byte ptr 8[rbp],1
+	jne	record_node_without_arguments_part
+
+	movzx	rdx,word ptr (-2+2)[rax]
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	sub	rdx,1
+ .else
+	test	rdx,rdx 
+	je	copy_record_arguments_3b
+	sub	rdx,1
+	je	copy_record_arguments_3abb
+ .endif
+
+	lea	rcx,(3*8)[rbp+rbx*8]
+	push	rcx 
+	push	rdx 
+
+	sub	rbx,rbx 
+	att_call	copy_lp2
+	
+	add	rbp,8
+	pop	rbx 
+	dec	rbx 
+	att_call	copy_lp2
+
+	pop	rbp 
+	att_jmp	copy_lp1
+
+ .if ! COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+copy_record_arguments_3abb:
+	push	rbx 
+	sub	rbx,rbx 
+	
+	call	copy_lp2
+	
+	pop	rbx 
+	
+	lea	rbp,(2*8)[rbp+rbx*8]
+	jmp	copy_lp1
+
+copy_record_arguments_3b:
+	lea	rbp,(3*8)[rbp+rbx*8]
+	jmp	copy_lp1
+ .endif
+
+record_node_without_arguments_part:
+	and	qword ptr 8[rbp],-2
+
+ .if ! COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	cmp	word ptr (-2+2)[rax],0
+	je	record_node_without_arguments_part_3b
+ .endif
+
+	sub	rbx,rbx 
+	att_call	copy_lp2
+
+	add	rbp,8
+	att_jmp	copy_lp1
+
+ .if ! COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+record_node_without_arguments_part_3b:
+	add	rbp,16
+	jmp	copy_lp1
+ .endif
+
+not_in_hnf_1:
+	movsxd	rbx,dword ptr (-4)[rax]
+	cmp	rbx,257
+	jge	copy_unboxed_closure_arguments
+	sub	rbx,1
+	att_jg	copy_lp2_lp1
+
+copy_node_arity1:
+	xor	rbx,rbx 
+	att_call	copy_lp2
+
+	add	rbp,8
+	att_jmp	copy_lp1
+
+copy_unboxed_closure_arguments:
+	je	copy_unboxed_closure_arguments1
+
+	xor	rax,rax
+	mov	al,bh 
+	and	rbx,255
+	sub	rbx,rax 
+
+	sub	rbx,1
+	jl	copy_unboxed_closure_arguments_without_pointers
+	
+	push	rax 
+	att_call	copy_lp2
+	pop	rax 
+
+copy_unboxed_closure_arguments_without_pointers:
+	lea	rbp,[rbp+rax*8]
+	att_jmp	copy_lp1
+
+copy_unboxed_closure_arguments1:
+	add	rbp,16
+	att_jmp	copy_lp1
+
+copy_array_21:
+	mov	rbx,qword ptr 8[rbp]
+	add	rbp,16
+	test	rbx,rbx 
+	je	copy_array_21_a
+
+	movzx	rax,word ptr (-2)[rbx]
+	movzx	rbx,word ptr (-2+2)[rbx]
+	sub	rax,256
+	test	rbx,rbx 
+	je	copy_array_21_b
+
+	cmp	rbx,rax 
+	je	copy_array_21_r_a
+
+copy_array_21_ab:
+	cmp	qword ptr (-16)[rbp],0
+	att_je	copy_lp1
+
+	sub	rax,rbx 
+	shl	rax,3
+	sub	rbx,1
+
+	push	rbx 
+	push	rax 
+	mov	rbx,qword ptr (-16)[rbp]
+	sub	rbx,1
+	push	rbx 
+
+copy_array_21_lp_ab:
+	mov	rbx,qword ptr 16[rsp]
+	att_call	copy_lp2
+
+	add	rbp,qword ptr 8[rsp]
+	sub	qword ptr [rsp],1
+	att_jnc	copy_array_21_lp_ab
+	
+	add	rsp,24
+	att_jmp	copy_lp1
+
+copy_array_21_b:
+	mov	rbx,qword ptr (-16)[rbp]
+	imul	rbx,rax 
+	lea	rbp,[rbp+rbx*8]
+	att_jmp	copy_lp1
+
+copy_array_21_r_a:
+	mov	rbx,qword ptr (-16)[rbp]
+	imul	rbx,rax
+	sub	rbx,1
+	att_jc	copy_lp1
+	att_jmp	copy_lp2_lp1
+
+copy_array_21_a:
+	mov	rbx,qword ptr (-16)[rbp]
+	sub	rbx,1
+	att_jc	copy_lp1
+	att_jmp	copy_lp2_lp1
+
+/* */
+/*	Copy nodes to the other semi-space */
+/* */
+
+copy_lp2:
+	mov	rdx,qword ptr [rbp]
+
+/* selectors: */
+continue_after_selector_2:
+	mov	rcx,qword ptr [rdx]
+	test	cl,2
+	je	not_in_hnf_2
+
+in_hnf_2:
+	movzx	rax,word ptr (-2)[rcx]
+	test	rax,rax
+	je	copy_arity_0_node2
+
+	cmp	rax,256
+	jae	copy_record_2
+
+	sub	rax,2
+	mov	[rbp],rdi 
+
+	lea	rbp,8[rbp ]
+	ja	copy_hnf_node2_3
+
+	mov	[rdi],rcx
+	jb	copy_hnf_node2_1
+
+	inc	rdi
+	mov	rcx,8[rdx]
+
+	mov	[rdx],rdi 
+	mov	rax,16[rdx]
+
+	sub	rbx,1
+	mov	(8-1)[rdi],rcx 
+
+	mov	(16-1)[rdi],rax 
+	lea	rdi,(24-1)[rdi]
+
+	att_jae	copy_lp2
+	ret
+
+copy_hnf_node2_1:
+	inc	rdi 
+	mov	rax,8[rdx]
+
+	sub	rbx,1
+	mov	[rdx],rdi
+
+	mov	(8-1)[rdi],rax 
+	lea	rdi,(16-1)[rdi]
+
+	att_jae	copy_lp2
+	ret
+
+copy_hnf_node2_3:
+	mov	[rdi],rcx 
+	inc	rdi 
+
+	mov	[rdx],rdi 
+	mov	rcx,8[rdx]
+
+	mov	(8-1)[rdi],rcx 
+	mov	rcx,16[rdx]
+
+	add	rdi,24-1
+	mov	rdx,[rcx]
+	
+	test	dl,1
+	jne	arguments_already_copied_2
+
+	mov	(-8)[rdi],rdi 
+	add	rcx,8
+
+	mov	[rdi],rdx 
+	inc	rdi
+
+	mov	(-8)[rcx],rdi 
+	add	rdi,8-1
+
+cp_hnf_arg_lp2:
+	mov	rdx,[rcx]
+	add	rcx,8
+
+	mov	[rdi],rdx 
+	add	rdi,8
+
+	dec	rax 
+	att_jne	cp_hnf_arg_lp2
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+arguments_already_copied_2:
+	mov	(-8)[rdi],rdx 
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_arity_0_node2:
+	lea	r9,dINT+2[rip]
+	cmp	rcx,r9
+	jb	copy_real_file_or_string_2
+
+	lea	r9,CHAR+2[rip]
+	cmp	rcx,r9
+	ja	copy_normal_hnf_0_2
+
+copy_int_bool_or_char_2:
+	mov	rax,8[rdx]
+	je	copy_char_2
+
+	lea	r9,dINT+2[rip]
+	cmp	rcx,r9
+	jne	no_small_int_or_char_2
+
+copy_int_2:
+	cmp	rax,33
+	att_jae	no_small_int_or_char_2
+
+	shl	rax,4
+	add	rbp,8
+
+	lea	r9,small_integers[rip]
+	add	rax,r9
+	sub	rbx,1
+
+	mov	(-8)[rbp],rax 
+	att_jae	copy_lp2
+	ret
+
+copy_char_2:	
+	and	rax,255
+
+	shl	rax,4
+	add	rbp,8
+
+	lea	r9,static_characters[rip]
+	add	rax,r9
+	sub	rbx,1
+
+	mov	(-8)[rbp],rax 
+	att_jae	copy_lp2
+	ret
+	
+no_small_int_or_char_2:
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+copy_record_node2_1_b:
+ .endif
+	mov	(-16)[rsi],rcx 
+	add	rbp,8
+
+	mov	(-8)[rsi],rax 
+	sub	rsi,15
+
+	mov	[rdx],rsi 
+	dec	rsi
+
+	mov	(-8)[rbp],rsi 
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_normal_hnf_0_2:
+	sub	rcx,2-(-8)
+	sub	rbx,1
+
+	mov	[rbp],rcx 
+	lea	rbp,8[rbp]
+	att_jae	copy_lp2
+	ret
+
+copy_real_file_or_string_2:
+	lea	r9,__STRING__+2[rip]
+	cmp	rcx,r9
+	jbe	copy_string_or_array_2
+
+copy_real_or_file_2:
+	mov	(-24)[rsi],rcx 
+	sub	rsi,24-1
+
+	mov	[rdx],rsi 
+	dec	rsi 
+
+	mov	rax,8[rdx]
+	mov	rcx,16[rdx]
+
+	mov	[rbp],rsi 
+	add	rbp,8
+
+	mov	8[rsi],rax 
+	sub	rbx,1
+
+	mov	16[rsi],rcx 
+
+	att_jae	copy_lp2
+	ret
+
+already_copied_2:
+	dec	rcx 
+	sub	rbx,1
+
+	mov	[rbp],rcx 
+	lea	rbp,8[rbp]
+
+	att_jae	copy_lp2
+	ret
+
+copy_record_2:
+	sub	rax,258
+	ja	copy_record_node2_3
+
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+ 	jb	copy_record_node2_1
+ 
+ 	cmp	word ptr (-2+2)[rcx],0
+	att_je	copy_real_or_file_2
+
+	mov	qword ptr [rbp],rdi
+	mov	qword ptr [rdi],rcx
+
+	lea	rcx,1[rdi]
+	mov	rax,qword ptr 8[rdx]
+
+	mov	qword ptr [rdx],rcx
+
+	mov	qword ptr 8[rdi],rax
+	mov	rax,qword ptr 16[rdx]
+
+	add	rbp,8
+	mov	qword ptr 16[rdi],rax
+
+	add	rdi,24
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_record_node2_1:
+	mov	rax,qword ptr 8[rdx]
+
+	cmp	word ptr (-2+2)[rcx],0
+	att_je	copy_record_node2_1_b
+
+	mov	qword ptr [rbp],rdi
+	mov	qword ptr [rdi],rcx
+
+	lea	rcx,1[rdi]
+	mov	qword ptr 8[rdi],rax
+
+	mov	qword ptr [rdx],rcx
+	add	rbp,8
+
+	add	rdi,16
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+ .else
+	mov	qword ptr [rbp],rdi
+	mov	qword ptr [rdi],rcx
+
+	lea	rcx,1[rdi]
+	mov	rax,qword ptr 8[rdx]
+
+	mov	qword ptr [rdx],rcx
+	jb	copy_record_node2_1
+
+	mov	qword ptr 8[rdi],rax
+	mov	rax,qword ptr 16[rdx]
+
+	add	rbp,8
+	mov	qword ptr 16[rdi],rax
+
+	add	rdi,24
+	sub	rbx,1
+	jae	copy_lp2
+	ret
+
+copy_record_node2_1:
+	add	rbp,8
+	mov	qword ptr 8[rdi],rax 
+
+	add	rdi,16
+	sub	rbx,1
+	jae	copy_lp2
+	ret
+ .endif
+
+copy_record_node2_3:
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+ 	cmp	word ptr (-2+2)[rcx],1
+ 	jbe	copy_record_node2_3_ab_or_b
+ .endif
+
+	push	rax 
+	lea	rax,1[rdi]
+	
+	mov	qword ptr [rdx],rax 
+	mov	rax,qword ptr 16[rdx]
+
+	mov	qword ptr [rdi],rcx 
+	mov	rdx,qword ptr 8[rdx]
+
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	mov	qword ptr 8[rdi],rdx 
+	mov	qword ptr [rbp],rdi
+	add	rbp,8
+
+	mov	rcx,rax
+	test	byte ptr [rax],1		
+	jne	record_arguments_already_copied_2
+ .else
+	mov	rcx,rax 
+	sub	rax,qword ptr heap_p1
+
+	shr	rax,4
+	mov	qword ptr 8[rdi],rdx 
+
+	mov	rdx,rax 
+	and	rax,31
+
+	shr	rdx,3
+	mov	qword ptr [rbp],rdi
+
+	and	rdx,-4
+	mov	eax,dword ptr (bit_set_table)[rax*4]
+
+	add	rdx,qword ptr heap_copied_vector
+	add	rbp,8
+
+	test	eax,[rdx]
+	jne	record_arguments_already_copied_2
+
+	or	[rdx],eax
+ .endif
+	lea	rdx,24[rdi]
+
+	pop	rax
+	mov	qword ptr 16[rdi],rdx
+
+	add	rdi,25
+	mov	rdx,qword ptr [rcx]
+
+	mov	qword ptr [rcx],rdi
+	add	rcx,8
+
+	mov	qword ptr (-1)[rdi],rdx
+	add	rdi,7
+
+cp_record_arg_lp2:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+
+	mov	qword ptr [rdi],rdx
+	add	rdi,8
+
+	sub	rax,1
+	att_jne	cp_record_arg_lp2
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+record_arguments_already_copied_2:
+	mov	rdx,qword ptr [rcx]
+	pop	rax 
+
+	mov	qword ptr 16[rdi],rdx 
+	add	rdi,24
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+copy_record_node2_3_ab_or_b:
+	jb	copy_record_node2_3_b
+
+	push	rax
+	lea	rax,1[rdi]
+
+	mov	qword ptr [rdx],rax 
+	mov	rax,qword ptr 16[rdx]
+
+	mov	qword ptr [rdi],rcx 
+	mov	rdx,qword ptr 8[rdx]
+
+	mov	rcx,rax 
+	sub	rax,qword ptr heap_p1[rip]
+
+	shr	rax,4
+	mov	qword ptr 8[rdi],rdx 
+
+	mov	rdx,rax 
+	and	rax,31
+
+	shr	rdx,3
+	mov	qword ptr [rbp],rdi
+
+	and	rdx,-4
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+
+	add	rdx,qword ptr heap_copied_vector[rip]
+	add	rbp,8
+
+	test	eax,[rdx]
+	att_jne	record_arguments_already_copied_2
+
+	or	[rdx],eax
+	pop	rax
+
+	sub	rsi,8
+
+	shl	rax,3
+	sub	rsi,rax
+	
+	push	rsi
+	add	rsi,1
+	
+	mov	qword ptr 16[rdi],rsi
+	add	rdi,24
+	
+	mov	rdx,qword ptr [rcx]
+	jmp	cp_record_arg_lp3_c
+
+copy_record_node2_3_b:
+	push	rax
+	lea	rax,(-24+1)[rsi]
+
+	mov	qword ptr [rdx],rax 
+	mov	rax,qword ptr 16[rdx]
+
+	mov	qword ptr (-24)[rsi],rcx 
+	mov	rdx,qword ptr 8[rdx]
+
+	mov	rcx,rax 
+	sub	rax,qword ptr heap_p1[rip]
+
+	shr	rax,4
+	mov	qword ptr (-16)[rsi],rdx 
+
+	mov	rdx,rax 
+	and	rax,31
+	sub	rsi,24
+
+	shr	rdx,3
+	mov	qword ptr [rbp],rsi
+
+	and	rdx,-4
+	lea	r9,bit_set_table[rip]
+	mov	eax,dword ptr [r9+rax*4]
+
+	add	rdx,qword ptr heap_copied_vector[rip]
+	add	rbp,8
+
+	test	eax,[rdx]
+	jne	record_arguments_already_copied_3_b
+
+	or	[rdx],eax
+	pop	rax
+
+	mov	rdx,rsi
+	sub	rsi,8
+
+	shl	rax,3
+	sub	rsi,rax
+
+	mov	qword ptr 16[rdx],rsi
+
+	mov	rdx,qword ptr [rcx]
+
+	push	rsi
+	add	rsi,1
+
+cp_record_arg_lp3_c:
+	mov	qword ptr [rcx],rsi
+	add	rcx,8
+	mov	qword ptr (-1) [rsi],rdx
+	add	rsi,7
+
+cp_record_arg_lp3:
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	
+	mov	qword ptr [rsi],rdx
+	add	rsi,8
+	
+	sub	rax,8
+	att_jne	cp_record_arg_lp3
+	
+	pop	rsi
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+record_arguments_already_copied_3_b:
+	mov	rdx,qword ptr [rcx]
+	pop	rax
+	
+	sub	rdx,1
+	mov	qword ptr 16[rsi],rdx
+	
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+ .endif
+
+not_in_hnf_2:
+	test	cl,1
+	att_jne	already_copied_2
+
+	movsxd	rax,dword ptr (-4)[rcx]
+	test	rax,rax 
+	jle	copy_arity_0_node2_
+
+copy_node2_1_:
+	and	rax,255
+	sub	rax,2
+	jl	copy_arity_1_node2
+copy_node2_3:
+	mov	[rbp],rdi 
+	add	rbp,8
+	mov	[rdi],rcx 
+	inc	rdi
+	mov	[rdx],rdi
+	mov	rcx,8[rdx]
+	add	rdx,16
+	mov	(8-1)[rdi],rcx 
+	add	rdi,16-1
+
+cp_arg_lp2:
+	mov	rcx,[rdx]
+	add	rdx,8
+	mov	[rdi],rcx 
+	add	rdi,8
+	sub	rax,1
+	att_jae	cp_arg_lp2
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_arity_1_node2:
+copy_arity_1_node2_:
+	mov	[rbp],rdi 
+	inc	rdi 
+
+	add	rbp,8
+	mov	[rdx],rdi 
+
+	mov	rax,8[rdx]
+	mov	(-1)[rdi],rcx 
+
+	mov	(8-1)[rdi],rax 
+	add	rdi,24-1
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_indirection_2:
+	mov	rax,rdx 
+	mov	rdx,8[rdx]
+
+	mov	rcx,[rdx]
+	test	cl,2
+	att_jne	in_hnf_2
+
+	test	cl,1
+	att_jne	already_copied_2
+
+	cmp	dword ptr (-4)[rcx],-2
+	je	skip_indirections_2
+
+	movsxd	rax,dword ptr(-4)[rcx]
+	test	rax,rax 
+	att_jle	copy_arity_0_node2_
+	att_jmp	copy_node2_1_
+
+skip_indirections_2:
+	mov	rdx,8[rdx]
+
+	mov	rcx,[rdx]
+	test	cl,2
+	jne	update_indirection_list_2
+	test	cl,1
+	att_jne	update_indirection_list_2
+
+	cmp	dword ptr (-4)[rcx],-2
+	att_je	skip_indirections_2
+
+update_indirection_list_2:
+	lea	rcx,8[rax]
+	mov	rax,8[rax]
+	mov	[rcx],rdx 
+	cmp	rdx,rax 
+	att_jne	update_indirection_list_2
+
+	att_jmp	continue_after_selector_2
+
+copy_selector_2:
+	cmp	rax,-2
+	att_je	copy_indirection_2
+	jl	copy_record_selector_2
+
+	mov	rax,8[rdx]
+
+	mov	r10,[rax]
+	test	r10b,2
+ 	att_je	copy_arity_1_node2_
+
+	movsxd	r11,dword ptr (-8)[rcx]
+
+	cmp	word ptr (-2)[r10],2
+	jbe	copy_selector_2_
+ 
+ 	mov	r10,16[rax]
+
+	test	byte ptr [r10],1
+	att_jne	copy_arity_1_node2_
+
+	movzx	r11,word ptr (4-8)[rcx+r11]
+	lea	r9,__indirection[rip]
+	mov	qword ptr [rdx],r9
+
+	cmp	r11,16
+	jl	copy_selector_2_1
+	je	copy_selector_2_2
+
+	mov	rcx,qword ptr (-24)[r10+r11]
+	mov	qword ptr 8[rdx],rcx
+	mov	rdx,rcx
+	att_jmp	continue_after_selector_2
+
+copy_selector_2_1:
+	mov	rcx,qword ptr 8[rax]
+	mov	qword ptr 8[rdx],rcx
+	mov	rdx,rcx
+	att_jmp	continue_after_selector_2
+
+copy_selector_2_2:
+	mov	rcx,qword ptr [r10]
+	mov	qword ptr 8[rdx],rcx
+	mov	rdx,rcx
+	att_jmp	continue_after_selector_2
+
+copy_selector_2_:
+	movzx	r11,word ptr (4-8)[rcx+r11]
+	lea	r9,__indirection[rip]
+	mov	qword ptr [rdx],r9
+
+	mov	rcx,qword ptr [rax+r11]
+	mov	qword ptr 8[rdx],rcx
+	mov	rdx,rcx
+	att_jmp	continue_after_selector_2
+
+copy_record_selector_2:
+	cmp	rax,-3
+ 	mov	rax,qword ptr 8[rdx]
+	mov	r10,qword ptr [rax]
+	je	copy_strict_record_selector_2
+
+ 	test	r10b,2
+	att_je	copy_arity_1_node2_
+
+	movsxd	r11,dword ptr (-8)[rcx]
+
+	cmp	word ptr (-2)[r10],258
+	jbe	copy_record_selector_2_
+
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	cmp	word ptr (-2+2)[r10],2
+	jae	copy_selector_2__
+ .endif
+
+	mov	r12,qword ptr 16[rax]
+
+	lea	r10,(-24)[r12]
+	sub	r12,qword ptr heap_p1[rip]
+
+	mov	r13,r12 
+	and	r12,31*16
+
+	shr	r13,7
+	
+	shr	r12,2
+	and	r13,-4
+
+	add	r13,qword ptr heap_copied_vector[rip]
+
+	lea	r9,bit_set_table[rip]
+	mov	r12d,dword ptr [r9+r12]
+
+	and	r12d,dword ptr [r13]
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	att_je	copy_record_selector_2_
+	att_jmp	copy_arity_1_node2_
+copy_selector_2__:
+	mov	r12,qword ptr 16[rax]
+	lea	r10,(-24)[r12]
+	test	byte ptr [r12],1
+	att_jne	copy_arity_1_node2_	
+ .else
+	jne	copy_arity_1_node2_
+ .endif
+copy_record_selector_2_:
+	movzx	r11,word ptr (4-8)[rcx+r11]
+	lea	r9,__indirection[rip]
+	mov	qword ptr [rdx],r9
+
+	cmp	r11,16
+	jle	copy_record_selector_3
+	mov	rax,r10
+copy_record_selector_3:
+	mov	rcx,qword ptr [rax+r11]
+	mov	qword ptr 8[rdx],rcx
+	mov	rdx,rcx
+	att_jmp	continue_after_selector_2
+
+copy_strict_record_selector_2:
+	test	r10b,2
+	att_je	copy_arity_1_node2_
+
+	movsxd	r11,dword ptr (-8)[rcx]
+
+	cmp	word ptr (-2)[r10],258
+	jbe	copy_strict_record_selector_2_
+
+ .if COPY_RECORDS_WITHOUT_POINTERS_TO_END_OF_HEAP
+	cmp	word ptr (-2+2)[r10],2
+	jb	copy_strict_record_selector_2_b
+
+	mov	r12,qword ptr 16[rax]
+	lea	r10,(-24)[r12]
+	test	byte ptr [r12],1
+	att_jne	copy_arity_1_node2_	
+
+	att_jmp	copy_strict_record_selector_2_
+
+copy_strict_record_selector_2_b:
+ .endif
+
+	mov	r12,qword ptr 16[rax]
+
+	lea	r10,(-24)[r12]
+	sub	r12,qword ptr heap_p1[rip]
+
+	mov	r13,r12 
+	and	r12,31*16
+
+	shr	r13,7
+
+	shr	r12,2
+	and	r13,-4
+
+	add	r13,qword ptr heap_copied_vector[rip]
+
+	lea	r9,bit_set_table[rip]
+	mov	r12d,dword ptr [r9+r12]
+	
+	and	r12d,[r13]
+
+	att_jne	copy_arity_1_node2_
+
+copy_strict_record_selector_2_:
+	add	r11,rcx
+	movzx	rcx,word ptr (4-8)[r11]
+	cmp	rcx,16
+	jle	copy_strict_record_selector_3
+	mov	rcx,qword ptr [r10+rcx]
+	jmp	copy_strict_record_selector_4
+copy_strict_record_selector_3:
+	mov	rcx,qword ptr [rax+rcx]
+copy_strict_record_selector_4:
+	mov	qword ptr 8[rdx],rcx
+
+	movzx	rcx,word ptr (6-8)[r11]
+	test	rcx,rcx
+	je	copy_strict_record_selector_6
+	cmp	rcx,16
+	jle	copy_strict_record_selector_5
+	mov	rax,r10
+copy_strict_record_selector_5:
+	mov	rcx,qword ptr [rax+rcx]
+	mov	qword ptr 16[rdx],rcx
+copy_strict_record_selector_6:
+
+	mov	rcx,qword ptr ((-8)-8)[r11]
+	mov	qword ptr [rdx],rcx
+	att_jmp	in_hnf_2
+
+copy_arity_0_node2_:
+	att_jl	copy_selector_2
+
+	mov	(-24)[rsi],rcx 
+	sub	rsi,24
+	mov	[rbp],rsi 
+	lea	rax,1[rsi]
+
+	add	rbp,8
+	mov	[rdx],rax 
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_string_or_array_2:
+	mov	rcx,rdx 
+	jne	copy_array_2
+	mov	rax,rcx 
+
+	sub	rax,heap_p1[rip]
+	cmp	rax,semi_space_size[rip]
+	jae	copy_string_constant
+
+	mov	rdx,8[rcx]
+	add	rbp,8
+
+	add	rdx,7
+	push	rbx
+
+	mov	rax,rdx 
+	and	rdx,-8
+	
+	shr	rax,3
+	sub	rsi,rdx 
+
+	mov	rbx,[rcx]
+	add	rcx,8
+
+	mov	(-16)[rsi],rbx 
+	sub	rsi,16
+
+	mov	(-8)[rbp],rsi 
+	lea	rdx,1[rsi]
+	
+	mov	(-8)[rcx],rdx 
+	lea	rdx,8[rsi]
+
+cp_s_arg_lp2:
+	mov	rbx,[rcx]
+	add	rcx,8
+
+	mov	[rdx],rbx 
+	add	rdx,8
+
+	sub	rax,1
+	att_jge	cp_s_arg_lp2
+
+	pop	rbx 
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_string_constant:
+	mov	qword ptr [rbp],rdx 
+	add	rbp,8
+
+	sub	rbx,1
+	att_jae	copy_lp2
+	ret
+
+copy_array_2:
+	push	rbx 
+
+	mov	rax,qword ptr 16[rcx]
+	test	rax,rax 
+	je	copy_array_a2
+
+	movzx 	rbx,word ptr (-2)[rax]
+
+	test	rbx,rbx 
+	je	copy_strict_basic_array_2
+	
+	sub	rbx,256
+	imul	rbx,qword ptr 8[rcx]
+	jmp	copy_array_a3
+
+copy_array_a2:
+	mov	rbx,qword ptr 8[rcx]
+copy_array_a3:
+	mov	rdx,rdi 
+	lea	rdi,24[rdi+rbx*8]
+
+	mov	qword ptr [rbp],rdx 
+	mov	rax,qword ptr [rcx]
+
+	add	rbp,8
+	mov	qword ptr [rdx],rax 
+
+	lea	rax,1[rdx]
+	add	rdx,8
+
+	mov	qword ptr [rcx],rax 
+	add	rcx,8
+
+	lea	rax,1[rbx]
+	att_jmp	cp_s_arg_lp2
+
+copy_strict_basic_array_2:
+	mov	rbx,qword ptr 8[rcx]
+
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jle	copy_int_or_real_array_2
+
+	lea	r9,BOOL+2[rip]
+	cmp	rax,r9
+	je	copy_bool_array_2
+
+copy_int32_or_real32_array_2:
+	add	rbx,1
+	shr	rbx,1
+
+copy_int_or_real_array_2:
+	shl	rbx,3
+	lea	rdx,(-24)[rsi]
+
+	sub	rdx,rbx 
+	mov	rax,qword ptr [rcx]
+
+	shr	rbx,3
+	mov	qword ptr [rbp],rdx 
+
+	add	rbp,8
+	mov	rsi,rdx 
+	
+	mov	qword ptr [rdx],rax 
+	lea	rax,1[rdx]
+
+	add	rdx,8
+	mov	qword ptr [rcx],rax 
+
+	add	rcx,8
+	lea	rax,1[rbx]
+	att_jmp	cp_s_arg_lp2
+
+copy_bool_array_2:
+	add	rbx,7
+	shr	rbx,3
+	att_jmp	copy_int_or_real_array_2
+
+end_copy1:
+	mov	heap_end_after_gc[rip],rsi 
+
+	lea	rcx,finalizer_list[rip]
+	lea	rdx,free_finalizer_list[rip]
+	mov	rbp,qword ptr finalizer_list[rip]
+
+determine_free_finalizers_after_copy:
+	mov	rax,qword ptr [rbp]
+	test	al,1
+	je	finalizer_not_used_after_copy
+
+	mov	rbp,qword ptr 8[rbp]
+	sub	rax,1
+	mov	qword ptr [rcx],rax 
+	lea	rcx,8[rax]
+	att_jmp	determine_free_finalizers_after_copy
+
+finalizer_not_used_after_copy:
+	lea	r9,__Nil-8[rip]
+	cmp	rbp,r9
+	je	end_finalizers_after_copy
+
+	mov	qword ptr [rdx],rbp 
+	lea	rdx,8[rbp]
+	mov	rbp,qword ptr 8[rbp]
+	att_jmp	determine_free_finalizers_after_copy	
+
+end_finalizers_after_copy:
+	mov	qword ptr [rcx],rbp 
+	mov	qword ptr [rdx],rbp 
diff --git a/macho64/afileIO3.s b/macho64/afileIO3.s
new file mode 100644
index 0000000..64a270e
--- /dev/null
+++ b/macho64/afileIO3.s
@@ -0,0 +1,915 @@
+
+	.intel_syntax noprefix
+
+	.macro att_jmp
+	.att_syntax
+	jmp	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_call
+	.att_syntax
+	call	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_ja
+	.att_syntax
+	ja	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jnc
+	.att_syntax
+	jnc	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jns
+	.att_syntax
+	jns	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.data
+	.align	3
+
+freadstring_error:
+	.ascii	"Error in freadsubstring parameters."
+	.byte	10,0
+	.byte	0,0,0
+fwritestring_error:
+	.ascii	"Error in fwritesubstring parameters."
+	.byte	10,0
+	.byte	0,0
+
+	.text
+
+	.globl	stdioF
+	.globl	stderrF
+	.globl	openF
+	.globl	closeF
+	.globl	reopenF
+	.globl	readFC
+	.globl	readFI
+	.globl	readFR
+	.globl	readFS
+	.globl	readFString
+	.globl	readLineF
+	.globl	writeFC
+	.globl	writeFI
+	.globl	writeFR
+	.globl	writeFS
+	.globl	writeFString
+	.globl	endF
+	.globl	errorF
+	.globl	positionF
+	.globl	seekF
+	.globl	shareF
+.if 0
+	.globl	flushF
+.endif
+	.globl	openSF
+	.globl	readSFC
+	.globl	readSFI
+	.globl	readSFR
+	.globl	readSFS
+	.globl	readLineSF
+	.globl	endSF
+	.globl	positionSF
+	.globl	seekSF
+
+	.globl	_open_file
+	.globl	_open_stdio
+	.globl	_open_stderr
+	.globl	_re_open_file
+	.globl	_close_file
+	.globl	_file_read_char
+	.globl	_file_read_int
+	.globl	_file_read_real
+	.globl	_file_read_characters
+	.globl	_file_read_line
+	.globl	_file_write_char
+	.globl	_file_write_int
+	.globl	_file_write_real
+	.globl	_file_write_characters
+	.globl	_file_end
+	.globl	_file_error
+	.globl	_file_position
+	.globl	_file_seek
+	.globl	_file_share
+.if 0
+	.globl	flush_file_buffer
+.endif
+	.globl	_open_s_file
+	.globl	_file_read_s_char
+	.globl	_file_read_s_int
+	.globl	_file_read_s_real
+	.globl	_file_read_s_string
+	.globl	_file_read_s_line
+	.globl	_file_s_end
+	.globl	_file_s_position
+	.globl	_file_s_seek
+
+	.globl	collect_0
+	.globl	collect_1
+
+	.globl	print_error
+
+	.globl	__STRING__
+
+stdioF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	att_call	_open_stdio
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rbx,rax
+	mov	rax,-1
+	ret
+
+stderrF:	
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	att_call	_open_stderr
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rbx,rax
+	mov	rax,-1
+	ret
+
+openF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	r14,rdi
+	lea	rdi,8[rcx]
+	att_call	_open_file
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	xor	r10,r10
+	test	rax,rax
+	setns	r10b
+	mov	rbx,rax
+	mov	rax,-1
+	ret
+
+closeF:	
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_close_file
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	ret
+
+reopenF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	rbx,r10
+	mov	r14,rdi
+	mov	rdi,r10
+	att_call	_re_open_file
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10d,eax	
+	mov	rax,-1
+	ret
+
+readFC:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_char
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	cmp	eax,-1
+	je	readFC_eof
+
+	mov	r10,rax
+	mov	rax,-1
+	mov	r11,1
+	ret
+
+readFC_eof:
+	xor	r10,r10
+	mov	rax,-1
+	xor	r11,r11
+	ret
+
+readFI:
+	mov	rbp,rsp
+	sub	rsp,8
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,[rsp]
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_int
+	mov	r10,[rsp]
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r11,rax
+	mov	rax,-1
+	ret
+
+readFR:
+	mov	rbp,rsp
+	sub	rsp,8
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,[rsp]
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_real
+
+	movlpd	xmm0,[rsp]	
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1	
+	ret
+
+readFString:
+	mov	rbp,8[rcx]
+	cmp	r11,rbp
+	jae	readFString_error
+
+	sub	rbp,r11
+	cmp	r10,rbp
+	att_ja	readFString_error
+
+	push	rcx
+
+	push	r11
+
+	lea	rdx,16[rcx+r11]
+	mov	rbp,rsp
+	or	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbp
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_characters
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	add	rsp,8
+	pop	rcx
+	
+	mov	r10,rax
+	mov	rax,-1
+	ret
+
+readFString_error:
+	lea	rbp,freadstring_error[rip]
+	att_jmp	print_error
+
+readFS:	lea	rbp,16+7[rax]
+	shr	rbp,3
+	sub	r15,rbp
+	jb	readFS_gc
+readFS_r_gc:
+	add	r15,rbp
+
+	lea	rbx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rbx
+	mov	8[rdi],rax
+	mov	rbx,r10
+
+	lea	rdx,16[rdi]
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,8[rdi]
+	mov	r14,rdi
+	mov	rdi,r10
+	att_call	_file_read_characters
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+readFS_end:
+	add	rax,16+7
+	mov	rcx,rdi
+	shr	rax,3
+	sub	r15,rax
+	lea	rdi,[rdi+rax*8]
+	mov	rax,-1
+	ret
+
+readFS_gc:	push	rbp
+	att_call	collect_0
+	pop	rbp
+	att_jmp	readFS_r_gc
+
+readLineF:
+	cmp	r15,32+2
+	jb	readLineF_gc
+
+readLineF_r_gc:
+	lea	rdx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rdx
+
+	lea	rdx,16[rdi]
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,-16[r15*8]
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_line
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	8[rdi],rax
+
+	test	rax,rax
+	att_jns	readFS_end
+
+	lea	rax,-16[r15*8]
+	mov	r12,rdi
+	mov	8[rdi],rax
+	add	rdi,16
+
+readLineF_lp:
+	add	rdi,rax
+
+	mov	r13,8[r12]
+	mov	rcx,r12
+	shr	r13,3
+	xor	r15,r15
+	add	r13,2+32
+	sub	r15,r13
+	
+	att_call	collect_1
+
+	add	r15,r13
+	mov	rax,8[rcx]
+	lea	rdx,16[rcx]
+	lea	rcx,7[rax]
+	shr	rcx,3
+	sub	r15,2
+	sub	r15,rcx
+
+	mov	r12,rdi
+
+	lea	rbp,__STRING__+2[rip]
+	mov	qword ptr [rdi],rbp
+
+	mov	8[rdi],rax
+	add	rdi,16
+	jmp	st_copy_string1
+
+copy_st_lp1:
+	mov	rbp,[rdx]
+	add	rdx,8
+	mov	[rdi],rbp
+	add	rdi,8
+st_copy_string1:
+	sub	rcx,1
+	att_jnc	copy_st_lp1
+
+	mov	rdx,rdi
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,[r15*8]
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_line
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	test	rax,rax
+	js	readLineF_again
+
+	add	8[r12],rax
+	add	rax,7
+
+	mov	rcx,r12
+
+	shr	rax,3
+	sub	r15,rax
+	lea	rdi,[rdi+rax*8]
+
+	mov	rax,-1
+	ret
+
+readLineF_gc:
+	sub	r15,32+2
+	att_call	collect_0
+	add	r15,32+2
+	att_jmp	readLineF_r_gc
+
+readLineF_again:
+	mov	rcx,8[r12]
+	lea	rax,[r15*8]
+	add	rcx,rax
+	mov	8[r12],rcx
+	att_jmp	readLineF_lp
+
+writeFC:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbx
+	mov	r14,rdi
+	mov	rdi,r10
+	att_call	_file_write_char
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rax,-1
+	ret
+
+writeFI:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbx
+	mov	r14,rdi
+	mov	rdi,r10
+	att_call	_file_write_int
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rax,-1
+	ret
+
+writeFR:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+ 	att_call	_file_write_real
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rax,-1
+	ret
+
+writeFS:
+	mov	rdx,rbx
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,8[rcx]
+	mov	r14,rdi
+	lea	rdi,16[rcx]
+	att_call	_file_write_characters
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rax,-1
+	ret
+
+writeFString:
+	mov	rbp,8[rcx]
+	cmp	r11,rbp
+	jae	writeFString_error
+
+	sub	rbp,r11
+	cmp	r10,rbp
+	att_ja	writeFString_error
+
+	mov	rdx,rbx
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,r10
+	mov	r14,rdi
+	lea	rdi,16[rcx+r11]
+	att_call	_file_write_characters
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	rax,-1
+
+	ret
+
+writeFString_error:
+	lea	rbp,fwritestring_error[rip]
+	att_jmp	print_error
+
+endF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_end
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1
+	ret
+
+errorF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_error
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1
+	ret
+
+positionF:
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rbp,rsp
+	mov	rdi,rbx
+	att_call	_file_position
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1
+	ret
+
+seekF:
+	mov	rdx,rax
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbx
+	mov	rbx,r11
+	mov	r14,rdi
+	mov	rdi,r11
+	att_call	_file_seek
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1
+	ret
+
+shareF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_share
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	
+	mov	rax,-1
+	ret
+.if 0
+flushF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rbx
+	call	flush_file_buffer
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax
+	mov	rax,-1
+	ret
+.endif
+
+openSF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	r14,rdi
+	lea	rdi,8[rcx]
+	att_call	_open_s_file
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	xor	r10,r10
+	test	rax,rax
+	setns	r10b
+
+	mov	rbx,rax
+	xor	rax,rax
+	ret
+
+readSFC:
+	push	rax
+
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbp
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_s_char
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	cmp	rax,-1
+	je	readSFC_eof
+
+	pop	rax
+	mov	r10,rax
+	mov	r11,1
+	ret
+
+readSFC_eof:
+	pop	rax
+	xor	r10,r10
+	xor	r11,r11
+	ret
+
+readSFI:
+	push	rax
+	mov	rdx,rsp
+	sub	rsp,8
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbp
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_s_int
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	pop	r10
+	mov	r11,rax
+	pop	rax
+	
+	ret
+
+readSFR:
+	push	rax
+	mov	rdx,rsp
+	sub	rsp,8
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rbp
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_s_real
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	movlpd	xmm0,[rsp]
+	mov	r10,rax
+	add	rsp,8
+	pop	rax
+
+	ret
+
+readSFS:
+	lea	rbp,16+7[rax]
+	shr	rbp,3
+	sub	r15,rbp
+	jb	readSFS_gc
+readSFS_r_gc:
+	add	r15,rbp
+
+	lea	rcx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rcx
+
+	push	rbx
+
+	mov	rbx,r10
+	
+	mov	rcx,rsp
+	lea	rdx,8[rdi]
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	r14,rdi
+	mov	rdi,r10
+	att_call	_file_read_s_string
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+readSFS_end:
+	add	rax,16+7
+	mov	rcx,rdi
+	shr	rax,3
+	sub	r15,rax
+	lea	rdi,[rdi+rax*8]
+	pop	rax
+	ret
+
+readSFS_gc:	push	rbp
+	att_call	collect_0
+	pop	rbp
+	att_jmp	readSFS_r_gc
+
+readLineSF:
+	cmp	r15,32+2
+	ja	readLineSF_gc
+
+readLineSF_r_gc:
+	push	rax
+
+	lea	rcx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rcx
+
+	mov	rcx,rsp
+	lea	rdx,16[rdi]
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	lea	rsi,-16[r15*8]
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_read_s_line
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	8[rdi],rax
+
+	test	rax,rax
+	att_jns	readSFS_end
+
+	lea	rax,-16[r15*8]
+	mov	r12,rdi
+	mov	8[rdi],rax
+	add	rdi,16
+
+readLineSF_lp:
+	add	rdi,rax
+
+	mov	r13,8[r12]
+	mov	rcx,r12
+	shr	r13,3
+	xor	r15,r15
+	add	r13,2+32
+	sub	r15,r13
+
+	att_call	collect_1
+
+	add	r15,r13
+	mov	rax,8[rcx]
+	lea	rdx,16[rcx]
+	lea	rcx,7[rax]
+	shr	rcx,3
+	sub	r15,2
+	sub	r15,rcx
+
+	mov	r12,rdi
+
+	lea	rbp,__STRING__+2[rip]
+	mov	qword ptr [rdi],rbp
+
+	mov	8[rdi],rax
+	add	rdi,16
+	jmp	st_copy_string2
+
+copy_st_lp2:
+	mov	rbp,[rdx]
+	add	rdx,8
+	mov	[rdi],rbp
+	add	rdi,8
+st_copy_string2:
+	sub	rcx,1
+	att_jnc	copy_st_lp2
+
+	mov	r9,rsp
+	mov	r8,rdi
+	lea	rdx,[r15*8]
+	mov	rcx,rbx
+
+	mov	rbp,rsp
+	or	rsp,8
+	sub	rsp,40
+	att_call	_file_read_s_line
+	mov	rsp,rbp
+
+	test	rax,rax
+	js	readLineSF_again
+
+	add	8[r12],rax
+	add	rax,7
+
+	mov	rcx,r12
+
+	shr	rax,3
+	sub	r15,rax
+	lea	rdi,[rdi+rax*8]
+
+	pop	rax
+	ret
+
+readLineSF_gc:
+	sub	r15,32+2
+	att_call	collect_0
+	add	r15,32+2
+	att_jmp	readLineSF_r_gc
+
+readLineSF_again:
+	mov	rcx,8[r12]
+	lea	rax,[r15*8]
+	add	rcx,rax
+	mov	8[r12],rcx
+	att_jmp	readLineSF_lp
+
+endSF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_s_end
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	ret
+
+positionSF:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,rax
+	mov	r14,rdi
+	mov	rdi,rbx
+	att_call	_file_s_position
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	ret
+
+seekSF:
+	push	rbx
+	mov	rcx,rsp
+	mov	rdx,rax
+
+	mov	rbx,r11
+
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	rsi,r10
+	mov	r14,rdi
+	mov	rdi,r11
+	att_call	_file_s_seek
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+
+	mov	r10,rax	
+	pop	rax
+
+	ret
diff --git a/macho64/amark.s b/macho64/amark.s
new file mode 100644
index 0000000..bb0d9dc
--- /dev/null
+++ b/macho64/amark.s
@@ -0,0 +1,1933 @@
+
+	mov	rax,qword ptr heap_size_65[rip]
+	xor	rbx,rbx 
+	
+	mov	qword ptr n_marked_words[rip],rbx
+	shl	rax,6
+
+	mov	qword ptr lazy_array_list[rip],rbx 
+	mov	qword ptr heap_size_64_65[rip],rax 
+	
+	lea	rsi,(-4000)[rsp]
+
+	mov	rax,qword ptr caf_list[rip]
+
+	mov	qword ptr end_stack[rip],rsi 
+
+	mov	r10,neg_heap_p3[rip]
+	mov	r11,heap_size_64_65[rip]
+	mov	r13,qword ptr end_stack[rip]
+	mov	r14,0
+
+	test	rax,rax 
+	je	_end_mark_cafs
+
+_mark_cafs_lp:
+	mov	rbx,qword ptr [rax]
+	mov	rbp,qword ptr (-8)[rax]
+
+	push	rbp
+	lea	rbp,8[rax]
+	lea	r12,8[rax+rbx*8]
+
+	call	_mark_stack_nodes
+
+	pop	rax 
+	test	rax,rax 
+	att_jne	_mark_cafs_lp
+
+_end_mark_cafs:
+	mov	rsi,qword ptr stack_top[rip]
+	mov	rbp,qword ptr stack_p[rip]
+
+	mov	r12,rsi 
+	att_call	_mark_stack_nodes
+
+continue_mark_after_pmark:
+	mov	qword ptr n_marked_words[rip],r14
+
+	mov	rcx,qword ptr lazy_array_list[rip]
+
+	test	rcx,rcx 
+	je	end_restore_arrays
+
+restore_arrays:
+	mov	rbx ,qword ptr [rcx]
+	lea	r9,__ARRAY__+2[rip]
+	mov	qword ptr [rcx],r9
+
+	cmp	rbx,1
+	je	restore_array_size_1
+
+	lea	rdx,[rcx+rbx*8]
+	mov	rax,qword ptr 16[rdx]
+	test	rax,rax 
+	je	restore_lazy_array
+
+	mov	rbp,rax 
+	push	rdx 
+
+	xor	rdx,rdx 
+	mov	rax,rbx 
+	movzx	rbx,word ptr (-2+2)[rbp]
+
+	div	rbx 
+	mov	rbx,rax 
+
+	pop	rdx
+	mov	rax,rbp 
+
+restore_lazy_array:
+	mov	rdi,qword ptr 16[rcx]
+	mov	rbp,qword ptr 8[rcx]
+	mov	qword ptr 8[rcx],rbx 
+	mov	rsi,qword ptr 8[rdx]
+	mov	qword ptr 16[rcx],rax 
+	mov	qword ptr 8[rdx],rbp 
+	mov	qword ptr 16[rdx],rdi 
+
+	test	rax,rax
+	je	no_reorder_array
+
+	movzx	rdx,word ptr (-2)[rax]
+	sub	rdx,256
+	movzx	rbp,word ptr (-2+2)[rax]
+	cmp	rbp,rdx 
+	att_je	no_reorder_array
+
+	add	rcx,24
+	imul	rbx,rdx 
+	mov	rax,rdx 
+	lea	rdx,[rcx+rbx*8]
+	mov	rbx,rbp 
+	sub	rax,rbp 
+
+	att_call	reorder
+
+no_reorder_array:
+	mov	rcx,rsi 
+	test	rcx,rcx 
+	att_jne	restore_arrays
+
+	att_jmp	end_restore_arrays
+
+restore_array_size_1:
+	mov	rbp,qword ptr 8[rcx]
+	mov	rdx,qword ptr 16[rcx]
+	mov	qword ptr 8[rcx],rbx 
+	mov	rax,qword ptr 24[rcx]
+	mov	qword ptr 24[rcx],rbp 
+	mov	qword ptr 16[rcx],rax 
+
+	mov	rcx,rdx 
+	test	rcx,rcx 
+	att_jne	restore_arrays
+
+end_restore_arrays:
+	mov	rdi,qword ptr heap_vector[rip]
+	lea	rcx,finalizer_list[rip]
+	lea	rdx,free_finalizer_list[rip]
+
+	mov	rbp,qword ptr [rcx]
+determine_free_finalizers_after_mark:
+	lea	r9,__Nil-8[rip]
+	cmp	rbp,r9
+	je	end_finalizers_after_mark
+
+	lea	rax,[r10+rbp]
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rax]
+	test	esi,dword ptr [rdi+rbx*4]
+	je	finalizer_not_used_after_mark
+
+	lea	rcx,8[rbp]
+	mov	rbp,qword ptr 8[rbp]
+	att_jmp	determine_free_finalizers_after_mark
+
+finalizer_not_used_after_mark:
+	mov	qword ptr [rdx],rbp 
+	lea	rdx,8[rbp]
+
+	mov	rbp,qword ptr 8[rbp]
+	mov	qword ptr [rcx],rbp 
+	att_jmp	determine_free_finalizers_after_mark
+
+end_finalizers_after_mark:
+	mov	qword ptr [rdx],rbp 
+
+	att_call	add_garbage_collect_time
+
+	mov	rax,qword ptr bit_vector_size[rip]
+
+	mov	rdi,qword ptr n_allocated_words[rip]
+	add	rdi,qword ptr n_marked_words[rip]
+	shl	rdi,3
+
+	mov	rsi,rax
+	shl	rsi,3
+
+	push	rdx 
+	push	rax 
+
+	mov	rax,rdi 
+	mul	qword ptr _heap_size_multiple[rip]
+	shrd	rax,rdx,8
+	shr	rdx,8
+
+	mov	rbx,rax 
+	test	rdx,rdx 
+	
+	pop	rax 
+	pop	rdx 
+	
+	je	not_largest_heap
+
+	mov	rbx,qword ptr heap_size_65[rip]
+	shl	rbx,6
+
+not_largest_heap:
+	cmp	rbx,rsi 
+	jbe	no_larger_heap
+	
+	mov	rsi,qword ptr heap_size_65[rip]
+	shl	rsi,6
+	cmp	rbx,rsi
+	jbe	not_larger_than_heap
+	mov	rbx,rsi 
+not_larger_than_heap:
+	mov	rax,rbx 
+	shr	rax,3
+	mov	qword ptr bit_vector_size[rip],rax
+no_larger_heap:
+
+	mov	rbp,rax
+
+	mov	rdi,qword ptr heap_vector[rip]
+
+	shr	rbp,5
+
+	test	al,31
+	je	no_extra_word
+
+	mov	dword ptr [rdi+rbp*4],0
+
+no_extra_word:
+	sub	rax,qword ptr n_marked_words[rip]
+	shl	rax,3
+	mov	qword ptr n_last_heap_free_bytes[rip],rax
+
+	mov	rax,qword ptr n_marked_words[rip]
+	shl	rax,3
+	add	qword ptr total_gc_bytes[rip],rax
+
+	test	qword ptr _flags[rip],2
+	je	_no_heap_use_message2
+
+	mov	r12,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+
+	lea	rdi,marked_gc_string_1[rip]
+ .else
+	sub	rsp,32
+
+	lea	rcx,marked_gc_string_1
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rdi,qword ptr n_marked_words[rip]
+	shl	rdi,3
+ .else
+	mov	rcx,qword ptr n_marked_words
+	shl	rcx,3
+ .endif
+	att_call	_ew_print_int
+
+ .if LINUX
+	lea	rdi,heap_use_after_gc_string_2[rip]
+ .else
+	lea	rcx,heap_use_after_gc_string_2
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,r12
+
+_no_heap_use_message2:
+	att_call	call_finalizers
+
+	mov	rsi,qword ptr n_allocated_words[rip]
+	xor	rbx,rbx 
+
+	mov	rcx,rdi 
+	mov	qword ptr n_free_words_after_mark[rip],rbx
+
+_scan_bits:
+	cmp	ebx,dword ptr [rcx]
+	je	_zero_bits
+	mov	dword ptr [rcx],ebx 
+	add	rcx,4
+	sub	rbp,1
+	att_jne	_scan_bits
+
+	jmp	_end_scan
+
+_zero_bits:
+	lea	rdx,4[rcx]
+	add	rcx,4
+	sub	rbp,1
+	jne	_skip_zero_bits_lp1
+	jmp	_end_bits
+
+_skip_zero_bits_lp:
+	test	rax,rax 
+	jne	_end_zero_bits
+_skip_zero_bits_lp1:
+	mov	eax,dword ptr [rcx]
+	add	rcx,4
+	sub	rbp,1
+	att_jne	_skip_zero_bits_lp
+
+	test	rax,rax
+	att_je	_end_bits
+	mov	rax,rcx
+	mov	dword ptr (-4)[rcx],ebx
+	sub	rax,rdx
+	jmp	_end_bits2	
+
+_end_zero_bits:
+	mov	rax,rcx 
+	sub	rax,rdx 
+	shl	rax,3
+	add	qword ptr n_free_words_after_mark[rip],rax
+	mov	dword ptr (-4)[rcx],ebx 
+
+	cmp	rax,rsi
+	att_jb	_scan_bits
+
+_found_free_memory:
+	mov	qword ptr bit_counter[rip],rbp
+	mov	qword ptr bit_vector_p[rip],rcx 
+
+	lea	rbx,(-4)[rdx]
+	sub	rbx,rdi 
+	shl	rbx,6
+	mov	rdi,qword ptr heap_p3[rip]
+	add	rdi,rbx 
+
+	mov	r15,rax
+	lea	rbx,[rdi+rax*8]
+
+	sub	r15,rsi
+	mov	rsi,qword ptr stack_top[rip]
+
+	mov	qword ptr heap_end_after_gc[rip],rbx
+
+	att_jmp	restore_registers_after_gc_and_return
+
+_end_bits:
+	mov	rax,rcx 
+	sub	rax,rdx 
+	add	rax,4
+_end_bits2:
+	shl	rax,3
+	add	qword ptr n_free_words_after_mark[rip],rax
+	cmp	rax,rsi 
+	att_jae	_found_free_memory
+
+_end_scan:
+	mov	qword ptr bit_counter[rip],rbp
+	att_jmp	compact_gc
+
+/* %rbp : pointer to stack element */
+/* %rdi : heap_vector */
+/* %rax ,%rbx ,%rcx ,%rdx ,%rsi : free */
+
+_mark_stack_nodes:
+	cmp	rbp,r12
+	je	_end_mark_nodes
+_mark_stack_nodes_:
+	mov	rcx,qword ptr [rbp]
+
+	add	rbp,8
+	lea	rdx,[r10+rcx]
+
+	cmp	rdx,r11
+	att_jnc	_mark_stack_nodes
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	_mark_stack_nodes
+
+	push	rbp
+	push	0
+	jmp	_mark_arguments
+
+_mark_hnf_2:
+	cmp	rsi,0x20000000
+	jbe	fits_in_word_6
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_6:
+	add	r14,3
+
+_mark_record_2_c:
+	mov	rbx,qword ptr 8[rcx]
+	push	rbx 
+
+	cmp	rsp,r13
+	jb	__mark_using_reversal
+
+_mark_node2:
+_shared_argument_part:
+	mov	rcx,qword ptr [rcx]
+
+_mark_node:
+	lea	rdx,[r10+rcx]
+	cmp	rdx,r11
+	jnc	_mark_next_node
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	_mark_next_node
+
+_mark_arguments:
+	mov	rax,qword ptr [rcx]
+	test	rax,2
+	je	_mark_lazy_node
+	
+	movzx	rbp,word ptr (-2)[rax]
+
+	test	rbp,rbp
+	je	_mark_hnf_0
+
+	or	dword ptr [rdi+rbx*4],esi 
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	_mark_record
+
+	sub	rbp,2
+	att_je	_mark_hnf_2
+	jb	_mark_hnf_1
+
+_mark_hnf_3:
+	mov	rdx,qword ptr 8[rcx]
+
+	cmp	rsi,0x20000000
+	jbe	fits_in_word_1
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_1:	
+
+	add	r14,3
+	lea	rax,[r10+rdx]
+	mov	rbx,rax
+	
+	and	rax,31*8
+	shr	rbx,8
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rax]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	_shared_argument_part
+
+_no_shared_argument_part:
+	or	dword ptr [rdi+rbx*4],esi 
+	add	rbp,1
+
+	add	r14,rbp 
+	lea	rax,[rax+rbp*8]
+	lea	rdx,(-8)[rdx+rbp*8]
+
+	cmp	rax,32*8
+	jbe	fits_in_word_2
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_2:
+
+	mov	rbx,qword ptr [rdx]
+	sub	rbp,2
+	push	rbx 
+
+_push_hnf_args:
+	mov	rbx,qword ptr (-8)[rdx]
+	sub	rdx,8
+	push	rbx 
+	sub	rbp,1
+	att_jge	_push_hnf_args
+
+	cmp	rsp,r13
+	att_jae	_mark_node2
+
+	att_jmp	__mark_using_reversal
+
+_mark_hnf_1:
+	cmp	rsi,0x40000000
+	jbe	fits_in_word_4
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_4:
+	add	r14,2
+	mov	rcx,qword ptr [rcx]
+	att_jmp	_mark_node
+
+_mark_lazy_node_1:
+	add	rcx,8
+	or	dword ptr [rdi+rbx*4],esi 
+	cmp	rsi,0x20000000
+	jbe	fits_in_word_3
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_3:
+	add	r14,3
+
+	cmp	rbp,1
+	att_je	_mark_node2
+
+_mark_selector_node_1:
+	add	rbp,2
+	mov	rdx,qword ptr [rcx]
+	je	_mark_indirection_node
+
+	lea	rsi,[r10+rdx]
+	mov	rbx,rsi 
+
+	shr	rbx,8
+	and	rsi,31*8
+
+	add	rbp,1
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rsi]
+	jle	_mark_record_selector_node_1
+
+	test	esi,dword ptr [rdi+rbx*4]
+	jne	_mark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	_mark_node3
+
+	cmp	word ptr (-2)[rbp],2
+	jbe	_small_tuple_or_record
+
+_large_tuple_or_record:
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r8,bit_set_table2[rip]
+	mov	ebp,dword ptr [r8+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	_mark_node3
+
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+
+	cmp	rax,16
+	jl	_mark_tuple_selector_node_1
+	mov	rdx,r9
+	je	_mark_tuple_selector_node_2
+	mov	rcx,qword ptr (-24)[r9+rax]
+	mov	qword ptr [rbp],rcx
+	att_jmp	_mark_node
+
+_mark_tuple_selector_node_2:
+	mov	rcx,qword ptr [r9]
+	mov	qword ptr [rbp],rcx
+	att_jmp	_mark_node	
+
+_small_tuple_or_record:
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+_mark_tuple_selector_node_1:
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rbp],rcx	
+	att_jmp	_mark_node
+
+_mark_record_selector_node_1:
+	je	_mark_strict_record_selector_node_1
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	_mark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	_mark_node3
+
+	cmp	word ptr (-2)[rbp],258
+	att_jbe	_small_tuple_or_record
+
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r8,bit_set_table2[rip]
+	mov	ebp,dword ptr [r8+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	_mark_node3
+
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+
+	cmp	rax,16
+	jle	_mark_record_selector_node_2
+	mov	rdx,r9
+	sub	rax,24
+_mark_record_selector_node_2:
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rbp],rcx
+	att_jmp	_mark_node
+
+_mark_strict_record_selector_node_1:
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	_mark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	_mark_node3
+
+	cmp	word ptr (-2)[rbp],258
+	jbe	_select_from_small_record
+
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp 
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r8,bit_set_table2[rip]
+	mov	ebp,dword ptr [r8+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	_mark_node3
+
+_select_from_small_record:
+	movsxd	rbx,dword ptr (-8)[rax]
+	add	rax,rbx
+	sub	rcx,8
+
+	movzx	ebx,word ptr (4-8)[rax]
+	cmp	rbx,16
+	jle	_mark_strict_record_selector_node_2
+	mov	rbx,qword ptr (-24)[r9+rbx]
+	jmp	_mark_strict_record_selector_node_3
+_mark_strict_record_selector_node_2:
+	mov	rbx,qword ptr [rdx+rbx]
+_mark_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rbx
+
+	movzx	ebx,word ptr (6-8)[rax]
+	test	rbx,rbx
+	je	_mark_strict_record_selector_node_5
+	cmp	rbx,16
+	jle	_mark_strict_record_selector_node_4
+	mov	rdx,r9
+	sub	rbx,24
+_mark_strict_record_selector_node_4:
+	mov	rbx,qword ptr [rdx+rbx]
+	mov	qword ptr 16[rcx],rbx
+_mark_strict_record_selector_node_5:
+
+	mov	rax,qword ptr ((-8)-8)[rax]
+	mov	qword ptr [rcx],rax
+	att_jmp	_mark_next_node
+
+_mark_indirection_node:
+_mark_node3:
+	mov	rcx,rdx 
+	att_jmp	_mark_node
+
+_mark_next_node:
+	pop	rcx 
+	test	rcx,rcx 
+	att_jne	_mark_node
+
+	pop	rbp 
+	cmp	rbp,r12
+	att_jne	_mark_stack_nodes_
+
+_end_mark_nodes:
+	ret
+
+_mark_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp 
+	je	_mark_real_or_file
+
+	cmp	rbp,1
+	att_jle	_mark_lazy_node_1
+
+	cmp	rbp,256
+	jge	_mark_closure_with_unboxed_arguments
+	inc	rbp
+	or	dword ptr [rdi+rbx*4],esi 
+
+	add	r14,rbp 
+	lea	rdx,[rdx+rbp*8]
+	lea	rcx,[rcx+rbp*8]
+
+	cmp	rdx,32*8
+	jbe	fits_in_word_7
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_7:
+	sub	rbp,3
+_push_lazy_args:
+	mov	rbx,qword ptr (-8)[rcx]
+	sub	rcx,8
+	push	rbx
+	sub	rbp,1
+	att_jge	_push_lazy_args
+
+	sub	rcx,8
+
+	cmp	rsp,r13
+	att_jae	_mark_node2
+	
+	att_jmp	__mark_using_reversal
+
+_mark_closure_with_unboxed_arguments:
+	mov	rax,rbp 
+	and	rbp,255
+	sub	rbp,1
+	att_je	_mark_real_or_file
+
+	shr	rax,8
+	add	rbp,2
+
+	or	dword ptr [rdi+rbx*4],esi 
+	add	r14,rbp 
+	lea	rdx,[rdx+rbp*8]
+
+	sub	rbp,rax
+
+	cmp	rdx,32*8
+	jbe	fits_in_word_7_
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_7_:
+	sub	rbp,2
+	att_jl	_mark_next_node
+
+	lea	rcx,16[rcx+rbp*8]
+	att_jne	_push_lazy_args
+
+_mark_closure_with_one_boxed_argument:
+	mov	rcx,qword ptr (-8)[rcx]
+	att_jmp	_mark_node
+
+_mark_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jb	_mark_real_file_or_string
+
+	or	dword ptr [rdi+rbx*4],esi 
+
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+	ja	_mark_normal_hnf_0
+
+_mark_bool:
+	add	r14,2
+
+	cmp	rsi,0x40000000
+	att_jbe	_mark_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	_mark_next_node
+
+_mark_normal_hnf_0:
+	inc	r14
+	att_jmp	_mark_next_node
+
+_mark_real_file_or_string:
+	lea	r9,__STRING__+2[rip]
+	cmp	rax,r9
+	jbe	_mark_string_or_array
+
+_mark_real_or_file:
+	or	dword ptr [rdi+rbx*4],esi 
+	add	r14,3
+
+	cmp	rsi,0x20000000
+	att_jbe	_mark_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	_mark_next_node
+
+_mark_record:
+	sub	rbp,258
+	je	_mark_record_2
+	jl	_mark_record_1
+
+_mark_record_3:
+	add	r14,3
+
+	cmp	rsi,0x20000000
+	jbe	fits_in_word_13
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_13:
+	mov	rdx,qword ptr 8[rcx]
+
+	movzx	rbx,word ptr (-2+2)[rax]
+	lea	rsi,[r10+rdx]
+
+	mov	rax,rsi 
+	and	rsi,31*8
+
+	shr	rax,8
+	sub	rbx,1
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rsi]
+	jb	_mark_record_3_bb
+
+	test	edx,dword ptr [rdi+rax*4]
+	att_jne	_mark_node2
+
+	add	rbp,1
+	or	dword ptr [rdi+rax*4],edx 
+	add	r14,rbp 
+	lea	rsi,[rsi+rbp*8]
+
+	cmp	rsi,32*8
+	jbe	_push_record_arguments
+	or	dword ptr 4[rdi+rax*4],1
+_push_record_arguments:
+	mov	rdx,qword ptr 8[rcx]
+	mov	rbp,rbx 
+	shl	rbx,3
+	add	rdx,rbx 
+	sub	rbp,1
+	att_jge	_push_hnf_args
+
+	att_jmp	_mark_node2
+
+_mark_record_3_bb:
+	test	edx,dword ptr [rdi+rax*4]
+	att_jne	_mark_next_node
+
+	add	rbp,1
+	or	dword ptr [rdi+rax*4],edx 
+	add	r14,rbp 
+	lea	rsi,[rsi+rbp*8]
+	
+	cmp	rsi,32*8
+	att_jbe	_mark_next_node
+
+	or	dword ptr 4[rdi+rax*4],1
+	att_jmp	_mark_next_node
+
+_mark_record_2:
+	cmp	rsi,0x20000000
+	jbe	fits_in_word_12
+	or	dword ptr 4[rdi+rbx*4],1
+fits_in_word_12:
+	add	r14,3
+
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	_mark_record_2_c
+	att_je	_mark_node2
+	att_jmp	_mark_next_node
+
+_mark_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	_mark_hnf_1
+
+	att_jmp	_mark_bool
+
+_mark_string_or_array:
+	je	_mark_string_
+
+_mark_array:
+	mov	rbp,qword ptr 16[rcx]
+	test	rbp,rbp
+	je	_mark_lazy_array
+
+	movzx	rax,word ptr (-2)[rbp]
+
+	test	rax,rax 
+	je	_mark_strict_basic_array
+
+	movzx	rbp,word ptr (-2+2)[rbp]
+	test	rbp,rbp 
+	je	_mark_b_record_array
+
+	cmp	rsp,r13
+	jb	_mark_array_using_reversal
+
+	sub	rax,256
+	cmp	rax,rbp 
+	je	_mark_a_record_array
+
+_mark_ab_record_array:
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rbp,qword ptr 8[rcx]
+
+	imul	rax,rbp 
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+
+	cmp	rbx,rax 
+	jae	_end_set_ab_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	_last_ab_array_bits
+
+_mark_ab_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx
+	cmp	rbx,rax 
+	att_jb	_mark_ab_array_lp
+
+_last_ab_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+
+_end_set_ab_array_bits:
+	mov	rax,qword ptr 8[rcx]
+	mov	rdx,qword ptr 16[rcx]
+	movzx	rbx,word ptr (-2+2)[rdx]
+	movzx	rdx,word ptr (-2)[rdx]
+	shl	rbx,3
+	lea	rdx,(-2048)[rdx*8]
+	push	rbx
+	push	rdx
+	lea	rbp,24[rcx]
+	push	r12
+	jmp	_mark_ab_array_begin
+	
+_mark_ab_array:
+	mov	rbx,qword ptr 16[rsp]
+	push	rax
+	push	rbp
+	lea	r12,[rbp+rbx]
+
+	att_call	_mark_stack_nodes
+
+	mov	rbx,qword ptr (8+16)[rsp]
+	pop	rbp
+	pop	rax
+	add	rbp,rbx 
+_mark_ab_array_begin:
+	sub	rax,1
+	att_jnc	_mark_ab_array
+
+	pop	r12
+	add	rsp,16
+	att_jmp	_mark_next_node
+
+_mark_a_record_array:
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rbp,qword ptr 8[rcx]
+
+	imul	rax,rbp 
+	push	rax
+
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	jae	_end_set_a_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	_last_a_array_bits
+
+_mark_a_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx 
+	cmp	rbx,rax
+	att_jb	_mark_a_array_lp
+
+_last_a_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp
+
+_end_set_a_array_bits:
+	pop	rax 
+	lea	rbp,24[rcx]
+
+	push	r12
+	lea	r12,24[rcx+rax*8]
+
+	att_call	_mark_stack_nodes
+
+	pop	r12
+	att_jmp	_mark_next_node
+
+_mark_lazy_array:
+	cmp	rsp,r13
+	att_jb	_mark_array_using_reversal
+
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rax,qword ptr 8[rcx]
+
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	jae	_end_set_lazy_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	_last_lazy_array_bits
+
+_mark_lazy_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp
+	inc	rbx
+	cmp	rbx,rax
+	att_jb	_mark_lazy_array_lp
+
+_last_lazy_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp
+
+_end_set_lazy_array_bits:
+	mov	rax,qword ptr 8[rcx]
+	lea	rbp,24[rcx]
+
+	push	r12
+	lea	r12,24[rcx+rax*8]
+
+	att_call	_mark_stack_nodes
+
+	pop	r12
+	att_jmp	_mark_next_node
+
+_mark_array_using_reversal:
+	push	0
+	mov	rsi,1
+	jmp	__mark_node
+
+_mark_strict_basic_array:
+	mov	rax,qword ptr 8[rcx]
+	lea	r9,dINT+2[rip]
+	cmp	rbp,r9
+	jle	_mark_strict_int_or_real_array
+	lea	r9,BOOL+2[rip]
+	cmp	rbp,r9
+	je	_mark_strict_bool_array
+_mark_strict_int32_or_real32_array:
+	add	rax,6+1
+	shr	rax,1
+	jmp	_mark_basic_array_
+_mark_strict_int_or_real_array:
+	add	rax,3
+	att_jmp	_mark_basic_array_
+_mark_strict_bool_array:
+	add	rax,24+7
+	shr	rax,3
+	att_jmp	_mark_basic_array_
+
+_mark_b_record_array:
+	mov	rbp,qword ptr 8[rcx]
+	sub	rax,256
+	imul	rax,rbp 
+	add	rax,3
+	att_jmp	_mark_basic_array_
+
+_mark_string_:
+	mov	rax,qword ptr 8[rcx]
+	add	rax,16+7
+	shr	rax,3
+
+_mark_basic_array_:
+	or	dword ptr [rdi+rbx*4],esi 
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	att_jae	_mark_next_node
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	_last_string_bits
+
+_mark_string_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx
+	cmp	rbx,rax 
+	att_jb	_mark_string_lp
+
+_last_string_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+	att_jmp	_mark_next_node
+
+__end_mark_using_reversal:
+	pop	rdx 
+	test	rdx,rdx 
+	att_je	_mark_next_node
+	mov	qword ptr [rdx],rcx 
+	att_jmp	_mark_next_node
+
+__mark_using_reversal:
+	push	rcx 
+	mov	rsi,1
+	mov	rcx,qword ptr [rcx]
+	att_jmp	__mark_node
+
+__mark_arguments:
+	mov	rax,qword ptr [rcx]
+	test	al,2
+	je	__mark_lazy_node
+
+	movzx	rbp,word ptr (-2)[rax]
+	test	rbp,rbp 
+	je	__mark_hnf_0
+
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	__mark__record
+
+	sub	rbp,2
+	je	__mark_hnf_2
+	jb	__mark_hnf_1
+
+__mark_hnf_3:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+
+	or	dword ptr [rdi+rbx*4],edx
+
+	cmp	rdx,0x20000000
+
+	mov	rax,qword ptr 8[rcx]
+
+	jbe	fits__in__word__1
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word__1:
+	add	rax,r10
+
+	mov	rbx,rax 
+	and	rax,31*8
+
+	shr	rbx,8
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rax]
+	test	edx,dword ptr [rdi+rbx*4]
+	jne	__shared_argument_part
+
+__no_shared_argument_part:
+	or	dword ptr [rdi+rbx*4],edx 
+	mov	rdx,qword ptr 8[rcx]
+
+	add	rbp,1
+	mov	qword ptr 8[rcx],rsi 
+
+	add	r14,rbp 
+	add	rcx,8
+
+	shl	rbp,3
+	or	qword ptr [rdx],1
+
+	add	rax,rbp
+	add	rdx,rbp
+
+	cmp	rax,32*8
+	jbe	fits__in__word__2
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word__2:
+
+	mov	rbp ,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rcx 
+	lea	rsi,(-8)[rdx]
+	mov	rcx,rbp 
+	att_jmp	__mark_node
+
+__mark_hnf_1:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,2
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x40000000
+	att_jbe	__shared_argument_part
+	or	dword ptr 4[rdi+rbx*4],1
+__shared_argument_part:
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi
+	lea	rsi,2[rcx]
+	mov	rcx,rbp 
+	att_jmp	__mark_node
+
+__mark_no_selector_2:
+	pop	rbx
+__mark_no_selector_1:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	att_jbe	__shared_argument_part
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	__shared_argument_part
+
+__mark_lazy_node_1:
+	att_je	__mark_no_selector_1
+
+__mark_selector_node_1:
+	add	rbp,2
+	je	__mark_indirection_node
+
+	add	rbp,1
+
+	push	rbx
+	mov	rbp,qword ptr [rcx]
+	push	rax 
+	lea	rax,[r10+rbp] 
+
+	jle	__mark_record_selector_node_1
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx*4]
+	pop	rax 
+	att_jne	__mark_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	__mark_no_selector_2
+
+	cmp	word ptr (-2)[rbx],2
+	jbe	__small_tuple_or_record
+
+__large_tuple_or_record:
+	mov	r8,qword ptr 16[rbp]
+
+	add	r8,r10
+	mov	rbx,r8
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+	
+	mov	r9,qword ptr 16[rbp]
+
+	att_jne	__mark_no_selector_2
+
+	movsxd	rdx,dword ptr (-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+
+	cmp	rax,16
+	jl	__mark_tuple_selector_node_1
+	je	__mark_tuple_selector_node_2
+	mov	rcx,qword ptr (-24)[r9+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	__mark_node
+
+__mark_tuple_selector_node_2:
+	mov	rcx,qword ptr [r9]
+	mov	qword ptr [r8],rcx
+	att_jmp	__mark_node
+
+__small_tuple_or_record:
+	movsxd	rdx,dword ptr (-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx 
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+__mark_tuple_selector_node_1:
+	mov	rcx,qword ptr [rbp+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	__mark_node
+	att_jmp	__mark_node
+
+__mark_record_selector_node_1:
+	je	__mark_strict_record_selector_node_1
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx*4]
+	pop	rax
+	att_jne	__mark_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	__mark_no_selector_2
+
+	cmp	word ptr (-2)[rbx],258
+	jbe	__small_record
+
+	mov	r8,qword ptr 16[rbp]
+
+	add	r8,r10
+	mov	rbx,r8
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+
+	mov	r9,qword ptr 16[rbp]
+
+	att_jne	__mark_no_selector_2
+
+__small_record:
+	movsxd	rdx,dword ptr(-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx 
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+
+	cmp	rax,16
+	jle	__mark_record_selector_node_2
+	mov	rbp,r9
+	sub	rax,24
+__mark_record_selector_node_2:
+	mov	rcx,qword ptr [rbp+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	__mark_node
+
+__mark_strict_record_selector_node_1:
+	mov	rbx,rax
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx *4]
+	pop	rax 
+	att_jne	__mark_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	__mark_no_selector_2
+
+	cmp	word ptr (-2)[rbx],258
+	jle	__select_from_small_record
+
+	mov	r8,qword ptr 16[rbp]
+
+	add	r8,r10
+	mov	rbx,r8 
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+
+	mov	r9,qword ptr 16[rbp]
+
+	att_jne	__mark_no_selector_2
+	
+__select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rax,rbx
+	sub	rcx,8
+
+	movzx	ebx,word ptr (4-8)[rax]
+	cmp	rbx,16
+	jle	__mark_strict_record_selector_node_2
+	mov	rbx,qword ptr (-24)[r9+rbx]
+	jmp	__mark_strict_record_selector_node_3
+__mark_strict_record_selector_node_2:
+	mov	rbx,qword ptr [rbp+rbx]
+__mark_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rbx
+
+	movzx	ebx,word ptr (6-8)[rax]
+	test	rbx,rbx
+	je	__mark_strict_record_selector_node_5
+	cmp	rbx,16
+	jle	__mark_strict_record_selector_node_4
+	mov	rbp,r9
+	sub	rbx,24
+__mark_strict_record_selector_node_4:
+	mov	rbx,qword ptr [rbp+rbx]
+	mov	qword ptr 16[rcx],rbx
+__mark_strict_record_selector_node_5:
+	pop	rbx
+
+	mov	rax,qword ptr ((-8)-8)[rax]
+	mov	qword ptr [rcx],rax
+	att_jmp	__mark_node
+
+__mark_indirection_node:
+	mov	rcx,qword ptr [rcx]
+	att_jmp	__mark_node
+
+__mark_hnf_2:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx
+	cmp	rdx,0x20000000
+	jbe	fits__in__word__6
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word__6:
+
+__mark_record_2_c:
+	mov	rax,qword ptr [rcx]
+	mov	rbp,qword ptr 8[rcx]
+	or	rax,2
+	mov	qword ptr 8[rcx],rsi 
+	mov	qword ptr [rcx],rax 
+	lea	rsi,8[rcx]
+	mov	rcx,rbp
+
+__mark_node:
+	lea	rdx,[r10+rcx]
+	cmp	rdx,r11
+	jae	__mark_next_node
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rdx]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_je	__mark_arguments
+
+__mark_next_node:
+	test	rsi,3
+	jne	__mark_parent
+
+	mov	rbp,qword ptr (-8)[rsi]
+	mov	rdx,qword ptr [rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr (-8)[rsi],rdx 
+	sub	rsi,8
+
+	mov	rcx,rbp 
+	and	rbp,3
+	and	rcx,-4
+	or	rsi,rbp 
+	att_jmp	__mark_node
+
+__mark_parent:
+	mov	rbx,rsi 
+	and	rsi,-4
+	att_je	__end_mark_using_reversal
+
+	and	rbx,3
+	mov	rbp,qword ptr [rsi]
+	mov	qword ptr [rsi],rcx 
+
+	sub	rbx,1
+	je	__argument_part_parent
+	
+	lea	rcx,(-8)[rsi]
+	mov	rsi,rbp 
+	att_jmp	__mark_next_node
+
+__argument_part_parent:
+	and	rbp,-4
+	mov	rdx,rsi 
+	mov	rcx,qword ptr (-8)[rbp]
+	mov	rbx,qword ptr [rbp]
+	mov	qword ptr (-8)[rbp],rbx 
+	mov	qword ptr [rbp],rdx 
+	lea	rsi,(2-8)[rbp]
+	att_jmp	__mark_node
+
+__mark_lazy_node:
+	movsxd	rbp,dword ptr(-4)[rax]
+	test	rbp,rbp 
+	je	__mark_real_or_file
+
+	add	rcx,8
+	cmp	rbp,1
+	att_jle	__mark_lazy_node_1
+	cmp	rbp,256
+	jge	__mark_closure_with_unboxed_arguments
+
+	add	rbp,1
+	mov	rax,rdx 
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,rbp
+
+	lea	rax,[rax+rbp*8]
+	sub	rbp,2
+
+	or	dword ptr [rdi+rbx*4],edx 
+
+	cmp	rax,32*8
+	jbe	fits__in__word__7
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word__7:
+__mark_closure_with_unboxed_arguments__2:
+	lea	rdx,[rcx+rbp*8]
+	mov	rax,qword ptr [rcx]
+	or	rax,2
+	mov	qword ptr [rcx],rax 
+	mov	rcx,qword ptr [rdx]
+	mov	qword ptr [rdx],rsi 
+	mov	rsi,rdx 
+	att_jmp	__mark_node
+
+__mark_closure_with_unboxed_arguments:
+	mov	rax,rbp 
+	and	rbp,255
+
+	sub	rbp,1
+	je	__mark_closure_1_with_unboxed_argument
+	add	rbp,2
+
+	shr	rax,8
+	add	r14,rbp 
+
+	push	rcx 
+	lea	rcx,[rdx+rbp*8]
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	sub	rbp,rax 
+
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rcx,32*8
+	jbe	fits__in_word_7_
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in_word_7_:
+	pop	rcx 
+	sub	rbp,2
+	att_jg	__mark_closure_with_unboxed_arguments__2
+	att_je	__shared_argument_part
+	sub	rcx,8
+	att_jmp	__mark_next_node
+
+__mark_closure_1_with_unboxed_argument:
+	sub	rcx,8
+	att_jmp	__mark_real_or_file
+
+__mark_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jne	__no_int_3
+
+	mov	rbp,qword ptr 8[rcx]
+	cmp	rbp,33
+	jb	____small_int
+
+__mark_bool_or_small_string:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,2
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x40000000
+	att_jbe	__mark_next_node
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	__mark_next_node
+
+____small_int:
+	shl	rbp,4
+	lea	rcx,small_integers[rip]
+	add	rcx,rbp
+	att_jmp	__mark_next_node
+
+__no_int_3:
+	jb	__mark_real_file_or_string
+
+ 	lea	r9,CHAR+2[rip]
+ 	cmp	rax,r9
+ 	jne	__no_char_3
+
+	movzx	rbp,byte ptr 8[rcx]
+	shl	rbp,4
+	lea	rcx,static_characters[rip]
+	add	rcx,rbp
+	att_jmp	__mark_next_node
+
+__no_char_3:
+	att_jb	__mark_bool_or_small_string
+
+	lea	rcx,((-8)-2)[rax]
+	att_jmp	__mark_next_node
+
+__mark_real_file_or_string:
+	lea	r9,__STRING__+2[rip]
+	cmp	rax,r9
+	jbe	__mark_string_or_array
+
+__mark_real_or_file:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+
+	or	dword ptr [rdi+rbx*4],edx
+
+	cmp	rdx,0x20000000
+	att_jbe	__mark_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	__mark_next_node
+
+__mark__record:
+	sub	rbp,258
+	je	__mark_record_2
+	jl	__mark_record_1
+
+__mark_record_3:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	jbe	fits__in__word__13
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word__13:
+	movzx	rbx,word ptr (-2+2)[rax]
+
+	mov	rdx,qword ptr 8[rcx]
+	add	rdx,r10
+	mov	rax,rdx 
+	and	rdx,31*8
+	shr	rax,8
+
+	push	rsi 
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+	test	esi,dword ptr [rdi+rax*4]
+	jne	__shared_record_argument_part
+
+	add	rbp,1
+	or	dword ptr [rdi+rax *4],esi
+
+	lea	rdx,[rdx+rbp*8]
+	add	r14,rbp 
+
+	pop	rsi 
+
+	cmp	rdx,32*8
+	jbe	fits__in__word__14
+	or	dword ptr 4[rdi+rax*4],1
+fits__in__word__14:
+	sub	rbx,1
+	mov	rdx,qword ptr 8[rcx]
+	jl	__mark_record_3_bb
+	att_je	__shared_argument_part
+
+	mov	qword ptr 8[rcx],rsi 
+	add	rcx,8
+
+	sub	rbx,1
+	je	__mark_record_3_aab
+
+	lea	rsi,[rdx+rbx*8]
+	mov	rax,qword ptr [rdx]
+	or	rax,1
+	mov	rbp,qword ptr [rsi]
+	mov	qword ptr [rdx],rax 
+	mov	qword ptr [rsi],rcx 
+	mov	rcx,rbp 
+	att_jmp	__mark_node
+
+__mark_record_3_bb:
+	sub	rcx,8
+	att_jmp	__mark_next_node
+
+__mark_record_3_aab:
+	mov	rbp,qword ptr [rdx]
+	mov	qword ptr [rdx],rcx 
+	lea	rsi,1[rdx]
+	mov	rcx,rbp 
+	att_jmp	__mark_node
+
+__shared_record_argument_part:
+	mov	rdx,qword ptr 8[rcx]
+
+	pop	rsi
+
+	test	rbx,rbx 
+	att_jne	__shared_argument_part
+	sub	rcx,8
+	att_jmp	__mark_next_node
+
+__mark_record_2:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	jbe	fits__in__word_12
+	or	dword ptr 4[rdi+rbx*4],1
+fits__in__word_12:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	__mark_record_2_c
+	att_je	__shared_argument_part
+	sub	rcx,8
+	att_jmp	__mark_next_node
+
+__mark_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	__mark_hnf_1
+	sub	rcx,8
+	att_jmp	__mark_bool_or_small_string
+
+__mark_string_or_array:
+	je	__mark_string_
+
+__mark_array:
+	mov	rbp,qword ptr 16[rcx]
+	test	rbp,rbp 
+	je	__mark_lazy_array
+
+	movzx	rax,word ptr (-2)[rbp]
+	test	rax,rax 
+	je	__mark_strict_basic_array
+
+	movzx	rbp,word ptr (-2+2)[rbp]
+	test	rbp,rbp 
+	je	__mark_b_record_array
+
+	sub	rax,256
+	cmp	rax,rbp 
+	je	__mark_a_record_array
+
+__mark__ab__record__array:
+	push	rdx 
+	push	rbx 
+	mov	rbx,rbp 
+
+	mov	rbp,qword ptr 8[rcx]
+	add	rcx,16
+	push	rcx 
+
+	shl	rbp,3
+	mov	rdx,rax 
+	imul	rdx,rbp 
+
+	sub	rax,rbx 
+	add	rcx,8
+	add	rdx,rcx 
+
+	att_call	reorder
+	
+	pop	rcx 
+
+	xchg	rax,rbx 
+	mov	rbp,qword ptr (-8)[rcx]
+	imul	rax,rbp 
+	imul	rbx,rbp 
+	add	r14,rbx 
+	add	rbx,rax 
+
+	shl	rbx,3
+	lea	rbp,[r10+rcx]
+	add	rbp,rbx 
+
+	pop	rbx 
+	pop	rdx 
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	or	dword ptr [rdi+rbx*4],edx
+
+	lea	rdx,[rcx+rax*8]
+	jmp	__mark_r_array
+
+__mark_a_record_array:
+	imul	rax,qword ptr 8[rcx]
+	add	rcx,16
+	jmp	__mark_lr_array
+
+__mark_lazy_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+
+__mark_lr_array:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	mov	rbp,r10
+	or	dword ptr [rdi+rbx*4],edx 
+	lea	rdx,[rcx+rax*8]
+	add	rbp,rdx 
+__mark_r_array:
+	shr	rbp,8
+
+	cmp	rbx,rbp 
+	jae	__skip_mark_lazy_array_bits
+
+	inc	rbx 
+
+__mark_lazy_array_bits:
+	or	dword ptr [rdi+rbx*4],1
+	inc	rbx
+	cmp	rbx,rbp 
+	att_jbe	__mark_lazy_array_bits
+
+__skip_mark_lazy_array_bits:
+	add	r14,3
+	add	r14,rax 
+
+	cmp	rax,1
+	jbe	__mark_array_length_0_1
+
+	mov	rbp,qword ptr [rdx]
+	mov	rbx,qword ptr [rcx]
+	mov	qword ptr [rdx],rbx 
+	mov	qword ptr [rcx],rbp 
+	
+	mov	rbp,qword ptr (-8)[rdx]
+	sub	rdx,8
+	mov	rbx,qword ptr lazy_array_list[rip]
+	add	rbp,2
+	mov	qword ptr [rdx],rbx 
+	mov	qword ptr (-8)[rcx],rbp 
+	mov	qword ptr (-16)[rcx],rax 
+	sub	rcx,16
+	mov	qword ptr lazy_array_list[rip],rcx
+
+	mov	rcx,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rsi 
+	lea	rsi,(-8)[rdx]
+	att_jmp	__mark_node
+
+__mark_array_length_0_1:
+	lea	rcx,(-16)[rcx]
+	att_jb	__mark_next_node
+
+	mov	rbx,qword ptr 24[rcx]
+	mov	rbp,qword ptr 16[rcx]
+	mov	rdx,qword ptr lazy_array_list[rip]
+	mov	qword ptr 24[rcx],rbp 
+	mov	qword ptr 16[rcx],rdx 
+	mov	qword ptr [rcx],rax 
+	mov	qword ptr lazy_array_list[rip],rcx 
+	mov	qword ptr 8[rcx],rbx 
+	add	rcx,8
+
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi 
+	lea	rsi,2[rcx]
+	mov	rcx,rbp 
+	att_jmp	__mark_node
+
+__mark_b_record_array:
+	mov	rbp,qword ptr 8[rcx]
+	sub	rax,256
+	imul	rax,rbp 
+	add	rax,3
+	jmp	__mark_basic_array
+
+__mark_strict_basic_array:
+	mov	rax,qword ptr 8[rcx]
+	lea	r9,dINT+2[rip]
+	cmp	rbp,r9
+	jle	__mark__strict__int__or__real__array
+	lea	r9,BOOL+2[rip]
+	cmp	rbp,r9
+	je	__mark__strict__bool__array
+__mark__strict__int32__or__real32__array:
+	add	rax,6+1
+	shr	rax,1
+	att_jmp	__mark_basic_array
+__mark__strict__int__or__real__array:
+	add	rax,3
+	att_jmp	__mark_basic_array
+__mark__strict__bool__array:
+	add	rax,24+7
+	shr	rax,3
+	att_jmp	__mark_basic_array
+
+__mark_string_:
+	mov	rax,qword ptr 8[rcx]
+	add	rax,16+7
+	shr	rax,3
+
+__mark_basic_array:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,rax 
+
+	or	dword ptr [rdi+rbx*4],edx 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+
+	cmp	rbx,rax 
+	att_jae	__mark_next_node
+
+	inc	rbx 
+	mov	rbp,1
+
+	cmp	rbx,rax 
+	att_jae	__last__string__bits
+
+__mark_string_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx
+	cmp	rbx,rax 
+	att_jb	__mark_string_lp
+
+__last__string__bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+	att_jmp	__mark_next_node
diff --git a/macho64/amark_prefetch.s b/macho64/amark_prefetch.s
new file mode 100644
index 0000000..63b0360
--- /dev/null
+++ b/macho64/amark_prefetch.s
@@ -0,0 +1,1757 @@
+
+
+	.data
+n_queue_items:
+	.quad	0
+queue_first:
+	.quad	0
+queue:
+	.quad	0,0,0,0,0,0,0,0
+	.quad	0,0,0,0,0,0,0,0
+
+	.text
+
+pmark:
+	mov	rax,qword ptr heap_size_65[rip]
+	xor	rbx,rbx 
+
+	mov	qword ptr n_marked_words[rip],rbx
+	shl	rax,6
+
+	mov	qword ptr heap_size_64_65[rip],rax
+	mov	qword ptr lazy_array_list[rip],rbx 
+	
+	lea	rsi,(-4000)[rsp]
+
+	mov	rax,qword ptr caf_list[rip]
+
+	mov	qword ptr end_stack[rip],rsi 
+
+	mov	r15,0
+	mov	r8,0
+
+	mov	r10,neg_heap_p3[rip]
+	mov	r11,heap_size_64_65[rip]
+	mov	r13,qword ptr end_stack[rip]
+	mov	r14,0
+
+	test	rax,rax
+	je	end_pmark_cafs
+
+pmark_cafs_lp:
+	mov	rbx,qword ptr [rax]
+	mov	rbp,qword ptr (-8)[rax]
+
+	push	rbp
+	lea	rbp,8[rax]
+	lea	r12,8[rax+rbx*8]
+
+	call	pmark_stack_nodes
+
+	pop	rax
+	test	rax,rax 
+	att_jne	pmark_cafs_lp
+
+end_pmark_cafs:
+	mov	rsi,qword ptr stack_top[rip]
+	mov	rbp,qword ptr stack_p[rip]
+
+	mov	r12,rsi 
+	att_call	pmark_stack_nodes
+	att_jmp	continue_mark_after_pmark
+
+/* %rbp : pointer to stack element */
+/* %rdi : heap_vector */
+/* %rax ,%rbx ,%rcx ,%rdx ,%rsi : free */
+
+pmark_stack_nodes:
+	cmp	rbp,r12
+	je	end_pmark_nodes
+pmark_stack_nodes_:
+	mov	rcx,qword ptr [rbp]
+
+	add	rbp,8
+	lea	rdx,[r10+rcx] 
+
+	cmp	rdx,r11
+	att_jnc	pmark_stack_nodes
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_stack_nodes
+
+	push	rbp 
+
+	push	0
+
+	jmp	pmark_node_
+
+pmark_hnf_2:
+	cmp	rsi,0x20000000
+	jbe	pmark_fits_in_word_6
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_6:
+	add	r14,3
+
+pmark_record_2_c:
+	mov	rbx,qword ptr 8[rcx]
+	push	rbx 
+
+	cmp	rsp,r13
+	jb	pmarkr_using_reversal
+
+pmark_node2:
+pmark_shared_argument_part:
+	mov	rcx,qword ptr [rcx]
+
+pmark_node:
+	lea	rdx,[r10+rcx]
+	cmp	rdx,r11
+	jnc	pmark_next_node
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_next_node
+
+pmark_node_:
+
+	prefetch	[rcx]
+
+	lea	r9,queue[rip]
+	mov	qword ptr [r9+r8],rcx 
+	lea	rdx,[r8+r15*8]
+	add	r8,8
+
+	and	r8,15*8
+	and	rdx,15*8
+
+	cmp	r15,-4
+	je	pmark_last_item_in_queue
+
+pmark_add_items:
+	mov	rcx,qword ptr [rsp]
+	test	rcx,rcx 
+	jne	pmark_add_stacked_item
+
+pmark_add_items2:
+	mov	rbp,qword ptr 8[rsp]
+	cmp	rbp,r12
+	att_je	pmark_last_item_in_queue
+
+	mov	rcx,qword ptr [rbp]
+	add	rbp,8
+	mov	qword ptr 8[rsp],rbp 
+
+	lea	rbp,[r10+rcx]
+	cmp	rbp,r11
+	att_jnc	pmark_add_items2
+	mov	rax,rbp 
+	and	rbp,31*8
+	shr	rax,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rbp]
+	test	esi,dword ptr [rdi+rax*4]
+	att_jne	pmark_add_items2
+
+	prefetch	[rcx]
+
+	lea	r9,queue[rip]
+	mov	qword ptr [r9+r8],rcx 
+	add	r8,8
+	and	r8,15*8
+
+	sub	r15,1
+	
+	cmp	r15,-4
+	att_jne	pmark_add_items2
+	att_jmp	pmark_last_item_in_queue
+
+pmark_add_stacked_item:
+	add	rsp ,8
+
+	lea	rbp,[r10+rcx]
+	cmp	rbp,r11
+	att_jnc	pmark_add_items
+	mov	rax,rbp 
+	and	rbp,31*8
+	shr	rax,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rbp]
+	test	esi,dword ptr [rdi+rax*4]
+	att_jne	pmark_add_items
+
+	prefetch	[rcx]
+
+	lea	r9,queue[rip]
+	mov	qword ptr [r9+r8],rcx 
+	add	r8,8
+	and	r8,15*8
+
+	sub	r15,1
+
+	cmp	r15,-4 
+	att_jne	pmark_add_items
+
+pmark_last_item_in_queue:
+	lea	r9,queue[rip]
+	mov	rcx,qword ptr [r9+rdx]
+
+	lea	rdx,[r10+rcx]
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr[r9+rdx]
+		
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_next_node
+
+pmark_arguments:
+	mov	rax,qword ptr [rcx]
+	test	rax,2
+	je	pmark_lazy_node
+	
+	movzx	rbp,word ptr (-2)[rax]
+
+	test	rbp,rbp 
+	je	pmark_hnf_0
+
+	or	dword ptr [rdi+rbx*4],esi 
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	pmark_record
+
+	sub	rbp,2
+	att_je	pmark_hnf_2
+	jb	pmark_hnf_1
+
+pmark_hnf_3:
+	mov	rdx,qword ptr 8[rcx]
+
+	cmp	rsi,0x20000000
+	jbe	pmark_fits_in_word_1
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_1:	
+
+	add	r14,3
+	lea	rax,[r10+rdx]
+	mov	rbx,rax 
+
+	and	rax,31*8
+	shr	rbx,8
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rax]
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_shared_argument_part
+
+pmark_no_shared_argument_part:
+	or	dword ptr [rdi+rbx*4],esi 
+	add	rbp,1
+
+	add	r14,rbp 
+	lea	rax,[rax+rbp*8]
+	lea	rdx,(-8)[rdx+rbp*8]
+
+	cmp	rax,32*8
+	jbe	pmark_fits_in_word_2
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_2:
+
+	mov	rbx,qword ptr [rdx]
+	sub	rbp,2
+	push	rbx 
+
+pmark_push_hnf_args:
+	mov	rbx,qword ptr (-8)[rdx]
+	sub	rdx,8
+	push	rbx 
+	sub	rbp,1
+	att_jge	pmark_push_hnf_args
+
+	cmp	rsp,r13
+	att_jae	pmark_node2
+
+	att_jmp	pmarkr_using_reversal
+
+pmark_hnf_1:
+	cmp	rsi,0x40000000
+	jbe	pmark_fits_in_word_4
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_4:
+	add	r14,2
+	mov	rcx,qword ptr [rcx]
+	att_jmp	pmark_node
+
+pmark_lazy_node_1:
+	add	rcx,8
+	or	dword ptr [rdi+rbx*4],esi 
+	cmp	rsi,0x20000000
+	jbe	pmark_fits_in_word_3
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_3:
+	add	r14,3
+
+	cmp	rbp,1
+	att_je	pmark_node2
+
+pmark_selector_node_1:
+	add	rbp,2
+	mov	rdx,qword ptr [rcx]
+	je	pmark_indirection_node
+
+	lea	rsi,[r10+rdx]
+	mov	rbx,rsi 
+
+	shr	rbx,8
+	and	rsi,31*8
+
+	add	rbp,1
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rsi]
+	jle	pmark_record_selector_node_1
+
+	test	esi,dword ptr [rdi+rbx*4]
+	jne	pmark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	pmark_node3
+
+	cmp	word ptr (-2)[rbp],2
+	jbe	pmark_small_tuple_or_record
+
+pmark_large_tuple_or_record:
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	pmark_node3
+
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+
+	cmp	rax,16
+	jl	pmark_tuple_selector_node_1
+	mov	rdx,r9
+	je	pmark_tuple_selector_node_2
+	mov	rcx,qword ptr (-24)[r9+rax]
+	mov	qword ptr [rbp],rcx
+	att_jmp	pmark_node
+
+pmark_tuple_selector_node_2:
+	mov	rcx,qword ptr [r9]
+	mov	qword ptr [rbp],rcx
+	att_jmp	pmark_node	
+
+pmark_small_tuple_or_record:
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+pmark_tuple_selector_node_1:
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rbp],rcx	
+	att_jmp	pmark_node
+
+pmark_record_selector_node_1:
+	je	pmark_strict_record_selector_node_1
+
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	pmark_node3
+
+	cmp	word ptr (-2)[rbp],258
+	att_jbe	pmark_small_tuple_or_record
+
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	pmark_node3
+
+	movsxd	rbp,dword ptr(-8)[rax]
+	add	rax,rbp
+	lea	rbp,__indirection[rip]
+	mov	qword ptr (-8)[rcx],rbp
+	movzx	eax,word ptr (4-8)[rax]
+	mov	rbp,rcx
+
+	cmp	rax,16
+	jle	pmark_record_selector_node_2
+	mov	rdx,r9
+	sub	rax,24
+pmark_record_selector_node_2:
+	mov	rcx,qword ptr [rdx+rax]
+	mov	qword ptr [rbp],rcx
+	att_jmp	pmark_node
+
+pmark_strict_record_selector_node_1:
+	test	esi,dword ptr [rdi+rbx*4]
+	att_jne	pmark_node3
+
+	mov	rbp,qword ptr [rdx]
+	test	rbp,2
+	att_je	pmark_node3
+
+	cmp	word ptr (-2)[rbp],258
+	jbe	pmark_select_from_small_record
+
+	mov	rbp,qword ptr 16[rdx]
+	mov	r9,rbp
+
+	add	rbp,r10
+	mov	rbx,rbp 
+	and	rbp,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rbp]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_jne	pmark_node3
+	
+pmark_select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rax,rbx
+	sub	rcx,8
+
+	movzx	ebx,word ptr (4-8)[rax]
+	cmp	rbx,16
+	jle	pmark_strict_record_selector_node_2
+	mov	rbx,qword ptr (-24)[r9+rbx]
+	jmp	pmark_strict_record_selector_node_3
+pmark_strict_record_selector_node_2:
+	mov	rbx,qword ptr [rdx+rbx]
+pmark_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rbx
+
+	movzx	ebx,word ptr (6-8)[rax]
+	test	rbx,rbx
+	je	pmark_strict_record_selector_node_5
+	cmp	rbx,16
+	jle	pmark_strict_record_selector_node_4
+	mov	rdx,r9
+	sub	rbx,24
+pmark_strict_record_selector_node_4:
+	mov	rbx,qword ptr [rdx+rbx]
+	mov	qword ptr 16[rcx],rbx
+pmark_strict_record_selector_node_5:
+
+	mov	rax,qword ptr ((-8)-8)[rax]
+	mov	qword ptr [rcx],rax
+	att_jmp	pmark_next_node
+
+pmark_indirection_node:
+pmark_node3:
+	mov	rcx,rdx 
+	att_jmp	pmark_node
+
+pmark_next_node:
+	pop	rcx 
+	test	rcx,rcx 
+	att_jne	pmark_node
+
+	pop	rbp 
+	cmp	rbp,r12
+	att_jne	pmark_stack_nodes_
+
+end_pmark_nodes:
+	test	r15,r15
+	je	end_pmark_nodes_
+
+	push	rbp 
+
+	push	0
+
+	lea	rdx,[r8+r15*8]
+	add	r15,1
+
+	and	rdx,15*8
+
+	att_jmp	pmark_last_item_in_queue
+
+end_pmark_nodes_:
+	ret
+
+pmark_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp 
+	je	pmark_real_or_file
+
+	cmp	rbp,1
+	att_jle	pmark_lazy_node_1
+
+	cmp	rbp,256
+	jge	pmark_closure_with_unboxed_arguments
+	inc	rbp 
+	or	dword ptr [rdi+rbx*4],esi 
+
+	add	r14,rbp 
+	lea	rdx,[rdx+rbp*8]
+	lea	rcx,[rcx+rbp*8]
+
+	cmp	rdx,32*8
+	jbe	pmark_fits_in_word_7
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_7:
+	sub	rbp,3
+pmark_push_lazy_args:
+	mov	rbx,qword ptr (-8)[rcx]
+	sub	rcx,8
+	push	rbx 
+	sub	rbp,1
+	att_jge	pmark_push_lazy_args
+
+	sub	rcx,8
+
+	cmp	rsp,r13
+	att_jae	pmark_node2
+	
+	att_jmp	pmarkr_using_reversal
+
+pmark_closure_with_unboxed_arguments:
+	mov	rax,rbp 
+	and	rbp,255
+	sub	rbp,1
+	att_je	pmark_real_or_file
+
+	shr	rax,8
+	add	rbp,2
+	
+	or	dword ptr [rdi+rbx*4],esi 
+	add	r14,rbp 
+	lea	rdx,[rdx+rbp*8]
+
+	sub	rbp,rax 
+
+	cmp	rdx,32*8
+	jbe	pmark_fits_in_word_7_
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_7_:
+	sub	rbp,2
+	att_jl	pmark_next_node
+
+	lea	rcx,16[rcx+rbp*8]
+	att_jne	pmark_push_lazy_args
+
+pmark_closure_with_one_boxed_argument:
+	mov	rcx,qword ptr (-8)[rcx]
+	att_jmp	pmark_node
+
+pmark_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jb	pmark_real_file_or_string
+
+	or	dword ptr [rdi+rbx*4],esi 
+
+	lea	r9,CHAR+2[rip]
+	cmp	rax,r9
+	ja	pmark_normal_hnf_0
+
+pmark_bool:
+	add	r14,2
+
+	cmp	rsi,0x40000000
+	att_jbe	pmark_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	pmark_next_node
+
+pmark_normal_hnf_0:
+	inc	r14
+	att_jmp	pmark_next_node
+
+pmark_real_file_or_string:
+	lea	r9,__STRING__+2[rip]
+	cmp	rax,r9
+	jbe	pmark_string_or_array
+
+pmark_real_or_file:
+	or	dword ptr [rdi+rbx*4],esi 
+	add	r14,3
+
+	cmp	rsi,0x20000000
+	att_jbe	pmark_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	pmark_next_node
+
+pmark_record:
+	sub	rbp,258
+	je	pmark_record_2
+	jl	pmark_record_1
+
+pmark_record_3:
+	add	r14,3
+
+	cmp	rsi,0x20000000
+	jbe	pmark_fits_in_word_13
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_13:
+	mov	rdx,qword ptr 8[rcx]
+
+	movzx	rbx,word ptr (-2+2)[rax]
+	lea	rsi,[r10+rdx] 
+
+	mov	rax,rsi 
+	and	rsi,31*8
+
+	shr	rax,8
+	sub	rbx,1
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rsi]
+	jb	pmark_record_3_bb
+
+	test	edx,dword ptr [rdi+rax*4]
+	att_jne	pmark_node2
+
+	add	rbp,1
+	or	dword ptr [rdi+rax*4],edx 
+	add	r14,rbp 
+	lea	rsi,[rsi+rbp*8]
+
+	cmp	rsi,32*8
+	jbe	pmark_push_record_arguments
+	or	dword ptr 4[rdi+rax*4],1
+pmark_push_record_arguments:
+	mov	rdx,qword ptr 8[rcx]
+	mov	rbp,rbx 
+	shl	rbx,3
+	add	rdx,rbx 
+	sub	rbp,1
+	att_jge	pmark_push_hnf_args
+
+	att_jmp	pmark_node2
+
+pmark_record_3_bb:
+	test	edx,dword ptr [rdi+rax*4]
+	att_jne	pmark_next_node
+
+	add	rbp,1
+	or	dword ptr [rdi+rax*4],edx 
+	add	r14,rbp 
+	lea	rsi,[rsi+rbp*8]
+	
+	cmp	rsi,32*8
+	att_jbe	pmark_next_node
+
+	or	dword ptr 4[rdi+rax*4],1
+	att_jmp	pmark_next_node
+
+pmark_record_2:
+	cmp	rsi,0x20000000
+	jbe	pmark_fits_in_word_12
+	or	dword ptr 4[rdi+rbx*4],1
+pmark_fits_in_word_12:
+	add	r14,3
+
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	pmark_record_2_c
+	att_je	pmark_node2
+	att_jmp	pmark_next_node
+
+pmark_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	pmark_hnf_1
+
+	att_jmp	pmark_bool
+
+pmark_string_or_array:
+	je	pmark_string_
+
+pmark_array:
+	mov	rbp,qword ptr 16[rcx]
+	test	rbp,rbp 
+	je	pmark_lazy_array
+
+	movzx	rax,word ptr (-2)[rbp]
+
+	test	rax,rax 
+	je	pmark_strict_basic_array
+
+	movzx	rbp,word ptr (-2+2)[rbp]
+	test	rbp,rbp 
+	je	pmark_b_record_array
+
+	cmp	rsp,r13
+	jb	pmark_array_using_reversal
+
+	sub	rax,256
+	cmp	rax,rbp 
+	je	pmark_a_record_array
+
+pmark_ab_record_array:
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rbp,qword ptr 8[rcx]
+
+	imul	rax,rbp 
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	jae	pmark_end_set_ab_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	pmark_last_ab_array_bits
+
+pmark_ab_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx 
+	cmp	rbx,rax
+	att_jb	pmark_ab_array_lp
+
+pmark_last_ab_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+
+pmark_end_set_ab_array_bits:
+	mov	rax,qword ptr 8[rcx]
+	mov	rdx,qword ptr 16[rcx]
+	movzx	rbx,word ptr (-2+2)[rdx]
+	movzx	rdx,word ptr (-2)[rdx]
+	shl	rbx,3
+	lea	rdx,(-2048)[rdx*8]
+	push	rbx 
+	push	rdx 
+	lea	rbp,24[rcx]
+	push	r12
+	jmp	pmark_ab_array_begin
+
+pmark_ab_array:
+	mov	rbx,qword ptr 16[rsp]
+	push	rax 
+	push	rbp 
+	lea	r12,[rbp+rbx]
+
+	att_call	pmark_stack_nodes
+
+	mov	rbx,qword ptr (8+16)[rsp]
+	pop	rbp 
+	pop	rax 
+	add	rbp,rbx 
+pmark_ab_array_begin:
+	sub	rax,1
+	att_jnc	pmark_ab_array
+
+	pop	r12
+	add	rsp,16
+	att_jmp	pmark_next_node
+
+pmark_a_record_array:
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rbp,qword ptr 8[rcx]
+
+	imul	rax,rbp 
+	push	rax 
+
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	jae	pmark_end_set_a_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	pmark_last_a_array_bits
+
+pmark_a_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx 
+	cmp	rbx,rax 
+	att_jb	pmark_a_array_lp
+
+pmark_last_a_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+
+pmark_end_set_a_array_bits:
+	pop	rax 
+	lea	rbp,24[rcx]
+
+	push	r12
+	lea	r12,24[rcx+rax*8]
+
+	att_call	pmark_stack_nodes
+
+	pop	r12
+	att_jmp	pmark_next_node
+
+pmark_lazy_array:
+	cmp	rsp,r13
+	att_jb	pmark_array_using_reversal
+
+	or	dword ptr [rdi+rbx*4],esi 
+	mov	rax,qword ptr 8[rcx]
+
+	add	rax,3
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	jae	pmark_end_set_lazy_array_bits
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	pmark_last_lazy_array_bits
+
+pmark_lazy_array_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx 
+	cmp	rbx,rax
+	att_jb	pmark_lazy_array_lp
+
+pmark_last_lazy_array_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+
+pmark_end_set_lazy_array_bits:
+	mov	rax,qword ptr 8[rcx]
+	lea	rbp,24[rcx]
+
+	push	r12
+	lea	r12,24[rcx+rax*8]
+
+	att_call	pmark_stack_nodes
+
+	pop	r12
+	att_jmp	pmark_next_node
+
+pmark_array_using_reversal:
+	push	0
+	mov	rsi,1
+	jmp	pmarkr_node
+
+pmark_strict_basic_array:
+	mov	rax,qword ptr 8[rcx]
+	lea	r9,dINT+2[rip]
+	cmp	rbp,r9
+	jle	pmark_strict_int_or_real_array
+	lea	r9,BOOL+2[rip]
+	cmp	rbp,r9
+	je	pmark_strict_bool_array
+	add	rax,6+1
+	shr	rax,1
+	jmp	pmark_basic_array_
+pmark_strict_int_or_real_array:
+	add	rax,3
+	att_jmp	pmark_basic_array_
+pmark_strict_bool_array:
+	add	rax,24+7
+	shr	rax,3
+	att_jmp	pmark_basic_array_
+
+pmark_b_record_array:
+	mov	rbp,qword ptr 8[rcx]
+	sub	rax,256
+	imul	rax,rbp 
+	add	rax,3
+	att_jmp	pmark_basic_array_
+
+pmark_string_:
+	mov	rax,qword ptr 8[rcx]
+	add	rax,16+7
+	shr	rax,3
+
+pmark_basic_array_:
+	or	dword ptr [rdi+rbx*4],esi 
+
+	add	r14,rax 
+	lea	rax,(-8)[rcx+rax*8]
+
+	add	rax,r10
+	shr	rax,8
+	
+	cmp	rbx,rax 
+	att_jae	pmark_next_node
+
+	inc	rbx 
+	mov	rbp,1
+	cmp	rbx,rax 
+	jae	pmark_last_string_bits
+
+pmark_string_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx
+	cmp	rbx,rax 
+	att_jb	pmark_string_lp
+
+pmark_last_string_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+	att_jmp	pmark_next_node
+
+end_pmarkr_using_reversal:
+	pop	rdx 
+	test	rdx,rdx 
+	att_je	pmark_next_node
+	mov	qword ptr [rdx],rcx 
+	att_jmp	pmark_next_node
+
+
+pmarkr_using_reversal:
+	push	rcx 
+	mov	rsi,1
+	mov	rcx,qword ptr [rcx]
+	att_jmp	pmarkr_node
+
+pmarkr_arguments:
+	mov	rax,qword ptr [rcx]
+	test	al,2
+	je	pmarkr_lazy_node
+
+	movzx	rbp,word ptr (-2)[rax]
+	test	rbp,rbp 
+	je	pmarkr_hnf_0
+
+	add	rcx,8
+
+	cmp	rbp,256
+	jae	pmarkr_record
+
+	sub	rbp,2
+	je	pmarkr_hnf_2
+	jb	pmarkr_hnf_1
+
+pmarkr_hnf_3:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+
+	or	dword ptr [rdi+rbx*4],edx 
+
+	cmp	rdx,0x20000000
+
+	mov	rax,qword ptr 8[rcx]
+
+	jbe	pmarkr_fits_in_word_1
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_1:
+	add	rax,r10
+
+	mov	rbx,rax 
+	and	rax,31*8
+
+	shr	rbx,8
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rax]
+	test	edx,dword ptr [rdi+rbx*4]
+	jne	pmarkr_shared_argument_part
+
+pmarkr_no_shared_argument_part:
+	or	dword ptr [rdi+rbx*4],edx 
+	mov	rdx,qword ptr 8[rcx]
+
+	add	rbp,1
+	mov	qword ptr 8[rcx],rsi 
+
+	add	r14,rbp 
+	add	rcx,8
+
+	shl	rbp,3
+	or	qword ptr [rdx],1
+
+	add	rax,rbp
+	add	rdx,rbp
+
+	cmp	rax,32*8
+	jbe	pmarkr_fits_in_word_2
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_2:
+
+	mov	rbp ,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rcx 
+	lea	rsi,(-8)[rdx]
+	mov	rcx,rbp 
+	att_jmp	pmarkr_node
+
+pmarkr_hnf_1:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,2
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x40000000
+	att_jbe	pmarkr_shared_argument_part
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_shared_argument_part:
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi 
+	lea	rsi,2[rcx]
+	mov	rcx,rbp
+	att_jmp	pmarkr_node
+
+pmarkr_no_selector_2:
+	pop	rbx 
+pmarkr_no_selector_1:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	att_jbe	pmarkr_shared_argument_part
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	pmarkr_shared_argument_part
+
+pmarkr_lazy_node_1:
+	att_je	pmarkr_no_selector_1
+
+pmarkr_selector_node_1:
+	add	rbp,2
+	je	pmarkr_indirection_node
+
+	add	rbp,1
+
+	push	rbx 
+	mov	rbp,qword ptr [rcx]
+	push	rax 
+	lea	rax,[r10+rbp]
+
+	jle	pmarkr_record_selector_node_1
+
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx*4]
+	pop	rax 
+	att_jne	pmarkr_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	pmarkr_no_selector_2
+
+	cmp	word ptr (-2)[rbx],2
+	jbe	pmarkr_small_tuple_or_record
+
+pmarkr_large_tuple_or_record:
+	mov	r8,qword ptr 16[rbp]
+	mov	r9,r8
+
+	add	r8,r10
+	mov	rbx,r8
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+	att_jne	pmarkr_no_selector_2
+
+	movsxd	rdx,dword ptr (-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx 
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+
+	cmp	rax,16
+	jl	pmarkr_tuple_selector_node_1
+	mov	rdx,r9
+	je	pmarkr_tuple_selector_node_2
+	mov	rcx,qword ptr (-24)[r9+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	pmarkr_node
+
+pmarkr_tuple_selector_node_2:
+	mov	rcx,qword ptr [r9]
+	mov	qword ptr [r8],rcx
+	att_jmp	pmarkr_node
+
+pmarkr_small_tuple_or_record:
+	movsxd	rdx,dword ptr(-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx 
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+pmarkr_tuple_selector_node_1:
+	mov	rcx,qword ptr [rbp+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	pmarkr_node
+
+pmarkr_record_selector_node_1:
+	je	pmarkr_strict_record_selector_node_1
+
+	mov	rbx,rax
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx*4]
+	pop	rax 
+	att_jne	pmarkr_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	pmarkr_no_selector_2
+	
+	cmp	word ptr (-2)[rbx],258
+	jbe	pmarkr_small_record
+
+	mov	r8,qword ptr 16[rbp]
+
+	add	r8,r10
+	mov	rbx,r8
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+
+	mov	r9,qword ptr 16[rbp]
+
+	att_jne	pmarkr_no_selector_2
+
+pmarkr_small_record:
+	movsxd	rdx,dword ptr (-8)[rax]
+	add	rax,rdx
+	lea	rdx,__indirection[rip]
+	pop	rbx 
+
+	mov	qword ptr (-8)[rcx],rdx
+	movzx	eax,word ptr (4-8)[rax]
+	mov	r8,rcx
+
+	cmp	rax,16
+	jle	pmarkr_record_selector_node_2
+	mov	rdx,r9
+	sub	rax,24
+pmarkr_record_selector_node_2:
+	mov	rcx,qword ptr [rbp+rax]
+	mov	qword ptr [r8],rcx
+	att_jmp	pmarkr_node
+
+pmarkr_strict_record_selector_node_1:
+	mov	rbx,rax 
+	and	rax,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	eax,dword ptr [r9+rax]
+	test	eax,dword ptr [rdi+rbx*4]
+	pop	rax 
+	att_jne	pmarkr_no_selector_2
+
+	mov	rbx,qword ptr [rbp]
+	test	bl,2
+	att_je	pmarkr_no_selector_2
+
+	cmp	word ptr (-2)[rbx],258
+	jle	pmarkr_select_from_small_record
+
+	mov	r8,qword ptr 16[rbp]
+
+	add	r8,r10
+	mov	rbx,r8 
+	and	r8,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	r8d,dword ptr [r9+r8]
+	test	r8d,dword ptr [rdi+rbx*4]
+
+	mov	r9,qword ptr 16[rbp]
+
+	att_jne	pmarkr_no_selector_2
+
+pmarkr_select_from_small_record:
+	movsxd	rbx,dword ptr(-8)[rax]
+	add	rax,rbx
+	sub	rcx,8
+
+	movzx	ebx,word ptr (4-8)[rax]
+	cmp	rbx,16
+	jle	pmarkr_strict_record_selector_node_2
+	mov	rbx,qword ptr (-24)[r9+rbx]
+	jmp	pmarkr_strict_record_selector_node_3
+pmarkr_strict_record_selector_node_2:
+	mov	rbx,qword ptr [rdx+rbx]
+pmarkr_strict_record_selector_node_3:
+	mov	qword ptr 8[rcx],rbx
+
+	movzx	ebx,word ptr (6-8)[rax]
+	test	rbx,rbx
+	je	pmarkr_strict_record_selector_node_5
+	cmp	rbx,16
+	jle	pmarkr_strict_record_selector_node_4
+	mov	rbp,r9
+	sub	rbx,24
+pmarkr_strict_record_selector_node_4:
+	mov	rbx,qword ptr [rbp+rbx]
+	mov	qword ptr 16[rcx],rbx
+pmarkr_strict_record_selector_node_5:
+	pop	rbx
+
+	mov	rax,qword ptr ((-8-8))[rax]
+	mov	qword ptr [rcx],rax
+	att_jmp	pmarkr_node
+
+pmarkr_indirection_node:
+	mov	rcx,qword ptr [rcx]
+	att_jmp	pmarkr_node
+
+pmarkr_hnf_2:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	jbe	pmarkr_fits_in_word_6
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_6:
+
+pmarkr_record_2_c:
+	mov	rax,qword ptr [rcx]
+	mov	rbp,qword ptr 8[rcx]
+	or	rax,2
+	mov	qword ptr 8[rcx],rsi 
+	mov	qword ptr [rcx],rax 
+	lea	rsi,8[rcx]
+	mov	rcx,rbp 
+
+pmarkr_node:
+	lea	rdx,[r10+rcx] 
+
+	cmp	rdx,r11
+	jae	pmarkr_next_node
+
+	mov	rbx,rdx 
+	and	rdx,31*8
+	shr	rbx,8
+	lea	r9,bit_set_table2[rip]
+	mov	ebp,dword ptr [r9+rdx]
+	test	ebp,dword ptr [rdi+rbx*4]
+	att_je	pmarkr_arguments
+
+pmarkr_next_node:
+	test	rsi,3
+	jne	pmarkr_parent
+
+	mov	rbp,qword ptr (-8)[rsi]
+	mov	rdx,qword ptr [rsi]
+	mov	qword ptr [rsi],rcx 
+	mov	qword ptr (-8)[rsi],rdx 
+	sub	rsi,8
+
+	mov	rcx,rbp 
+	and	rbp,3
+	and	rcx,-4
+	or	rsi,rbp 
+	att_jmp	pmarkr_node
+
+pmarkr_parent:
+	mov	rbx,rsi 
+	and	rsi,-4
+	att_je	end_pmarkr_using_reversal
+
+	and	rbx,3
+	mov	rbp,qword ptr [rsi]
+	mov	qword ptr [rsi],rcx 
+
+	sub	rbx,1
+	je	pmarkr_argument_part_parent
+
+	lea	rcx,(-8)[rsi]
+	mov	rsi,rbp 
+	att_jmp	pmarkr_next_node
+	
+pmarkr_argument_part_parent:
+	and	rbp,-4
+	mov	rdx,rsi 
+	mov	rcx,qword ptr (-8)[rbp]
+	mov	rbx,qword ptr [rbp]
+	mov	qword ptr (-8)[rbp],rbx 
+	mov	qword ptr [rbp],rdx 
+	lea	rsi,(2-8)[rbp]
+	att_jmp	pmarkr_node
+
+pmarkr_lazy_node:
+	movsxd	rbp,dword ptr (-4)[rax]
+	test	rbp,rbp 
+	je	pmarkr_real_or_file
+
+	add	rcx,8
+	cmp	rbp,1
+	att_jle	pmarkr_lazy_node_1
+	cmp	rbp,256
+	jge	pmarkr_closure_with_unboxed_arguments
+
+	add	rbp,1
+	mov	rax,rdx 
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,rbp 
+
+	lea	rax,[rax+rbp*8]
+	sub	rbp,2
+
+	or	dword ptr [rdi+rbx*4],edx 
+
+	cmp	rax,32*8
+	jbe	pmarkr_fits_in_word_7
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_7:
+pmarkr_closure_with_unboxed_arguments_2:
+	lea	rdx,[rcx+rbp*8]
+	mov	rax,qword ptr [rcx]
+	or	rax,2
+	mov	qword ptr [rcx],rax 
+	mov	rcx,qword ptr [rdx]
+	mov	qword ptr [rdx],rsi 
+	mov	rsi,rdx 
+	att_jmp	pmarkr_node
+
+pmarkr_closure_with_unboxed_arguments:
+	mov	rax,rbp 
+	and	rbp,255
+
+	sub	rbp,1
+	je	pmarkr_closure_1_with_unboxed_argument
+	add	rbp,2
+
+	shr	rax,8
+	add	r14,rbp 
+
+	push	rcx 
+	lea	rcx,[rdx+rbp*8]
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	sub	rbp,rax 
+
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rcx,32*8
+	jbe	pmarkr_fits_in_word_7_
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_7_:
+	pop	rcx
+	sub	rbp,2
+	att_jg	pmarkr_closure_with_unboxed_arguments_2
+	att_je	pmarkr_shared_argument_part
+	sub	rcx,8
+	att_jmp	pmarkr_next_node
+
+pmarkr_closure_1_with_unboxed_argument:
+	sub	rcx,8
+	att_jmp	pmarkr_real_or_file
+
+pmarkr_hnf_0:
+	lea	r9,dINT+2[rip]
+	cmp	rax,r9
+	jne	pmarkr_no_int_3
+
+	mov	rbp,qword ptr 8[rcx]
+	cmp	rbp,33
+
+	jb	pmarkr_small_int
+
+pmarkr_bool_or_small_string:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,2
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x40000000
+	att_jbe	pmarkr_next_node
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	pmarkr_next_node
+
+pmarkr_small_int:
+	shl	rbp,4
+	lea	rcx,small_integers[rip]
+	add	rcx,rbp
+	att_jmp	pmarkr_next_node
+
+pmarkr_no_int_3:
+	jb	pmarkr_real_file_or_string
+
+ 	lea	r9,CHAR+2[rip]
+ 	cmp	rax,r9
+ 	jne	pmarkr_no_char_3
+
+	movzx	rbp,byte ptr 8[rcx]
+	shl	rbp,4
+	lea	rcx,static_characters[rip]
+	add	rcx,rbp
+	att_jmp	pmarkr_next_node
+
+pmarkr_no_char_3:
+	att_jb	pmarkr_bool_or_small_string
+
+	lea	rcx,((-8)-2)[rax]
+	att_jmp	pmarkr_next_node
+
+pmarkr_real_file_or_string:
+	lea	r9,__STRING__+2[rip]
+	cmp	rax,r9
+	jbe	pmarkr_string_or_array
+
+pmarkr_real_or_file:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+
+	or	dword ptr [rdi+rbx*4],edx 
+	
+	cmp	rdx,0x20000000
+	att_jbe	pmarkr_next_node
+
+	or	dword ptr 4[rdi+rbx*4],1
+	att_jmp	pmarkr_next_node
+
+pmarkr_record:
+	sub	rbp,258
+	je	pmarkr_record_2
+	jl	pmarkr_record_1
+
+pmarkr_record_3:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	jbe	pmarkr_fits_in_word_13
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_13:
+	movzx	rbx,word ptr (-2+2)[rax]
+
+	mov	rdx,qword ptr 8[rcx]
+	lea	rdx,[r10+rdx]
+	mov	rax,rdx 
+	and	rdx,31*8
+	shr	rax,8
+
+	push	rsi 
+
+	lea	r9,bit_set_table2[rip]
+	mov	esi,dword ptr [r9+rdx]
+	test	esi,dword ptr [rdi+rax*4]
+	jne	pmarkr_shared_record_argument_part
+
+	add	rbp,1
+	or	dword ptr [rdi+rax*4],esi 
+
+	lea	rdx,[rdx+rbp*8]
+	add	r14,rbp 
+
+	pop	rsi 
+
+	cmp	rdx,32*8
+	jbe	pmarkr_fits_in_word_14
+	or	dword ptr 4[rdi+rax*4],1
+pmarkr_fits_in_word_14:
+	sub	rbx,1
+	mov	rdx,qword ptr 8[rcx]
+	jl	pmarkr_record_3_bb
+	att_je	pmarkr_shared_argument_part
+
+	mov	qword ptr 8[rcx],rsi 
+	add	rcx,8
+
+	sub	rbx,1
+	je	pmarkr_record_3_aab
+
+	lea	rsi,[rdx+rbx*8]
+	mov	rax,qword ptr [rdx]
+	or	rax,1
+	mov	rbp,qword ptr [rsi]
+	mov	qword ptr [rdx],rax 
+	mov	qword ptr [rsi],rcx 
+	mov	rcx,rbp
+	att_jmp	pmarkr_node
+
+pmarkr_record_3_bb:
+	sub	rcx,8
+	att_jmp	pmarkr_next_node
+
+pmarkr_record_3_aab:
+	mov	rbp,qword ptr [rdx]
+	mov	qword ptr [rdx],rcx 
+	lea	rsi,1[rdx]
+	mov	rcx,rbp 
+	att_jmp	pmarkr_node
+
+pmarkr_shared_record_argument_part:
+	mov	rdx,qword ptr 8[rcx]
+
+	pop	rsi 
+
+	test	rbx,rbx 
+	att_jne	pmarkr_shared_argument_part
+	sub	rcx,8
+	att_jmp	pmarkr_next_node
+
+pmarkr_record_2:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,3
+	or	dword ptr [rdi+rbx*4],edx 
+	cmp	rdx,0x20000000
+	jbe	pmarkr_fits_in_word_12
+	or	dword ptr 4[rdi+rbx*4],1
+pmarkr_fits_in_word_12:
+	cmp	word ptr (-2+2)[rax],1
+	att_ja	pmarkr_record_2_c
+	att_je	pmarkr_shared_argument_part
+	sub	rcx,8
+	att_jmp	pmarkr_next_node
+
+pmarkr_record_1:
+	cmp	word ptr (-2+2)[rax],0
+	att_jne	pmarkr_hnf_1
+	sub	rcx,8
+	att_jmp	pmarkr_bool_or_small_string
+
+pmarkr_string_or_array:
+	je	pmarkr_string_
+
+pmarkr_array:
+	mov	rbp,qword ptr 16[rcx]
+	test	rbp,rbp 
+	je	pmarkr_lazy_array
+
+	movzx	rax,word ptr (-2)[rbp]
+	test	rax,rax 
+	je	pmarkr_strict_basic_array
+
+	movzx	rbp,word ptr (-2+2)[rbp]
+	test	rbp,rbp
+	je	pmarkr_b_record_array
+
+	sub	rax,256
+	cmp	rax,rbp 
+	je	pmarkr_a_record_array
+
+pmarkr_ab_record_array:
+	push	rdx 
+	push	rbx 
+	mov	rbx,rbp 
+
+	mov	rbp,qword ptr 8[rcx]
+	add	rcx,16
+	push	rcx 
+
+	shl	rbp,3
+	mov	rdx,rax 
+	imul	rdx,rbp 
+
+	sub	rax,rbx 
+	add	rcx,8
+	add	rdx,rcx 
+
+	att_call	reorder
+	
+	pop	rcx 
+
+	xchg	rax,rbx 
+	mov	rbp,qword ptr (-8)[rcx]
+	imul	rax,rbp 
+	imul	rbx,rbp 
+	add	r14,rbx 
+	add	rbx,rax 
+
+	shl	rbx,3
+	lea	rbp,[r10+rcx]
+	add	rbp,rbx 
+
+	pop	rbx 
+	pop	rdx 
+
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	or	dword ptr [rdi+rbx*4],edx 
+
+	lea	rdx,[rcx+rax*8]
+	jmp	pmarkr_r_array
+
+pmarkr_a_record_array:
+	imul	rax,qword ptr 8[rcx]
+	add	rcx,16
+	jmp	pmarkr_lr_array
+
+pmarkr_lazy_array:
+	mov	rax,qword ptr 8[rcx]
+	add	rcx,16
+
+pmarkr_lr_array:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	mov	rbp,r10
+	or	dword ptr [rdi+rbx*4],edx
+	lea	rdx,[rcx+rax*8]
+	add	rbp,rdx 
+pmarkr_r_array:
+	shr	rbp,8
+
+	cmp	rbx,rbp 
+	jae	pmarkr_skip_mark_lazy_array_bits
+
+	inc	rbx 
+
+pmarkr_lazy_array_bits:
+	or	dword ptr [rdi+rbx*4],1
+	inc	rbx 
+	cmp	rbx,rbp 
+	att_jbe	pmarkr_lazy_array_bits
+
+pmarkr_skip_mark_lazy_array_bits:
+	add	r14,3
+	add	r14,rax 
+
+	cmp	rax,1
+	jbe	pmarkr_array_length_0_1
+
+	mov	rbp,qword ptr [rdx]
+	mov	rbx,qword ptr [rcx]
+	mov	qword ptr [rdx],rbx 
+	mov	qword ptr [rcx],rbp 
+	
+	mov	rbp,qword ptr (-8)[rdx]
+	sub	rdx,8
+	mov	rbx,qword ptr lazy_array_list[rip]
+	add	rbp,2
+	mov	qword ptr [rdx],rbx 
+	mov	qword ptr (-8)[rcx],rbp 
+	mov	qword ptr (-16)[rcx],rax 
+	sub	rcx,16
+	mov	qword ptr lazy_array_list[rip],rcx 
+
+	mov	rcx,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rsi 
+	lea	rsi,(-8)[rdx]
+	att_jmp	pmarkr_node
+
+pmarkr_array_length_0_1:
+	lea	rcx,(-16)[rcx]
+	att_jb	pmarkr_next_node
+
+	mov	rbx,qword ptr 24[rcx]
+	mov	rbp,qword ptr 16[rcx]
+	mov	rdx,qword ptr lazy_array_list[rip]
+	mov	qword ptr 24[rcx],rbp 
+	mov	qword ptr 16[rcx],rdx 
+	mov	qword ptr [rcx],rax 
+	mov	qword ptr lazy_array_list[rip],rcx
+	mov	qword ptr 8[rcx],rbx
+	add	rcx,8
+
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rcx],rsi 
+	lea	rsi,2[rcx]
+	mov	rcx,rbp 
+	att_jmp	pmarkr_node
+	
+pmarkr_b_record_array:
+	mov	rbp,qword ptr 8[rcx]
+	sub	rax,256
+	imul	rax,rbp 
+	add	rax,3
+	jmp	pmarkr_basic_array
+
+pmarkr_strict_basic_array:
+	mov	rax,qword ptr 8[rcx]
+	lea	r9,dINT+2[rip]
+	cmp	rbp,r9
+	jle	pmarkr_strict_int_or_real_array
+	lea	r9,BOOL+2[rip]
+	cmp	rbp,r9
+	je	pmarkr_strict_bool_array
+	add	rax,6+1
+	shr	rax,1
+	att_jmp	pmarkr_basic_array
+pmarkr_strict_int_or_real_array:
+	add	rax,3
+	att_jmp	pmarkr_basic_array
+pmarkr_strict_bool_array:
+	add	rax,24+7
+	shr	rax,3
+	att_jmp	pmarkr_basic_array
+
+pmarkr_string_:
+	mov	rax,qword ptr 8[rcx]
+	add	rax,16+7
+	shr	rax,3
+
+pmarkr_basic_array:
+	lea	r9,bit_set_table2[rip]
+	mov	edx,dword ptr [r9+rdx]
+	add	r14,rax 
+
+	or	dword ptr [rdi+rbx*4],edx 
+	lea	rax,(-8)[rcx+rax*8]
+	
+	add	rax,r10
+	shr	rax,8
+
+	cmp	rbx,rax 
+	att_jae	pmarkr_next_node
+
+	inc	rbx 
+	mov	rbp,1
+
+	cmp	rbx,rax 
+	jae	pmarkr_last_string_bits
+
+pmarkr_string_lp:
+	or	dword ptr [rdi+rbx*4],ebp 
+	inc	rbx 
+	cmp	rbx,rax 
+	att_jb	pmarkr_string_lp
+
+pmarkr_last_string_bits:
+	or	dword ptr [rdi+rbx*4],ebp 
+	att_jmp	pmarkr_next_node
diff --git a/macho64/areals.s b/macho64/areals.s
new file mode 100644
index 0000000..363f824
--- /dev/null
+++ b/macho64/areals.s
@@ -0,0 +1,8965 @@
+
+	.intel_syntax noprefix
+
+	.macro att_jp
+	.att_syntax
+	jp	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.text
+
+	.globl	sin_real
+
+sin_real:
+	ucomisd	xmm0,qword ptr real_pi_d_4[rip]
+	jbe		sin_real_1					# x<=pi/4 | NAN
+	ucomisd	xmm0,qword ptr real_3_pi_d_4[rip]
+	jbe		sin_real_2
+	ucomisd	xmm0,qword ptr real_5_pi_d_4[rip]
+	jbe		sin_real_3
+	ucomisd	xmm0,qword ptr real_7_pi_d_4[rip]
+	jbe		sin_real_4
+	ucomisd	xmm0,qword ptr real_9_pi_d_4[rip]
+	jbe		sin_real_5
+
+	ucomisd	xmm0,qword ptr real_36825084_pi[rip]
+	jbe		sin_real_0
+
+	ucomisd	xmm0,qword ptr real_2_p_53[rip]
+	jae		sin_cos_or_tan_real_too_large	# x>=-2^53
+
+	call	rem_36825084_pi
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+	subsd	xmm13,xmm7
+	jmp		sin_real_0_
+
+sin_real_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+
+	movq	xmm4,qword ptr round_even_c[rip]
+	
+	addsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi+1
+	
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi+1)
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm6,xmm4
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+sin_real_0_:
+	ucomisd	xmm1,xmm5
+	jne		sin_real_cos
+
+sin_real_sin:
+	ucomisd	xmm1,xmm6
+	jne		sin_real_n_sin
+
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+
+	jmp		sin_real_p
+
+sin_real_n_sin:
+	movsd	xmm1,xmm3
+	subsd	xmm0,xmm2
+	subsd	xmm1,xmm0
+	jmp		sin_real_m
+
+sin_real_cos:
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+
+	ucomisd	xmm1,xmm6
+	je		cos_real_p
+	jmp		cos_real_m
+
+sin_real_n:
+	ucomisd	xmm0,qword ptr real_m_3_pi_d_4[rip]
+	jae		sin_real_n_2
+	ucomisd	xmm0,qword ptr real_m_5_pi_d_4[rip]
+	jae		sin_real_n_3
+	ucomisd	xmm0,qword ptr real_m_7_pi_d_4[rip]
+	jae		sin_real_n_4
+	ucomisd	xmm0,qword ptr real_m_9_pi_d_4[rip]
+	jae		sin_real_n_5
+
+	ucomisd	xmm0,qword ptr real_m_36825084_pi[rip]
+	jae		sin_real_n_0
+
+	ucomisd	xmm0,qword ptr real_m_2_p_53[rip]
+	jbe		sin_cos_or_tan_real_too_small_or_nan	# x<=-2^53 | NAN
+
+	call	rem_n_36825084_pi
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+	subsd	xmm13,xmm7
+	jmp		sin_real_n_0_
+
+sin_real_n_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+	
+	movq	xmm4,qword ptr m_round_even_c[rip]
+
+	subsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi-1
+	
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi-1)
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm6,xmm4
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+sin_real_n_0_:
+	ucomisd	xmm1,xmm5
+	att_je		sin_real_sin
+
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+
+	ucomisd	xmm1,xmm6
+	att_jne		cos_real_p
+	att_jmp		cos_real_m
+
+sin_real_1:
+	ucomisd	xmm0,qword ptr real_m_pi_d_4[rip]
+	att_jb		sin_real_n				# # x<-pi/4 | NAN
+
+	ucomisd	xmm0,qword ptr real_0_43540000008249979402[rip]
+	jae		sin_real_p_0_6
+	ucomisd	xmm0,qword ptr real_m_0_43540000008249979402[rip]
+	jbe		sin_real_m_0_6
+
+	movsd	xmm1,xmm0				# x
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	je		sin_real_0_0			# sin -0.0 = -0.0
+
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm5,qword ptr sin_p_0[rip]
+	movlpd	xmm6,qword ptr sin_p_1[rip]
+
+	movsd	xmm2,xmm1				# x
+	mulsd	xmm1,xmm0				# x3
+	
+	movsd	xmm3,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+#	ucomisd	xmm3,qword ptr real_0_25 # x2>=0.25 (|x|>0.5)
+
+	mulsd	xmm3,xmm1				# x5
+
+	movlpd	xmm7,qword ptr sin_p_2[rip]
+	movlpd	xmm8,qword ptr sin_p_3[rip]
+	
+	mulsd	xmm5,xmm1				# p3
+	mulsd	xmm1,xmm0				# x7
+
+	mulsd	xmm6,xmm3				# p5
+	mulsd	xmm3,xmm0				# x9
+
+	mulsd	xmm7,xmm1				# p7
+	mulsd	xmm1,xmm0				# x11
+	
+	mulsd	xmm8,xmm3				# p9
+	mulsd	xmm3,xmm0				# x13
+
+	mulsd	xmm1,qword ptr sin_p_4[rip]	# p11
+	mulsd	xmm3,qword ptr sin_p_5[rip]	# p13
+	
+	movsd	xmm0,xmm2				# x
+
+	addsd	xmm1,xmm3				# p11+p13
+	addsd	xmm1,xmm8				# p9+p11+p13
+	
+#	jae		sin_real_1_l
+
+	addsd	xmm1,xmm7				# p7+p9+p11+p13
+	addsd	xmm1,xmm6				# p5+p7+p9+p11+p13
+	addsd	xmm1,xmm5				# p3+p5+p7+p9+p11+p13
+	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+sin_real_0_0:
+	ret
+
+#sin_real_1_l:
+#	movlpd	xmm2,qword ptr real_47_bits
+#	andpd	xmm2,xmm5				# p3h
+
+#	addsd	xmm1,xmm7				# p7+p9+p11+p13
+
+#	subsd	xmm5,xmm2				# p3l
+#	addsd	xmm0,xmm2				# x+p3h
+
+#	addsd	xmm1,xmm6				# p5+p7+p9+p11+p13
+#	addsd	xmm1,xmm5				# p3l+p5+p7+p9+p11+p13
+#	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+#	ret
+
+sin_real_m_0_6:
+	movlpd	xmm1,qword ptr real_m_0_600000000082499762577[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	lea		rcx,sin_p3_c[rip]
+	lea		rdx,sin_p4_c[rip]
+
+	cmova	rcx,rdx
+	
+	jmp		sin_real_0_6
+
+sin_real_p_0_6:
+	movlpd	xmm1,qword ptr real_0_600000000082499762577[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	lea		rcx,sin_p1_c[rip]
+	lea		rdx,sin_p2_c[rip]
+
+	cmovb	rcx,rdx
+
+sin_real_0_6:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm7,qword ptr (sin_p1_3-sin_p1_c)[rcx]
+	movlpd	xmm8,qword ptr (sin_p1_2-sin_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm7,xmm2				# p3
+	mulsd	xmm8,xmm2				# p2
+
+	movlpd	xmm9,qword ptr (sin_p1_5-sin_p1_c)[rcx]
+
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+	
+	mulsd	xmm9,xmm4				# p5
+	mulsd	xmm4,qword ptr (sin_p1_4-sin_p1_c)[rcx] # p4
+
+	movlpd	xmm11,qword ptr (sin_p1_7-sin_p1_c)[rcx]
+	movlpd	xmm12,qword ptr (sin_p1_9-sin_p1_c)[rcx]
+
+	mulsd	xmm11,xmm2				# p7
+	mulsd	xmm2,qword ptr (sin_p1_6-sin_p1_c)[rcx] # p6
+
+	mulsd	xmm12,xmm0				# p9
+	mulsd	xmm0,qword ptr (sin_p1_8-sin_p1_c)[rcx] # p8
+		
+	addsd	xmm11,xmm12				# p9+p7
+	addsd	xmm0,xmm2				# p8+p6
+
+	movlpd	xmm2,qword ptr real_26_bits[rip]
+	movsd	xmm3,xmm1				# x
+	movlpd	xmm6,qword ptr (sin_p1_1l-sin_p1_c)[rcx]
+	movlpd	xmm5,qword ptr (sin_p1_1h-sin_p1_c)[rcx]
+
+	addsd	xmm9,xmm11				# p9+p7+p5
+	addsd	xmm0,xmm4				# p8+p6+p4
+
+	andpd	xmm2,xmm1				# x_h
+	mulsd	xmm6,xmm1				# x*c1l
+	movlpd	xmm13,qword ptr (sin_p1_0h-sin_p1_c)[rcx]
+
+	addsd	xmm7,xmm9				# p9+p7+p5+p3
+	addsd	xmm0,xmm8				# p8+p6+p4+p2
+
+	subsd	xmm3,xmm2				# x_l
+	mulsd	xmm2,xmm5				# x_h*c1h
+
+	mulsd	xmm7,xmm1				# (p9+p7+p5+p3)(*x)
+
+	movsd	xmm14,xmm13
+	addsd	xmm13,xmm2				# x_h*c1h+c0h
+	
+	addsd	xmm0,xmm7				# p9+p8+..+p2
+
+	subsd	xmm14,xmm13				# c0h-(x_h*c1h+c0h)
+	mulsd	xmm3,xmm5				# x_l*c1h
+
+#	addsd	xmm0,qword ptr (sin_p1_0l-sin_p1_c)[rcx]
+	
+	addsd	xmm14,xmm2				# (c0h-(x_h*c1h+c0h))+x_h*c1h
+	addsd	xmm3,xmm6				# x_l*c1h+x*c1l
+	
+	addsd	xmm3,xmm14				# (c0h-(x_h*c1h+c0h))+x_h*c1h+x_l*c1h+x*c1l
+
+	addsd	xmm0,xmm3
+	addsd	xmm0,xmm13
+	ret
+
+sin_real_n_5:
+	addsd	xmm0,qword ptr real_2_pi[rip]
+	movlpd	xmm13,qword ptr real_m_2_pi_l[rip]
+	att_jmp		sin_real_p
+
+sin_real_5:
+	subsd	xmm0,qword ptr real_2_pi[rip]
+	movlpd	xmm13,qword ptr real_2_pi_l[rip]
+	att_jmp		sin_real_p
+
+cos_real_n_2:
+	addsd	xmm0,qword ptr real_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_m_pi_d_2_l[rip]
+	att_jmp		sin_real_p
+
+cos_real_4:
+	subsd	xmm0,qword ptr real_3_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_3_pi_d_2_l[rip]
+
+sin_real_p:
+	ucomisd	xmm0,qword ptr real_0_43540000008249979402[rip]
+
+	movsd	xmm12,xmm0
+
+	jae		sin_real_p_p_0_6
+	ucomisd	xmm0,qword ptr real_m_0_43540000008249979402[rip]
+	jb		sin_real_p_m_0_6
+
+	subsd	xmm0,xmm13
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	subsd	xmm12,xmm1
+
+	movlpd	xmm5,qword ptr sin_p_0[rip]
+	movlpd	xmm6,qword ptr sin_p_1[rip]
+
+	movsd	xmm2,xmm1				# x
+	mulsd	xmm1,xmm0				# x3
+	
+	movsd	xmm3,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	subsd	xmm12,xmm13				# x_l
+	movsd	xmm14,xmm3				# x2
+
+#	ucomisd	xmm3,qword ptr real_0_25 # x2>=0.25 (|x|>0.5)
+
+	mulsd	xmm3,xmm1				# x5
+
+	movlpd	xmm7,qword ptr sin_p_2[rip]
+	movlpd	xmm8,qword ptr sin_p_3[rip]
+		
+	mulsd	xmm5,xmm1				# p3
+	mulsd	xmm1,xmm0				# x7
+
+	mulsd	xmm6,xmm3				# p5
+	mulsd	xmm3,xmm0				# x9
+
+	mulsd	xmm7,xmm1				# p7
+	mulsd	xmm1,xmm0				# x11
+	
+	mulsd	xmm8,xmm3				# p9
+	mulsd	xmm3,xmm0				# x13
+
+	mulsd	xmm14,xmm12				# x_l*x2
+
+	mulsd	xmm1,qword ptr sin_p_4[rip]	# p11
+	mulsd	xmm3,qword ptr sin_p_5[rip]	# p13
+	
+	movsd	xmm0,xmm2				# x
+
+	mulsd	xmm14,qword ptr real_0_5[rip] # 0.5*x_l*x2
+
+	addsd	xmm1,xmm3				# p11+p13
+
+	subsd	xmm12,xmm14				# x_l-0.5*x_l*x2
+
+	addsd	xmm1,xmm8				# p9+p11+p13
+
+#	jae		sin_real_p_l
+
+	addsd	xmm1,xmm7				# p7+p9+p11+p13
+	addsd	xmm1,xmm6				# p5+p7+p9+p11+p13
+	addsd	xmm1,xmm5				# p3+p5+p7+p9+p11+p13
+
+	addsd	xmm1,xmm12
+
+	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+	ret
+
+#sin_real_p_l:
+#	movlpd	xmm2,qword ptr real_47_bits
+#	andpd	xmm2,xmm5				# p3h
+
+#	addsd	xmm1,xmm7				# p7+p9+p11+p13
+
+#	subsd	xmm5,xmm2				# p3l
+#	addsd	xmm0,xmm2				# x+p3h
+
+#	addsd	xmm1,xmm6				# p5+p7+p9+p11+p13
+#	addsd	xmm1,xmm5				# p3+p5+p7+p9+p11+p13
+
+#	addsd	xmm1,xmm12
+
+#	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+#	ret
+
+sin_real_p_m_0_6:
+	movlpd	xmm1,qword ptr real_m_0_600000000082499762577[rip]
+
+	subsd	xmm0,xmm13
+
+	ucomisd	xmm0,xmm1
+	movsd	xmm15,xmm0				# x1
+	subsd	xmm0,xmm1
+
+	lea		rcx,sin_p3_c[rip]
+	lea		rdx,sin_p4_c[rip]
+
+	cmova	rcx,rdx
+	jmp		sin_real_p_pm_0_6
+
+sin_real_p_p_0_6:
+	movlpd	xmm1,qword ptr real_0_600000000082499762577[rip]
+
+	subsd	xmm0,xmm13
+
+	ucomisd	xmm0,xmm1
+	movsd	xmm15,xmm0				# x1
+	subsd	xmm0,xmm1
+
+	lea		rcx,sin_p1_c[rip]
+	lea		rdx,sin_p2_c[rip]
+
+	cmovb	rcx,rdx
+
+sin_real_p_pm_0_6:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+	subsd	xmm12,xmm15
+	mulsd	xmm15,xmm15				# x1*x1
+
+	movlpd	xmm7,qword ptr (sin_p1_3-sin_p1_c)[rcx]
+	movlpd	xmm8,qword ptr (sin_p1_2-sin_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm7,xmm2				# p3
+	mulsd	xmm8,xmm2				# p2
+	subsd	xmm12,xmm13				# x_l
+
+sin_real_pm_0_6:
+	mulsd	xmm15,qword ptr real_0_5[rip] # 0.5*x1*x1
+
+	movlpd	xmm9,qword ptr (sin_p1_5-sin_p1_c)[rcx]
+
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+	
+	mulsd	xmm9,xmm4				# p5
+	mulsd	xmm4,qword ptr (sin_p1_4-sin_p1_c)[rcx] # p4
+
+	mulsd	xmm15,xmm12				# x_l*0.5*x1*x1
+
+	movlpd	xmm11,qword ptr (sin_p1_7-sin_p1_c)[rcx]
+	movlpd	xmm10,qword ptr (sin_p1_9-sin_p1_c)[rcx]
+
+	mulsd	xmm11,xmm2				# p7
+	mulsd	xmm2,qword ptr (sin_p1_6-sin_p1_c)[rcx] # p6
+
+	mulsd	xmm10,xmm0				# p9
+	mulsd	xmm0,qword ptr (sin_p1_8-sin_p1_c)[rcx] # p8
+
+	addsd	xmm11,xmm10				# p9+p7
+	addsd	xmm0,xmm2				# p8+p6
+
+	subsd	xmm12,xmm15				# x_l-x_l*0.5*x1*x1
+
+	movlpd	xmm2,qword ptr real_26_bits[rip]
+	movsd	xmm3,xmm1				# x
+	movlpd	xmm6,qword ptr (sin_p1_1l-sin_p1_c)[rcx]
+	movlpd	xmm5,qword ptr (sin_p1_1h-sin_p1_c)[rcx]
+
+	addsd	xmm9,xmm11				# p9+p7+p5
+	addsd	xmm0,xmm4				# p8+p6+p4
+
+	andpd	xmm2,xmm1				# x_h
+	mulsd	xmm6,xmm1				# x*c1l
+	movlpd	xmm13,qword ptr (sin_p1_0h-sin_p1_c)[rcx]
+
+	addsd	xmm7,xmm9				# p9+p7+p5+p3
+	addsd	xmm0,xmm8				# p8+p6+p4+p2
+
+	subsd	xmm3,xmm2				# x_l
+	mulsd	xmm2,xmm5				# x_h*c1h
+
+	mulsd	xmm7,xmm1				# (p9+p7+p5+p3)(*x)
+
+	movsd	xmm14,xmm13
+	addsd	xmm13,xmm2				# x_h*c1h+c0h
+	
+	addsd	xmm0,xmm7				# p9+p8+..+p2
+
+	subsd	xmm14,xmm13				# c0h-(x_h*c1h+c0h)
+	mulsd	xmm3,xmm5				# x_l*c1h
+
+#	addsd	xmm0,qword ptr (sin_p1_0l-sin_p1_c)[rcx]
+	addsd	xmm0,xmm12
+	
+	addsd	xmm14,xmm2				# (c0h-(x_h*c1h+c0h))+x_h*c1h
+	addsd	xmm3,xmm6				# x_l*c1h+x*c1l
+	
+	addsd	xmm3,xmm14				# (c0h-(x_h*c1h+c0h))+x_h*c1h+x_l*c1h+x*c1l
+
+	addsd	xmm0,xmm3
+	addsd	xmm0,xmm13
+	ret
+
+sin_real_3:
+	movlpd	xmm1,qword ptr real_pi[rip]
+	movlpd	xmm13,qword ptr real_pi_l[rip]
+	subsd	xmm1,xmm0
+	att_jmp		sin_real_m
+
+cos_real_n_4:
+	movlpd	xmm1,qword ptr real_m_3_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_m_3_pi_d_2_l[rip]
+	subsd	xmm1,xmm0
+	att_jmp		sin_real_m
+
+sin_real_n_3:
+	movlpd	xmm1,qword ptr real_m_pi[rip]
+	movlpd	xmm13,qword ptr real_m_pi_l[rip]
+	subsd	xmm1,xmm0
+	att_jmp		sin_real_m
+
+cos_real_2:
+	movlpd	xmm1,qword ptr real_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_pi_d_2_l[rip]
+	subsd	xmm1,xmm0
+
+sin_real_m:
+	ucomisd	xmm1,qword ptr real_0_43540000008249979402[rip]
+
+	movsd	xmm12,xmm1
+
+	jae		sin_real_m_p_0_6
+	ucomisd	xmm1,qword ptr real_m_0_43540000008249979402[rip]
+	jbe		sin_real_m_m_0_6
+
+	addsd	xmm1,xmm13
+
+	movsd	xmm0,xmm1				# x
+	mulsd	xmm1,xmm1				# x2
+
+	subsd	xmm12,xmm0
+
+	movlpd	xmm5,qword ptr sin_p_0[rip]
+	movlpd	xmm6,qword ptr sin_p_1[rip]
+
+	movsd	xmm2,xmm0				# x
+	mulsd	xmm0,xmm1				# x3
+
+#	ucomisd	xmm1,qword ptr real_0_25 # x2>=0.25 (|x|>0.5)
+	
+	movsd	xmm3,xmm1				# x2
+	mulsd	xmm1,xmm1				# x4
+
+	addsd	xmm12,xmm13
+
+	movsd	xmm14,xmm3
+	mulsd	xmm3,xmm0				# x5
+
+	movlpd	xmm7,qword ptr sin_p_2[rip]
+	movlpd	xmm8,qword ptr sin_p_3[rip]
+		
+	mulsd	xmm5,xmm0				# p3
+	mulsd	xmm0,xmm1				# x7
+
+	mulsd	xmm6,xmm3				# p5
+	mulsd	xmm3,xmm1				# x9
+
+	mulsd	xmm7,xmm0				# p7
+	mulsd	xmm0,xmm1				# x11
+	
+	mulsd	xmm8,xmm3				# p9
+	mulsd	xmm3,xmm1				# x13
+
+	mulsd	xmm14,xmm12
+
+	mulsd	xmm0,qword ptr sin_p_4[rip]	# p11
+	mulsd	xmm3,qword ptr sin_p_5[rip]	# p13
+	
+	movsd	xmm1,xmm2				# x
+
+	mulsd	xmm14,qword ptr real_0_5[rip]
+
+	addsd	xmm0,xmm3				# p11+p13
+
+	subsd	xmm12,xmm14
+
+	addsd	xmm0,xmm8				# p9+p11+p13
+
+#	jae		sin_real_m_l
+
+	addsd	xmm0,xmm7				# p7+p9+p11+p13
+	addsd	xmm0,xmm6				# p5+p7+p9+p11+p13
+	addsd	xmm0,xmm5				# p3+p5+p7+p9+p11+p13
+
+	addsd	xmm0,xmm12
+
+	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+	ret
+
+#sin_real_m_l:
+#	movlpd	xmm2,qword ptr real_47_bits
+#	andpd	xmm2,xmm5				# p3h
+
+#	addsd	xmm0,xmm7				# p7+p9+p11+p13
+
+#	subsd	xmm5,xmm2				# p3l
+#	addsd	xmm1,xmm2				# x+p3h
+
+#	addsd	xmm0,xmm6				# p5+p7+p9+p11+p13
+#	addsd	xmm0,xmm5				# p3+p5+p7+p9+p11+p13
+
+#	addsd	xmm0,xmm12
+
+#	addsd	xmm0,xmm1				# x+p3+p5+p7+p9+p11+p13
+
+#	ret
+
+sin_real_m_m_0_6:
+	addsd	xmm1,xmm13
+
+	movlpd	xmm0,qword ptr real_0_600000000082499762577[rip]
+
+	ucomisd	xmm1,qword ptr real_m_0_600000000082499762577[rip]
+	movsd	xmm15,xmm1				# x1
+	addsd	xmm0,xmm1
+
+	lea		rcx,sin_p3_c[rip]
+	lea		rdx,sin_p4_c[rip]
+
+	cmova	rcx,rdx
+	jmp		sin_real_m_pm_0_6
+
+sin_real_m_p_0_6:
+	addsd	xmm1,xmm13
+
+	movlpd	xmm0,qword ptr real_m_0_600000000082499762577[rip]
+
+	ucomisd	xmm1,qword ptr real_0_600000000082499762577[rip]
+	movsd	xmm15,xmm1				# x1
+	addsd	xmm0,xmm1
+
+	lea		rcx,sin_p1_c[rip]
+	lea		rdx,sin_p2_c[rip]
+
+	cmovb	rcx,rdx
+
+sin_real_m_pm_0_6:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+	subsd	xmm12,xmm15
+	mulsd	xmm15,xmm15				# x1*x1
+
+	movlpd	xmm7,qword ptr (sin_p1_3-sin_p1_c)[rcx]
+	movlpd	xmm8,qword ptr (sin_p1_2-sin_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm7,xmm2				# p3
+	mulsd	xmm8,xmm2				# p2
+
+	addsd	xmm12,xmm13				# x_l
+	att_jmp		sin_real_pm_0_6
+
+	.globl	cos_real
+
+cos_real:
+	ucomisd	xmm0,qword ptr real_pi_d_4[rip]
+	jbe		cos_real_1				# x<=pi/4 | NAN
+	ucomisd	xmm0,qword ptr real_3_pi_d_4[rip]
+	att_jbe		cos_real_2
+	ucomisd	xmm0,qword ptr real_5_pi_d_4[rip]
+	jbe		cos_real_3
+	ucomisd	xmm0,qword ptr real_7_pi_d_4[rip]
+	att_jbe		cos_real_4
+	ucomisd	xmm0,qword ptr real_9_pi_d_4[rip]
+	jbe		cos_real_5
+
+	ucomisd	xmm0,qword ptr real_36825084_pi[rip]
+	jbe		cos_real_0
+
+	ucomisd	xmm0,qword ptr real_2_p_53[rip]
+	att_jae		sin_cos_or_tan_real_too_large	# x>=-2^53
+
+	att_call	rem_36825084_pi
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+	subsd	xmm13,xmm7
+	jmp		cos_real_0_
+
+cos_real_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+	
+	addsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi+1
+	
+	movq	xmm4,qword ptr round_even_c[rip]
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi+1)
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm6,xmm4
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+cos_real_0_:
+	ucomisd	xmm1,xmm5
+	jne		cos_real_sin
+
+cos_real_cos:
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+
+	ucomisd	xmm1,xmm6
+	att_je		cos_real_p
+	att_jmp		cos_real_m
+
+cos_real_sin:
+	ucomisd	xmm1,xmm6
+	jne		cos_real_p_sin
+
+	movsd	xmm1,xmm3
+	subsd	xmm0,xmm2
+	subsd	xmm1,xmm0
+	att_jmp		sin_real_m
+
+cos_real_p_sin:
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+	att_jmp		sin_real_p
+
+cos_real_n:
+	ucomisd	xmm0,qword ptr real_m_3_pi_d_4[rip]
+	att_jae		cos_real_n_2
+	ucomisd	xmm0,qword ptr real_m_5_pi_d_4[rip]
+	jae		cos_real_n_3
+	ucomisd	xmm0,qword ptr real_m_7_pi_d_4[rip]
+	att_jae		cos_real_n_4
+	ucomisd	xmm0,qword ptr real_m_9_pi_d_4[rip]
+	jae		cos_real_n_5
+
+	ucomisd	xmm0,qword ptr real_m_36825084_pi[rip]
+	jae		cos_real_n_0
+
+	ucomisd	xmm0,qword ptr real_m_2_p_53[rip]
+	att_jbe		sin_cos_or_tan_real_too_small_or_nan	# x<=-2^53 | NAN
+
+	att_call	rem_n_36825084_pi
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+	subsd	xmm13,xmm7
+	jmp		cos_real_n_0_
+
+cos_real_n_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+	
+	movq	xmm4,qword ptr m_round_even_c[rip]
+
+	subsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi-1
+	
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi-1)
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm13,qword ptr real_pi_d_4_52_l[rip]
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm6,xmm4
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm13,xmm1
+
+	andpd	xmm5,xmm1
+	andpd	xmm6,xmm1
+
+cos_real_n_0_:
+	ucomisd	xmm1,xmm5
+	att_je		cos_real_cos
+
+	ucomisd	xmm1,xmm6
+	je		cos_real_n_p_sin
+
+	movsd	xmm1,xmm3
+	subsd	xmm0,xmm2
+	subsd	xmm1,xmm0
+	att_jmp		sin_real_m
+
+cos_real_n_p_sin:
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+	att_jmp		sin_real_p
+
+cos_real_1:
+	ucomisd	xmm0,qword ptr real_m_pi_d_4[rip]
+	att_jb		cos_real_n				# x<-pi/4 | NAN
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm9,qword ptr real_17_bits[rip]
+	movlpd	xmm5,qword ptr cos_p_0[rip]
+
+	movsd	xmm8,xmm1				# x
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	andpd	xmm9,xmm1				# x_17_h
+	movlpd	xmm10,qword ptr real_0_5[rip]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movlpd	xmm7,qword ptr cos_p_2[rip]
+
+	mulsd	xmm5,xmm4				# p4
+	mulsd	xmm4,xmm2				# x10
+
+	movsd	xmm6,xmm2				# x6
+	mulsd	xmm2,xmm2				# x12
+
+	mulsd	xmm7,xmm0				# p8
+	mulsd	xmm0,xmm6				# x14
+
+	mulsd	xmm6,qword ptr cos_p_1[rip]	# p6	
+	mulsd	xmm4,qword ptr cos_p_3[rip]	# p10
+	mulsd	xmm2,qword ptr cos_p_4[rip]	# p12
+	mulsd	xmm0,qword ptr cos_p_5[rip]	# p14
+
+	subsd	xmm8,xmm9				# x_17_l
+	mulsd	xmm10,xmm9				# 0.5*x_17_h
+
+	addsd	xmm0,xmm2				# p12+p14
+	
+	mulsd	xmm10,xmm9				# 0.5*x_17_h*x_17_h
+	movlpd	xmm11,qword ptr real_1_0[rip]
+
+	addsd	xmm0,xmm4				# p10+p12+p14
+
+	mulsd	xmm8,qword ptr real_0_5[rip]	# 0.5*x_17_l
+	addsd	xmm9,xmm1				# x+x_17_h
+
+	addsd	xmm0,xmm7				# p8+p10+p12+p14
+
+	mulsd	xmm8,xmm9				# 0.5*x_17_l*(x+x_17_h)
+
+	addsd	xmm0,xmm6				# p6+p8+p10+p12+p14
+
+	ucomisd	xmm10,qword ptr real_2_p_m_18[rip] # 0.5*x_17_h*x_17_h<2**-18
+	jb		cos_real_1_s
+
+	subsd	xmm11,xmm10				# 1.0-0.5*x_17_h*x_17_h
+
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	subsd	xmm0,xmm8
+	addsd	xmm0,xmm11
+
+	ret
+
+cos_real_1_s:
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	subsd	xmm0,xmm8
+	subsd	xmm0,xmm10
+	addsd	xmm0,xmm11
+
+	ret
+
+cos_real_n_5:
+	addsd	xmm0,qword ptr real_2_pi[rip]
+	movlpd	xmm13,qword ptr real_m_2_pi_l[rip]
+	att_jmp		cos_real_p
+
+cos_real_5:
+	subsd	xmm0,qword ptr real_2_pi[rip]
+	movlpd	xmm13,qword ptr real_2_pi_l[rip]
+	att_jmp		cos_real_p
+
+sin_real_n_4:
+	addsd	xmm0,qword ptr real_3_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_m_3_pi_d_2_l[rip]
+	att_jmp		cos_real_p
+
+sin_real_2:
+	subsd	xmm0,qword ptr real_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_pi_d_2_l[rip]
+
+cos_real_p:
+	movsd	xmm12,xmm0
+	subsd	xmm0,xmm13
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	subsd	xmm12,xmm1
+
+	movlpd	xmm9,qword ptr real_17_bits[rip]
+	movlpd	xmm5,qword ptr cos_p_0[rip]
+
+	movsd	xmm8,xmm1				# x
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	subsd	xmm12,xmm13
+
+	andpd	xmm9,xmm1				# x_17_h
+	movlpd	xmm10,qword ptr real_0_5[rip]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movlpd	xmm7,qword ptr cos_p_2[rip]
+
+	mulsd	xmm12,xmm8
+
+	mulsd	xmm5,xmm4				# p4
+	mulsd	xmm4,xmm2				# x10
+
+	movsd	xmm6,xmm2				# x6
+	mulsd	xmm2,xmm2				# x12
+
+	mulsd	xmm7,xmm0				# p8
+	mulsd	xmm0,xmm6				# x14
+
+	mulsd	xmm6,qword ptr cos_p_1[rip]	# p6	
+	mulsd	xmm4,qword ptr cos_p_3[rip]	# p10
+	mulsd	xmm2,qword ptr cos_p_4[rip]	# p12
+	mulsd	xmm0,qword ptr cos_p_5[rip]	# p14
+
+	subsd	xmm8,xmm9				# x_17_l
+	mulsd	xmm10,xmm9				# 0.5*x_17_h
+
+	addsd	xmm0,xmm2				# p12+p14
+	
+	mulsd	xmm10,xmm9				# 0.5*x_17_h*x_17_h
+	movlpd	xmm11,qword ptr real_1_0[rip]
+
+	addsd	xmm0,xmm4				# p10+p12+p14
+
+	mulsd	xmm8,qword ptr real_0_5[rip]	# 0.5*x_17_l
+	addsd	xmm9,xmm1				# x+x_17_h
+
+	addsd	xmm0,xmm7				# p8+p10+p12+p14
+
+	mulsd	xmm8,xmm9				# 0.5*x_17_l*(x+x_17_h)
+
+	addsd	xmm0,xmm6				# p6+p8+p10+p12+p14
+
+	ucomisd	xmm10,qword ptr real_2_p_m_18[rip] # 0.5*x_17_h*x_17_h<2**-18
+	jb		cos_real__s
+
+	subsd	xmm11,xmm10				# 1.0-0.5*x_17_h*x_17_h
+
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	addsd	xmm8,xmm12
+
+	subsd	xmm0,xmm8
+	addsd	xmm0,xmm11
+
+	ret
+
+cos_real__s:
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	addsd	xmm8,xmm12
+
+	subsd	xmm0,xmm8
+	subsd	xmm0,xmm10
+	addsd	xmm0,xmm11
+
+	ret
+
+sin_real_4:
+	subsd	xmm0,qword ptr real_3_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_3_pi_d_2_l[rip]
+	att_jmp		cos_real_m
+
+cos_real_n_3:
+	addsd	xmm0,qword ptr real_pi[rip]
+	movlpd	xmm13,qword ptr real_m_pi_l[rip]
+	att_jmp		cos_real_m
+
+sin_real_n_2:
+	addsd	xmm0,qword ptr real_pi_d_2[rip]
+	movlpd	xmm13,qword ptr real_m_pi_d_2_l[rip]
+	att_jmp		cos_real_m
+
+cos_real_3:
+	subsd	xmm0,qword ptr real_pi[rip]
+	movlpd	xmm13,qword ptr real_pi_l[rip]
+
+cos_real_m:
+	movsd	xmm12,xmm0
+	subsd	xmm0,xmm13
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	subsd	xmm12,xmm1
+
+	movlpd	xmm9,qword ptr real_17_bits[rip]
+	movlpd	xmm5,qword ptr cos_p_0[rip]
+
+	movsd	xmm8,xmm1				# x
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	subsd	xmm12,xmm13
+
+	andpd	xmm9,xmm1				# x_17_h
+	movlpd	xmm10,qword ptr real_0_5[rip]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movlpd	xmm7,qword ptr cos_p_2[rip]
+
+	mulsd	xmm12,xmm8
+
+	mulsd	xmm5,xmm4				# p4
+	mulsd	xmm4,xmm2				# x10
+
+	movsd	xmm6,xmm2				# x6
+	mulsd	xmm2,xmm2				# x12
+
+	mulsd	xmm7,xmm0				# p8
+	mulsd	xmm0,xmm6				# x14
+
+	mulsd	xmm6,qword ptr cos_p_1[rip]	# p6	
+	mulsd	xmm4,qword ptr cos_p_3[rip]	# p10
+	mulsd	xmm2,qword ptr cos_p_4[rip]	# p12
+	mulsd	xmm0,qword ptr cos_p_5[rip]	# p14
+
+	subsd	xmm8,xmm9				# x_17_l
+	mulsd	xmm10,xmm9				# 0.5*x_17_h
+
+	addsd	xmm0,xmm2				# p12+p14
+	
+	mulsd	xmm10,xmm9				# 0.5*x_17_h*x_17_h
+	movlpd	xmm11,qword ptr real_m_1_0[rip]
+
+	addsd	xmm0,xmm4				# p10+p12+p14
+
+	mulsd	xmm8,qword ptr real_0_5[rip]	# 0.5*x_17_l
+	addsd	xmm9,xmm1				# x+x_17_h
+
+	addsd	xmm0,xmm7				# p8+p10+p12+p14
+
+	mulsd	xmm8,xmm9				# 0.5*x_17_l*(x+x_17_h)
+
+	addsd	xmm0,xmm6				# p6+p8+p10+p12+p14
+	
+	ucomisd	xmm10,qword ptr real_2_p_m_18[rip] # 0.5*x_17_h*x_17_h<2**-18
+	jb		cos_real_m_s
+
+	addsd	xmm11,xmm10				# -1.0+0.5*x_17_h*x_17_h
+
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	addsd	xmm8,xmm12
+
+	subsd	xmm8,xmm0
+	movsd	xmm0,xmm11
+	addsd	xmm0,xmm8
+	ret
+
+cos_real_m_s:
+	addsd	xmm0,xmm5				# p4+p6+p8+p10+p12+p14
+
+	addsd	xmm8,xmm12
+
+	subsd	xmm8,xmm0
+
+	movsd	xmm0,xmm11
+	addsd	xmm8,xmm10
+
+	addsd	xmm0,xmm8
+	ret
+
+	.globl	tan_real
+
+tan_real:
+	ucomisd	xmm0,qword ptr real_m_0_338[rip]
+	jb		tan_real_n				# x<-0.338 | NAN
+	ucomisd	xmm0,qword ptr real_0_89[rip]
+	jbe		tan_real_1
+	ucomisd	xmm0,qword ptr real_pi_m_0_699[rip]
+	jbe		tan_real_2
+	ucomisd	xmm0,qword ptr real_pi_p_0_89[rip]
+	jbe		tan_real_3
+	ucomisd	xmm0,qword ptr real_2pi_m_0_699[rip]
+	jbe		tan_real_4
+	ucomisd	xmm0,qword ptr real_2pi_p_0_89[rip]
+	jbe		tan_real_5
+
+	ucomisd	xmm0,qword ptr real_36825084_pi[rip]
+	jbe		tan_real_0
+
+	ucomisd	xmm0,qword ptr real_2_p_53[rip]
+	att_jae		sin_cos_or_tan_real_too_large	# x>=-2^53
+
+	att_call	rem_36825084_pi
+
+tan_real_pn_l:
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm12,qword ptr real_pi_d_4_52_l[rip]
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm12,xmm1
+
+	subsd	xmm12,xmm7
+	jmp		tan_real_
+
+tan_real_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+	
+	addsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi+1
+	
+	movq	xmm4,qword ptr round_even_c[rip]
+tan_real_pn_0:
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi+1)
+
+	movlpd	xmm2,qword ptr real_pi_d_4_26[rip]
+	movlpd	xmm3,qword ptr real_pi_d_4_26_26[rip]
+	movlpd	xmm12,qword ptr real_pi_d_4_52_l[rip]
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+
+	mulsd	xmm2,xmm1
+	mulsd	xmm3,xmm1
+	mulsd	xmm12,xmm1
+
+tan_real_:
+	andpd	xmm5,xmm1
+	ucomisd	xmm1,xmm5
+	jne		tan_real_0_2
+
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+	
+	ucomisd	xmm0,qword ptr real_0_338[rip]
+	ja		tan_real_0_1p
+	ucomisd	xmm0,qword ptr real_m_0_338[rip]
+	jae		tan_real_3_
+
+	ucomisd	xmm0,qword ptr real_m_0_542[rip]
+	ja		tan_real_0_5_n
+	ucomisd	xmm0,qword ptr real_m_0_699[rip]
+	ja		tan_real_s_0_5_n
+	jmp		tan_real_1_0_3_n
+
+tan_real_0_1p:
+	ucomisd	xmm0,qword ptr real_0_542[rip]
+	jb		tan_real_0_5_p
+	ucomisd	xmm0,qword ptr real_0_699[rip]
+	jb		tan_real_s_0_5_p
+	jmp		tan_real_1_0_3
+
+tan_real_0_2:
+	subsd	xmm2,xmm0
+	movsd	xmm1,xmm3
+	addsd	xmm1,xmm2
+
+	ucomisd	xmm1,qword ptr real_0_5_pi_m_1_18[rip]
+	ja		tan_real_0_2p
+	ucomisd	xmm1,qword ptr real_1_18_m_0_5_pi[rip]
+	jae		tan_real_2_
+
+	ucomisd	xmm1,qword ptr real_1_04_m_0_5_pi[rip]
+	ja		tan_real_2_0_n
+	ucomisd	xmm1,qword ptr real_0_89_m_0_5_pi[rip]
+	ja		tan_real_s_2_n
+	jmp		tan_real_1_0_n
+
+tan_real_0_2p:
+	ucomisd	xmm1,qword ptr real_0_5_pi_m_1_04[rip]
+	jb		tan_real_2_0_p
+	ucomisd	xmm1,qword ptr real_0_5_pi_m_0_89[rip]
+	jb		tan_real_s_2_p
+	jmp		tan_real_1_0_2_p
+
+tan_real_n:
+	ucomisd	xmm0,qword ptr real_m_0_89[rip]
+	jae		tan_real_n_1
+	ucomisd	xmm0,qword ptr real_n_pi_m_0_699[rip]
+	jae		tan_real_n_2
+	ucomisd	xmm0,qword ptr real_n_pi_p_0_89[rip]
+	jae		tan_real_n_3
+	ucomisd	xmm0,qword ptr real_n_2pi_m_0_699[rip]
+	jae		tan_real_n_4
+	ucomisd	xmm0,qword ptr real_n_2pi_p_0_89[rip]
+	jae		tan_real_n_5
+
+	ucomisd	xmm0,qword ptr real_m_36825084_pi[rip]
+	jae		tan_real_n_0
+
+	ucomisd	xmm0,qword ptr real_m_2_p_53[rip]
+	att_jbe		sin_cos_or_tan_real_too_small_or_nan	# x<=-2^53 | NAN
+
+	att_call	rem_n_36825084_pi
+	att_jmp		tan_real_pn_l
+
+tan_real_n_0:
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+	
+	subsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi-1
+	
+	movq	xmm4,qword ptr m_round_even_c[rip]
+	att_jmp		tan_real_pn_0
+
+tan_real_1:
+	ucomisd	xmm0,qword ptr real_0_699[rip]
+	ja		tan_real_1_2
+	ucomisd	xmm0,qword ptr real_0_338[rip]
+	ja		tan_real_0_5
+
+tan_real_1_0:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm6,qword ptr tan_q_1[rip]
+	movlpd	xmm5,qword ptr tan_q_2[rip]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm7,qword ptr tan_p_1[rip]
+	mulsd	xmm6,xmm2				# q2
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	mulsd	xmm5,xmm4				# q4
+	mulsd	xmm2,qword ptr tan_q_3[rip]	# q6
+	
+	mulsd	xmm4,qword ptr tan_p_2[rip]	# p4
+
+	addsd	xmm0,xmm2				# x8+q6
+
+	mulsd	xmm7,xmm3				# p2
+
+	addsd	xmm0,xmm5				# x8+q6+q4
+
+	addsd	xmm4,xmm7				# p4+p2
+
+	addsd	xmm0,xmm6				# x8+q6+q4+q2
+
+	addsd	xmm4,qword ptr tan_p_0[rip]	# p4+p2+p0
+
+	addsd	xmm0,qword ptr tan_q_0[rip]	# x8+q6+q4+q2+q0
+
+	divsd	xmm4,xmm0				# p/q
+
+	mulsd	xmm3,xmm1				# x3
+	movsd	xmm0,xmm1				# x
+
+	mulsd	xmm3,xmm4
+	addsd	xmm0,xmm3
+
+	ret
+
+tan_real_m_0_5:
+	ucomisd	xmm0,qword ptr real_m_0_542[rip]
+	jb		tan_real_n_s_0_5
+
+	movlpd	xmm1,qword ptr atan_0_5_53[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	lea		rdx,tan_n_0_5_t[rip]
+	addsd	xmm1,qword ptr atan_0_5_53_l[rip] # y=x+atan 0.5
+	jmp		tan_real_0_5_a_s_0_5_a_s_2
+
+tan_real_n_s_0_5:
+	movlpd	xmm1,qword ptr atan_sqrt_0_5[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+	
+	lea		rdx,tan_n_s_0_5_t[rip]
+	addsd	xmm1,qword ptr atan_sqrt_0_5_l[rip] # y=x+atan sqrt 0.5
+	att_jmp		tan_real_0_5_a_s_0_5_a_s_2
+
+tan_real_0_5:
+	ucomisd	xmm0,qword ptr real_0_542[rip]
+	ja		tan_real_s_0_5
+
+	movlpd	xmm1,qword ptr m_atan_0_5_53[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	lea		rdx,tan_0_5_t[rip]
+	subsd	xmm1,qword ptr atan_0_5_53_l[rip] # y=x-atan 0.5
+	att_jmp		tan_real_0_5_a_s_0_5_a_s_2
+
+tan_real_s_0_5:
+	movlpd	xmm1,qword ptr m_atan_sqrt_0_5[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+	
+	lea		rdx,tan_s_0_5_t[rip]
+	subsd	xmm1,qword ptr atan_sqrt_0_5_l[rip] # y=x-atan sqrt 0.5
+
+tan_real_0_5_a_s_0_5_a_s_2:
+	subsd	xmm0,xmm10				# x_l
+
+	movsd	xmm2,xmm1				# y
+	mulsd	xmm1,xmm1				# y2
+
+	mulsd	xmm10,qword ptr [rdx]	# 1.25|1.5|3*x_h
+
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm1				# y2
+	mulsd	xmm1,xmm1				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm1				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm1				# y4
+	mulsd	xmm1,xmm1				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm1,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm1,xmm6				# y8+q6+q4
+
+	mulsd	xmm0,qword ptr [rdx]	# 1.25|1.5|3*x_l
+
+	movlpd	xmm6,qword ptr 8[rdx]	# 0.5|sqrt 0.5|sqrt 2
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm1,xmm7				# y8+q6+q4+q2
+
+	mulsd	xmm6,xmm2				# 0.5|sqrt 0.5|sqrt 2*y
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm1,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm1				# p+q
+	
+	mulsd	xmm4,xmm6				# 0.5|sqrt 0.5|sqrt 2*y*(p+q)
+
+	mulsd	xmm2,qword ptr [rdx]	# 1.25|1.5|3*y
+
+	addsd	xmm3,xmm4				# p+y*(p+q)
+	subsd	xmm1,xmm4				# q-y*(p+q)
+
+	divsd	xmm3,xmm1				# (p-y*(p+q))/(q+y*(p+q))
+	
+	addsd	xmm0,qword ptr 16[rdx]	# (0.5-1.25atan0.5)_49_l | (sqrt0.5-1.5atan(sqrt0.5))_l | (sqrt2-3atan(sqrt2))_l
+
+	mulsd	xmm2,xmm3				# 1.25|1.5|3 * y * (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm10,qword ptr 24[rdx]	# (0.5-1.25atan0.5)_49 | sqrt0.5-1.5atan(sqrt0.5) | sqrt2-3atan(sqrt2)
+
+	addsd	xmm0,xmm2
+
+	addsd	xmm0,xmm10				# 0.5|sqrt 0.5|sqrt 2+1.25|1.5|3y+1.25|1.5|3y * ..
+
+	ret
+
+tan_real_n_1:
+	ucomisd	xmm0,qword ptr real_m_0_699[rip]
+	att_ja		tan_real_m_0_5
+
+tan_real_n_1_2:
+	movlpd	xmm1,qword ptr real_pi_d_4[rip]
+
+	addsd	xmm1,xmm0
+	addsd	xmm0,xmm0				# 2x
+
+	addsd	xmm1,qword ptr real_pi_d_4_l[rip] # y=pi/4+x
+
+	addsd	xmm0,qword ptr real_pi_d_2_m_1_0_52[rip] # 1+2x-pi/2=1-2y
+	movlpd	xmm10,qword ptr real_m_pi_d_2_m_1_0_52_l[rip]
+	movlpd	xmm11,qword ptr real_m_2_0[rip]
+	jmp		tan_real_1_2_
+
+tan_real_1_2:
+	movlpd	xmm1,qword ptr real_pi_d_4[rip]
+
+	subsd	xmm1,xmm0
+	addsd	xmm0,xmm0				# 2x
+
+	addsd	xmm1,qword ptr real_pi_d_4_l[rip] # y=pi/4-x
+	subsd	xmm0,qword ptr real_pi_d_2_m_1_0_52[rip] # 1+2x-pi/2=1-2y
+	movlpd	xmm10,qword ptr real_pi_d_2_m_1_0_52_l[rip]
+	movlpd	xmm11,qword ptr real_2_0[rip]
+
+tan_real_1_2_:
+	movsd	xmm2,xmm1				# y
+	mulsd	xmm1,xmm1				# y2
+
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm1				# y2
+	mulsd	xmm1,xmm1				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm1				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm1				# y4
+	mulsd	xmm1,xmm1				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm1,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm1,xmm6				# y8+q6+q4
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm1,xmm7				# y8+q6+q4+q2
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm1,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm1				# p+q
+
+	mulsd	xmm4,xmm2				# y*(p+q)
+	mulsd	xmm2,xmm11				# 2y
+
+	subsd	xmm3,xmm4				# p-y*(p+q)
+	addsd	xmm1,xmm4				# q+y*(p+q)
+
+	divsd	xmm3,xmm1				# (p-y*(p+q))/(q+y*(p+q))
+	
+	mulsd	xmm2,xmm3				# 2y * (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm2,xmm10
+	
+	subsd	xmm0,xmm2				# 1-2y-2y * ..
+
+	ret
+
+tan_real_n_4:
+	ucomisd	xmm0,qword ptr real_n_pi_p_1_18[rip]
+
+	movlpd	xmm1,qword ptr real_m_3_pi_d_2[rip]
+	movlpd	xmm12,qword ptr real_m_3_pi_d_2_l[rip]
+
+	ja		tan_real_n_2_0_4_n
+
+	ucomisd	xmm0,qword ptr real_n_2pi_m_0_89[rip]
+	jb		tan_real_n_1_0_2
+	ucomisd	xmm0,qword ptr real_n_2pi_m_1_18[rip]
+	jb		tan_real_n_2_0_4_p
+	jmp		tan_real_24
+
+tan_real_4:
+	ucomisd	xmm0,qword ptr real_pi_p_1_18[rip]
+
+	movlpd	xmm1,qword ptr real_3_pi_d_2[rip]
+	movlpd	xmm12,qword ptr real_3_pi_d_2_l[rip]
+
+	jb		tan_real_2_0_2_p
+
+	ucomisd	xmm0,qword ptr real_2pi_m_0_89[rip]
+	ja		tan_real_1_0_2
+	ucomisd	xmm0,qword ptr real_2pi_m_1_18[rip]
+	ja		tan_real_2_0_4_n
+	att_jmp		tan_real_24
+
+tan_real_n_2:
+	ucomisd	xmm0,qword ptr real_m_1_18[rip]
+	ja		tan_real_n_s_2
+
+	movlpd	xmm1,qword ptr real_m_pi_d_2[rip]
+	movlpd	xmm12,qword ptr real_m_pi_d_2_l[rip]
+
+	ucomisd	xmm0,qword ptr real_n_pi_m_0_89[rip]
+	att_jb		tan_real_n_1_0_2
+	ucomisd	xmm0,qword ptr real_n_pi_m_1_18[rip]
+	jb		tan_real_n_2_0_2_n
+
+	att_jmp		tan_real_24
+
+tan_real_2:
+	ucomisd	xmm0,qword ptr real_1_18[rip]
+	jb		tan_real_s_2
+
+	movlpd	xmm1,qword ptr real_pi_d_2[rip]
+	movlpd	xmm12,qword ptr real_pi_d_2_l[rip]
+
+	ucomisd	xmm0,qword ptr real_pi_m_0_89[rip]
+	att_ja		tan_real_1_0_2
+	ucomisd	xmm0,qword ptr real_pi_m_1_18[rip]
+	ja		tan_real_2_0_2_n
+
+tan_real_24:
+	subsd	xmm1,xmm0				# y_1
+
+tan_real_2_:
+	movlpd	xmm2,qword ptr real_18_bits[rip]
+	movsd	xmm3,xmm12
+
+	andpd	xmm2,xmm1				# y_1_h
+	addsd	xmm3,xmm1				# y
+
+	movsd	xmm6,xmm1				# y_1
+	subsd	xmm1,xmm2				# y_1-y_1_h
+
+	movsd	xmm0,xmm2				# y_1_h
+	addsd	xmm2,xmm3				# y+y_1_h
+	addsd	xmm1,xmm12				# y_1_l
+
+	mulsd	xmm0,xmm0				# y_1_h^2
+	mulsd	xmm1,xmm2				# (y+y_1_h)*y_1_l
+
+	movsd	xmm13,xmm0				# y_1_h^2
+	addsd	xmm0,xmm1				# y^2
+
+	movsd	xmm14,xmm1				# (y_1^2)_l
+	movsd	xmm1,xmm3				# y
+	subsd	xmm3,xmm6				# y-y_1
+
+	movlpd	xmm5,qword ptr tan2_q_1[rip]
+	movlpd	xmm9,qword ptr tan2_p_0[rip]
+
+	subsd	xmm12,xmm3				# y_s
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm5,xmm2				# q2
+	mulsd	xmm9,xmm2				# p2
+
+	movlpd	xmm6,qword ptr tan2_q_2[rip]
+	movlpd	xmm11,qword ptr tan2_p_1[rip]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm1				# x3
+
+	movsd	xmm8,xmm1				# x
+
+	movlpd	xmm10,qword ptr real_0_5[rip]
+
+	mulsd	xmm6,xmm0				# q4
+	mulsd	xmm11,xmm0				# p4
+
+	addsd	xmm5,xmm6				# q4+q2
+
+	movlpd	xmm7,qword ptr real_1_0[rip]
+
+	mulsd	xmm10,xmm3				# 0.5*x2
+	mulsd	xmm3,xmm0				# x6	
+
+	addsd	xmm5,qword ptr tan2_q_0[rip]	# q4+q2+q0
+
+	mulsd	xmm0,xmm0				# x8
+	mulsd	xmm3,qword ptr tan2_p_2[rip]	# p6
+
+	mulsd	xmm5,xmm2				# (q4+q2+q0)*x3
+
+	mulsd	xmm0,qword ptr tan2_p_3[rip]	# p8
+
+	movsd	xmm2,xmm5				# (q4+q2+q0)*x3
+	addsd	xmm5,xmm8				# (q4+q2+q0)*x3+x
+
+	addsd	xmm0,xmm3				# p8+p6
+
+	divsd	xmm7,xmm5				# 1/(q4+q2+q0)*x3+x
+
+	movlpd	xmm4,qword ptr real_25_bits[rip]
+
+	addsd	xmm0,xmm11				# p8+p6+p4
+	
+	andpd	xmm5,xmm4				# q_h
+	
+	addsd	xmm0,xmm9				# p8+p6+p4+p2
+
+	movlpd	xmm9,qword ptr real_18_bits[rip]
+
+#	ucomisd	xmm8,qword ptr real_0_4
+	ucomisd	xmm10,qword ptr real_2_p_m_31[rip]
+
+	subsd	xmm8,xmm5				# x-q_h	
+
+	addsd	xmm8,xmm2				# q_l=q-q_h
+	movlpd	xmm6,qword ptr real_1_0[rip]
+
+	jb		tan_real_2_s			# x<1.25*2**-16
+
+tan_real_2_1:
+	movsd	xmm3,xmm0				# p
+
+	mulsd	xmm14,qword ptr real_0_5[rip] # (0.5*x2)_l
+	mulsd	xmm13,qword ptr real_0_5[rip] # (0.5*x2)_h
+
+	subsd	xmm3,xmm10				# p-0.5*x2
+	addsd	xmm8,xmm12				# q_l+y_s
+
+	subsd	xmm0,xmm14				# p-(0.5*x2)_l
+
+	andpd	xmm9,xmm7				# (1/q)_h
+
+	mulsd	xmm5,xmm9				# q_h*(1/q)_h
+	mulsd	xmm8,xmm9				# q_l*(1/q)_h
+
+	mulsd	xmm13,xmm9				# (0.5*x2)_h*(1/q)_h
+	mulsd	xmm0,xmm9				# (p-(0.5*x2)_l)*(1/q)_h
+	movlpd	xmm1,qword ptr real_21_bits[rip]
+
+	subsd	xmm6,xmm5				# 1-q_h*(1/q)_h
+
+	andpd	xmm1,xmm0				# ((p-(0.5*x2)_l)*(1/q)_h)_h
+	
+	subsd	xmm6,xmm8				# 1-(q_h+q_l)*(1/q)_h
+
+	subsd	xmm0,xmm1				# ((p-(0.5*x2)_l)*(1/q)_h)_l
+	subsd	xmm1,xmm13				# ((p-(0.5*x2)_l)*(1/q)_h)_h-(0.5*x2)_h*(1/q)_h
+	movlpd	xmm2,qword ptr real_21_bits[rip]
+
+	mulsd	xmm7,xmm6				# (1/q)_l
+
+	andpd	xmm2,xmm1				# (((p-(0.5*x2)_l)*(1/q)_h)_h-(0.5*x2)_h*(1/q)_h)_h
+
+	mulsd	xmm3,xmm7				# (p-0.5x2)*(1/q)_l
+
+	subsd	xmm1,xmm2				# (((p-(0.5*x2)_l)*(1/q)_h)_h-(0.5*x2)_h*(1/q)_l
+	addsd	xmm2,xmm9				# (((p-(0.5*x2)_l)*(1/q)_h)_h-(0.5*x2)_h*(1/q)_h)_h+(1/q)_h
+
+	addsd	xmm0,xmm3				# ((p-(0.5*x2)_l)*(1/q)_h)_l+(p-0.5x2)*(1/q)_l
+
+	addsd	xmm0,xmm7				# ((p-(0.5*x2)_l)*(1/q)_h)_l+(p-0.5x2)*(1/q)_l+(1/q)_l
+
+	addsd	xmm0,xmm1				# ((p-(0.5*x2)_l)*(1/q)_h)+(p-0.5x2)*(1/q)_l+(1/q)_l
+									# -(0.5*x2)_h*(1/q)_l
+	addsd	xmm0,xmm2
+
+	ret	
+
+tan_real_2_s:
+	subsd	xmm0,xmm10				# p-0.5*x2
+	addsd	xmm8,xmm12				# q_l+y_s
+
+	andpd	xmm9,xmm7				# (1/q)_h
+
+	mulsd	xmm0,xmm7				# (p-0.5*x2)*(1/q)
+
+	mulsd	xmm5,xmm9				# q_h*(1/q)_h
+	mulsd	xmm8,xmm9				# q_l*(1/q)_h
+	
+	subsd	xmm6,xmm5				# 1-q_h*(1/q)_h
+	
+	subsd	xmm6,xmm8				# 1-(q_h+q_l)*(1/q)_h
+
+	mulsd	xmm6,xmm7				# (1/q)_l
+
+	addsd	xmm0,xmm6				# (p/q)_l+(1/q)_l
+	
+	addsd	xmm0,xmm9				# p/q+1/q
+
+	ret
+
+# x<0.4
+#	subsd	xmm0,xmm10				# p-0.5*x2
+#	addsd	xmm8,xmm12				# q_l+y_s
+
+#	andpd	xmm9,xmm7				# (1/q)_h
+
+#	mulsd	xmm0,xmm7				# (p-0.5*x2)*(1/q)
+
+#	movlpd	xmm4,qword ptr real_21_bits
+
+#	mulsd	xmm5,xmm9				# q_h*(1/q)_h
+#	mulsd	xmm8,xmm9				# q_l*(1/q)_h
+
+#	andpd	xmm4,xmm0				# (p/q)_h
+	
+#	subsd	xmm6,xmm5				# 1-q_h*(1/q)_h
+
+#	subsd	xmm0,xmm4				# (p/q)_l
+#	addsd	xmm4,xmm9				# (p/q)_h+(1/q)_h
+	
+#	subsd	xmm6,xmm8				# 1-(q_h+q_l)*(1/q)_h
+
+#	mulsd	xmm6,xmm7				# (1/q)_l
+
+#	addsd	xmm0,xmm6				# (p/q)_l+(1/q)_l	
+	
+#	addsd	xmm0,xmm4				# p/q+1/q
+
+#	ret
+
+tan_real_n_s_2:
+	ucomisd	xmm0,qword ptr real_m_1_04[rip]
+	jb		tan_real_n_2_0
+
+	movlpd	xmm1,qword ptr atan_sqrt_2[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	lea		rdx,tan_n_s_2_t[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	addsd	xmm1,qword ptr atan_sqrt_2_l[rip] # y=x+atan sqrt 2
+	att_jmp		tan_real_0_5_a_s_0_5_a_s_2
+
+tan_real_s_2:
+	ucomisd	xmm0,qword ptr real_1_04[rip]
+	ja		tan_real_2_0
+
+	movlpd	xmm1,qword ptr m_atan_sqrt_2[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	lea		rdx,tan_s_2_t[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	subsd	xmm1,qword ptr atan_sqrt_2_l[rip] # y=x-atan sqrt 2
+	att_jmp		tan_real_0_5_a_s_0_5_a_s_2
+
+tan_real_n_2_0:
+	movlpd	xmm1,qword ptr atan_2_53[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	addsd	xmm1,qword ptr atan_2_53_l[rip] # y=x+atan 2
+	subsd	xmm0,xmm10				# x_l
+
+	movlpd	xmm11,qword ptr real_5_atan_2_m_2_53_l[rip]
+	movlpd	xmm12,qword ptr real_5_atan_2_m_2_53[rip]
+	movlpd	xmm13,qword ptr real_m_2_0[rip]
+	jmp		tan_real_2_0_
+
+tan_real_2_0:
+	movlpd	xmm1,qword ptr m_atan_2_53[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	addsd	xmm1,xmm0
+	andpd	xmm10,xmm0				# x_h
+
+	subsd	xmm1,qword ptr atan_2_53_l[rip] # y=x-atan 2
+	subsd	xmm0,xmm10				# x_l
+
+	movlpd	xmm11,qword ptr real_n_2_m_5_atan_2_53_l[rip]
+	movlpd	xmm12,qword ptr real_n_2_m_5_atan_2_53[rip]
+	movlpd	xmm13,qword ptr real_2_0[rip]
+
+tan_real_2_0_:
+	movsd	xmm2,xmm1				# y
+	mulsd	xmm1,xmm1				# y2
+
+	mulsd	xmm10,qword ptr real_5_0[rip] # 5*x_h
+
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm1				# y2
+	mulsd	xmm1,xmm1				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm1				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm1				# y4
+	mulsd	xmm1,xmm1				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm1,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm1,xmm6				# y8+q6+q4
+
+	mulsd	xmm0,qword ptr real_5_0[rip]	# 5*x_l
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm1,xmm7				# y8+q6+q4+q2
+
+	mulsd	xmm13,xmm2				# 2*y
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm1,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm1				# p+q
+	
+	mulsd	xmm4,xmm13				# 2*y*(p+q)
+
+#	mulsd	xmm2,qword ptr real_5_0	# 5*y
+
+	addsd	xmm3,xmm4				# p+y*(p+q)
+	subsd	xmm1,xmm4				# q-y*(p+q)
+
+	divsd	xmm3,xmm1				# (p-y*(p+q))/(q+y*(p+q))
+	
+	addsd	xmm0,xmm11
+
+	mulsd	xmm2,xmm3				# y * (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm10,xmm12				# 2+5*x_h-5*atan 2=2-5*y
+
+	mulsd	xmm2,qword ptr real_5_0[rip]	# 5*y*..
+
+	addsd	xmm0,xmm2
+
+	addsd	xmm0,xmm10				# 2+5*y+5*y * ..
+
+	ret
+
+tan_real_n_2_0_4_p:
+	ucomisd	xmm0,qword ptr real_n_2pi_m_1_04[rip]
+	subsd	xmm1,xmm0
+	att_jb		tan_real_s_2_p
+	att_jmp		tan_real_2_0_p
+
+tan_real_n_2_0_4_n:
+	ucomisd	xmm0,qword ptr real_n_pi_p_1_04[rip]
+	subsd	xmm1,xmm0
+	att_ja		tan_real_s_2_n
+	att_jmp		tan_real_2_0_n
+
+tan_real_n_2_0_2_n:
+	ucomisd	xmm0,qword ptr real_n_pi_m_1_04[rip]
+	subsd	xmm1,xmm0
+	att_jb		tan_real_s_2_p
+	att_jmp		tan_real_2_0_p
+
+tan_real_2_0_2_p:
+	ucomisd	xmm0,qword ptr real_pi_p_1_04[rip]
+	subsd	xmm1,xmm0
+	att_jb		tan_real_s_2_p
+
+tan_real_2_0_p:
+	movlpd	xmm0,qword ptr real_h_pi_m_atan_2_52[rip]
+	movlpd	xmm2,qword ptr real_h_pi_m_atan_2_52_l[rip]
+	movlpd	xmm10,qword ptr real_48_bits[rip]
+
+	subsd	xmm0,xmm1				# (0.5pi-atan 2)_h-x_h
+	subsd	xmm2,xmm12				# (0.5pi-atan 2)_l-x_l
+	andpd	xmm10,xmm1				# x_h
+
+	movlpd	xmm11,qword ptr real_m_5_0[rip]
+	addsd	xmm2,xmm0				# y=0.5pi-atan 2-x
+	subsd	xmm1,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# -5x_h
+
+	movsd	xmm0,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	addsd	xmm1,xmm12				# x_h_l+x_l
+	addsd	xmm10,qword ptr real_2_p_2_5_pi_m_5_atan_2_53[rip] # -5x+(2+2.5pi-5atan 2)=2+5y
+
+	movlpd	xmm12,qword ptr real_2_p_2_5_pi_m_5_atan_2_53_l[rip]
+	movlpd	xmm13,qword ptr real_5_0[rip]
+	jmp		tan_real_2_0_pn
+
+tan_real_2_0_4_n:
+	ucomisd	xmm0,qword ptr real_2pi_m_1_04[rip]
+	subsd	xmm1,xmm0
+	att_ja		tan_real_s_2_n
+	att_jmp		tan_real_2_0_n
+
+tan_real_2_0_2_n:
+	ucomisd	xmm0,qword ptr real_pi_m_1_04[rip]
+	subsd	xmm1,xmm0
+	att_ja		tan_real_s_2_n
+
+tan_real_2_0_n:
+	movlpd	xmm0,qword ptr real_h_pi_m_atan_2_52[rip]
+	movlpd	xmm2,qword ptr real_h_pi_m_atan_2_52_l[rip]
+	movlpd	xmm10,qword ptr real_48_bits[rip]
+
+	addsd	xmm0,xmm1				# (0.5pi-atan 2)_h+x_h
+	addsd	xmm2,xmm12				# (0.5pi-atan 2)_l+x_l
+	andpd	xmm10,xmm1				# x_h
+
+	movlpd	xmm11,qword ptr real_m_5_0[rip]
+	addsd	xmm2,xmm0				# y=0.5pi-atan 2+x
+	subsd	xmm1,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# -5x_h
+	
+	movsd	xmm0,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	addsd	xmm1,xmm12				# x_h_l+x_l
+	subsd	xmm10,qword ptr real_2_p_2_5_pi_m_5_atan_2_53[rip] # -5x-(2+2.5pi-5atan 2)=-2-5y
+
+	movlpd	xmm12,qword ptr real_5_atan_2_m_2_m_2_5_pi_53_l[rip]
+	movlpd	xmm13,qword ptr real_m_5_0[rip]
+
+tan_real_2_0_pn:
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm2				# y2
+	mulsd	xmm2,xmm2				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm2				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm2				# y4
+	mulsd	xmm2,xmm2				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm2,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm2,xmm6				# y8+q6+q4
+
+	movlpd	xmm6,qword ptr real_2_0[rip]
+	mulsd	xmm1,xmm11				# -5x_l
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm2,xmm7				# y8+q6+q4+q2
+
+	mulsd	xmm6,xmm0				# 2*y
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm2,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm2				# p+q
+	
+	mulsd	xmm4,xmm6				# 2*y*(p+q)
+
+	addsd	xmm3,xmm4				# p+y*(p+q)
+	subsd	xmm2,xmm4				# q-y*(p+q)
+
+	divsd	xmm3,xmm2				# (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm1,xmm12
+		
+	mulsd	xmm0,xmm3				# y * (p-y*(p+q))/(q+y*(p+q))
+
+	mulsd	xmm0,xmm13				# -5*y*..
+
+	addsd	xmm0,xmm1
+
+	addsd	xmm0,xmm10				# -2-5*y-5*y * ..
+
+	ret
+
+tan_real_s_2_p:
+	movlpd	xmm0,qword ptr real_h_pi_m_atan_s_2_53[rip]
+	movlpd	xmm2,qword ptr real_h_pi_m_atan_s_2_53_l[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	subsd	xmm0,xmm1				# (0.5pi-atan sqrt 2)_h-x_h
+	subsd	xmm2,xmm12				# (0.5pi-atan sqrt 2)_l-x_l
+	andpd	xmm10,xmm1				# x_h
+
+	movlpd	xmm11,qword ptr real_m_3_0[rip]
+	addsd	xmm2,xmm0				# y=0.5pi-atan 2+x
+	subsd	xmm1,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# -3x_h
+	
+	movsd	xmm0,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	addsd	xmm1,xmm12				# x_h_l+x_l
+	addsd	xmm10,qword ptr real_sqrt_2_p_1_5_pi_m_3_atan_sqrt_2_53[rip] # -3x+(2+2.5pi-5atan 2)=sqrt 2+3y
+
+	movlpd	xmm12,qword ptr real_sqrt_2_p_1_5_pi_m_3_atan_sqrt_2_53_l[rip]
+	movlpd	xmm13,qword ptr real_3_0[rip]
+	jmp		tan_real_s_2_pn
+
+tan_real_s_2_n:
+	movlpd	xmm0,qword ptr real_h_pi_m_atan_s_2_53[rip]
+	movlpd	xmm2,qword ptr real_h_pi_m_atan_s_2_53_l[rip]
+	movlpd	xmm10,qword ptr real_49_bits[rip]
+
+	addsd	xmm0,xmm1				# (0.5pi-atan sqrt 2)_h+x
+	addsd	xmm2,xmm12				# (0.5pi-atan sqrt 2)_l+x_l
+	andpd	xmm10,xmm1				# x_h
+
+	movlpd	xmm11,qword ptr real_m_3_0[rip]
+	addsd	xmm2,xmm0				# y=0.5pi-atan 2+x
+	subsd	xmm1,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# -3x_h
+	
+	movsd	xmm0,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	addsd	xmm1,xmm12				# x_h_l+x_l
+	subsd	xmm10,qword ptr real_sqrt_2_p_1_5_pi_m_3_atan_sqrt_2_53[rip] # -3x-(2+2.5pi-5atan 2)=-sqrt 2-3y
+
+	movlpd	xmm12,qword ptr real_3_atan_sqrt_2_m_sqrt_2_m_1_5_pi_m_53_l[rip]
+	movlpd	xmm13,qword ptr real_m_3_0[rip]
+
+tan_real_s_2_pn:
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm2				# y2
+	mulsd	xmm2,xmm2				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm2				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm2				# y4
+	mulsd	xmm2,xmm2				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm2,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm2,xmm6				# y8+q6+q4
+
+	movlpd	xmm6,qword ptr sqrt_2_0[rip]
+	mulsd	xmm1,xmm11				# -3x_l
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm2,xmm7				# y8+q6+q4+q2
+
+	mulsd	xmm6,xmm0				# sqrt 2*y
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm2,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm2				# p+q
+	
+	mulsd	xmm4,xmm6				# sqrt 2*y*(p+q)
+	mulsd	xmm0,xmm13				# -3*y
+
+	addsd	xmm3,xmm4				# p+sqrt 2*y*(p+q)
+	subsd	xmm2,xmm4				# q-sqrt 2*y*(p+q)
+
+	divsd	xmm3,xmm2				# (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm1,xmm12
+	
+	mulsd	xmm0,xmm3				# -3*y * (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm0,xmm1
+
+	addsd	xmm0,xmm10				# -sqrt 2-3*y-3*y * ..
+
+	ret
+
+tan_real_1_0_2:
+	subsd	xmm1,xmm0
+
+tan_real_1_0_n:
+	movlpd	xmm0,qword ptr real_m_pi_d_4[rip]
+	movlpd	xmm2,qword ptr real_m_pi_d_4_l[rip]
+
+	subsd	xmm0,xmm1				# -(pi/4)_h-x_h
+	subsd	xmm2,xmm12				# -(pi/4)_l-x_l
+	mulsd	xmm1,qword ptr real_m_2_0[rip] # -2x
+
+	addsd	xmm2,xmm0				# y=-pi/4-x
+	addsd	xmm12,xmm12				# 2x_l
+	subsd	xmm1,qword ptr real_pi_d_2_p_1_0[rip] # -2x-(pi/2+1)=-1-2y
+
+	movlpd	xmm10,qword ptr real_pi_d_2_p_1_0_l[rip]
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	jmp		tan_real_1_0_pn
+
+tan_real_n_1_0_2:
+	subsd	xmm1,xmm0
+
+tan_real_1_0_2_p:
+	movlpd	xmm0,qword ptr real_m_pi_d_4[rip]
+	movlpd	xmm2,qword ptr real_m_pi_d_4_l[rip]
+	movlpd	xmm11,qword ptr real_m_2_0[rip]
+
+	addsd	xmm0,xmm1				# -(pi/4)_h+x_h
+	addsd	xmm2,xmm12				# -(pi/4)_l+x_l
+	mulsd	xmm1,xmm11				# -2x
+
+	addsd	xmm2,xmm0				# y=pi/4-x
+	addsd	xmm12,xmm12				# 2x_l
+	addsd	xmm1,qword ptr real_pi_d_2_p_1_0[rip] # (pi/2+1)-2x-=1+2y
+
+	movlpd	xmm10,qword ptr real_m_pi_d_2_p_1_0_l[rip]
+	att_jmp		tan_real_1_0_pn
+
+tan_real_n_5:
+	ucomisd	xmm0,qword ptr real_n_2pi_m_0_338[rip]
+	movlpd	xmm12,qword ptr real_m_2_pi_l[rip]
+	movlpd	xmm1,qword ptr real_2_pi[rip]
+	ja		tan_real_n_s_0_5_5_p
+	ucomisd	xmm0,qword ptr real_n_2pi_p_0_338[rip]
+	ja		tan_real_5_n
+	ucomisd	xmm0,qword ptr real_n_2pi_p_0_699[rip]
+	ja		tan_real_n_s_0_5_5_n
+	addsd	xmm0,xmm1
+	att_jmp		tan_real_1_0_3_n
+
+tan_real_5:
+	ucomisd	xmm0,qword ptr real_2pi_m_0_338[rip]
+	movlpd	xmm12,qword ptr real_2_pi_l[rip]
+	movlpd	xmm1,qword ptr real_2_pi[rip]
+	jb		tan_real_s_0_5_5
+	ucomisd	xmm0,qword ptr real_2pi_p_0_338[rip]
+	jb		tan_real_5_0
+	ucomisd	xmm0,qword ptr real_2pi_p_0_699[rip]
+	jb		tan_real_s_0_5_5_p
+	subsd	xmm0,xmm1
+	att_jmp		tan_real_1_0_3
+
+tan_real_n_3:
+	ucomisd	xmm0,qword ptr real_n_pi_m_0_338[rip]
+	movlpd	xmm12,qword ptr real_m_pi_l[rip]
+	movlpd	xmm1,qword ptr real_pi[rip]
+	ja		tan_real_n_s_0_5_3_p
+	ucomisd	xmm0,qword ptr real_n_pi_p_0_338[rip]
+	ja		tan_real_3_n
+	ucomisd	xmm0,qword ptr real_n_pi_p_0_699[rip]
+	ja		tan_real_n_s_0_5_3_n
+	addsd	xmm0,xmm1
+	att_jmp		tan_real_1_0_3_n
+
+tan_real_3:
+	ucomisd	xmm0,qword ptr real_pi_m_0_338[rip]
+	movlpd	xmm1,qword ptr real_pi[rip]
+	movlpd	xmm12,qword ptr real_pi_l[rip]
+	jb		tan_real_s_0_5_3
+	ucomisd	xmm0,qword ptr real_pi_p_0_338[rip]
+	jb		tan_real_3_0
+	ucomisd	xmm0,qword ptr real_pi_p_0_699[rip]
+	jb		tan_real_s_0_5_3_p
+	subsd	xmm0,xmm1
+	att_jmp		tan_real_1_0_3
+
+tan_real_5_n:
+tan_real_3_n:
+	addsd	xmm0,xmm1
+	att_jmp		tan_real_3_
+
+tan_real_5_0:
+tan_real_3_0:
+	subsd	xmm0,xmm1
+
+tan_real_3_:
+	movlpd	xmm2,qword ptr real_26_bits[rip]
+
+	andpd	xmm2,xmm0				# y_1_h
+	movsd	xmm1,xmm0				# y_1
+
+	subsd	xmm1,xmm12				# y
+	movsd	xmm8,xmm0				# y_1
+	subsd	xmm0,xmm2				# y_1-y_1_h
+
+	movsd	xmm3,xmm2				# y_1_h
+	addsd	xmm2,xmm1				# y+y_1_h
+	subsd	xmm0,xmm12				# y_1_l
+
+	mulsd	xmm3,xmm3				# y_1_h^2
+	mulsd	xmm0,xmm2				# (y+y_1_h)*y_1_l
+
+	addsd	xmm0,xmm3				# y^2
+
+	subsd	xmm8,xmm1				# y_1-y
+
+	movlpd	xmm6,qword ptr tan_q_1[rip]
+	movlpd	xmm5,qword ptr tan_q_2[rip]
+
+	subsd	xmm8,xmm12				# y_s
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm7,qword ptr tan_p_1[rip]
+	mulsd	xmm6,xmm2				# q2
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	mulsd	xmm5,xmm4				# q4
+	mulsd	xmm2,qword ptr tan_q_3[rip]	# q6
+	
+	mulsd	xmm4,qword ptr tan_p_2[rip]	# p4
+
+	addsd	xmm0,xmm2				# x8+q6
+
+	mulsd	xmm7,xmm3				# p2
+
+	addsd	xmm0,xmm5				# x8+q6+q4
+
+	addsd	xmm4,xmm7				# p4+p2
+
+	addsd	xmm0,xmm6				# x8+q6+q4+q2
+
+	addsd	xmm4,qword ptr tan_p_0[rip]	# p4+p2+p0
+
+	addsd	xmm0,qword ptr tan_q_0[rip]	# x8+q6+q4+q2+q0
+
+	divsd	xmm4,xmm0				# p/q
+
+	mulsd	xmm3,xmm1				# x3	
+	movsd	xmm0,xmm1				# x
+
+#	movlpd	xmm2,qword ptr real_0_125
+
+	mulsd	xmm3,xmm4
+	
+#	ucomisd	xmm3,xmm2
+#	jae		tan_real_3_1
+
+	addsd	xmm3,xmm8
+
+	addsd	xmm0,xmm3
+
+	ret
+
+#tan_real_3_1:
+#	subsd	xmm3,xmm2
+#	addsd	xmm0,xmm2
+
+#	addsd	xmm3,xmm8
+
+#	addsd	xmm0,xmm3
+
+#	ret
+
+tan_real_1_0_3:
+	movlpd	xmm1,qword ptr real_pi_d_4[rip]
+	movlpd	xmm2,qword ptr real_pi_d_4_l[rip]
+
+	subsd	xmm1,xmm0				# (pi/4)_h-x
+	addsd	xmm2,xmm12				# (pi/4)_l+n_x_l
+	mulsd	xmm0,qword ptr real_2_0[rip] # 2x
+
+	addsd	xmm2,xmm1				# y=pi/4-x
+	movlpd	xmm1,qword ptr real_1_0_m_pi_d_2_52[rip]
+	addsd	xmm12,xmm12				# 2n_x_l
+	addsd	xmm1,xmm0				# 2x+(1-pi/2)=1-2y
+
+	movlpd	xmm11,qword ptr real_m_2_0[rip]
+	movlpd	xmm10,qword ptr real_pi_d_2_m_1_0_52_l[rip]
+
+	att_jmp		tan_real_1_0_pn
+
+tan_real_1_0_3_n:
+	movlpd	xmm1,qword ptr real_pi_d_4[rip]
+	movlpd	xmm2,qword ptr real_pi_d_4_l[rip]
+
+	addsd	xmm1,xmm0
+	subsd	xmm2,xmm12				# (pi/4)_l-n_x_l
+	mulsd	xmm0,qword ptr real_2_0[rip] # 2x
+
+	addsd	xmm2,xmm1				# y=pi/4+x
+	movlpd	xmm1,qword ptr real_pi_d_2_m_1_0_52[rip]
+	addsd	xmm12,xmm12				# 2n_x_l
+	addsd	xmm1,xmm0				# 2x+(pi/2-1)=-1+2y
+
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	movlpd	xmm10,qword ptr real_m_pi_d_2_m_1_0_52_l[rip]
+
+tan_real_1_0_pn:
+	movsd	xmm0,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm2				# y2
+	mulsd	xmm2,xmm2				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm2				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm2				# y4
+	mulsd	xmm2,xmm2				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm2,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm2,xmm6				# y8+q6+q4
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm2,xmm7				# y8+q6+q4+q2
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm2,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm2				# p+q
+
+	mulsd	xmm4,xmm0				# y*(p+q)
+	mulsd	xmm0,xmm11				# -2y | 2y
+
+	subsd	xmm3,xmm4				# p-y*(p+q)
+	addsd	xmm2,xmm4				# q+y*(p+q)
+
+	divsd	xmm3,xmm2				# (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm12,xmm10
+	
+	mulsd	xmm0,xmm3				# -2y * (p-y*(p+q))/(q+y*(p+q)) | 2y * (p-y*(p+q))/(q+y*(p+q))
+
+	subsd	xmm0,xmm12
+
+	addsd	xmm0,xmm1				# 1-2y-2y * .. | -1+2y+2y * ..
+
+	ret
+
+tan_real_n_s_0_5_5_p:
+	ucomisd	xmm0,qword ptr real_n_2pi_m_0_542[rip]
+	addsd	xmm0,xmm1
+	att_jb		tan_real_0_5_p
+	att_jmp		tan_real_s_0_5_p
+
+tan_real_s_0_5_5_p:
+	ucomisd	xmm0,qword ptr real_2pi_p_0_542[rip]
+	subsd	xmm0,xmm1
+	att_jb		tan_real_0_5_p
+	att_jmp		tan_real_s_0_5_p
+
+tan_real_n_s_0_5_3_p:
+	ucomisd	xmm0,qword ptr real_n_pi_m_0_542[rip]
+	addsd	xmm0,xmm1
+	att_jb		tan_real_0_5_p
+	att_jmp		tan_real_s_0_5_p
+
+tan_real_s_0_5_3_p:
+	ucomisd	xmm0,qword ptr real_pi_p_0_542[rip]
+	subsd	xmm0,xmm1
+	att_jb		tan_real_0_5_p
+
+tan_real_s_0_5_p:
+	movlpd	xmm1,qword ptr atan_sqrt_0_5[rip]
+	movlpd	xmm2,qword ptr atan_sqrt_0_5_l[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	subsd	xmm1,xmm0				# atan sqrt 0.5_h-x
+	addsd	xmm2,xmm12				# atan sqrt 0.5_l+n_x_l
+	andpd	xmm10,xmm0				# x_h
+
+	movlpd	xmm11,qword ptr real_1_5[rip]
+	addsd	xmm2,xmm1				# y=atan sqrt 0.5-x
+	subsd	xmm0,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# 1.5x_h
+
+	movsd	xmm1,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	subsd	xmm0,xmm12				# x_h_l-n_x_l
+	addsd	xmm10,qword ptr sqrtn_0_5_m_1_5_atann_sqrt_0_5[rip] # 1.5x+(sqrt 0.5-1.5atan sqrt 0.5)=sqrt 0.5+1.5y
+
+	movlpd	xmm12,qword ptr sqrtn_0_5_m_1_5_atann_sqrt_0_5_l[rip]
+	movlpd	xmm13,qword ptr real_m_1_5[rip]
+	movlpd	xmm14,qword ptr sqrt_0_5[rip]
+	jmp		tan_real_0_5_a_s_0_5_3_pn
+
+tan_real_n_s_0_5_5_n:
+	ucomisd	xmm0,qword ptr real_n_2pi_p_0_542[rip]
+	addsd	xmm0,xmm1
+	att_ja		tan_real_0_5_n
+	att_jmp		tan_real_s_0_5_n
+
+tan_real_s_0_5_5:
+	ucomisd	xmm0,qword ptr real_2pi_m_0_542[rip]
+	subsd	xmm0,xmm1
+	att_ja		tan_real_0_5_n
+	att_jmp		tan_real_s_0_5_n
+
+tan_real_n_s_0_5_3_n:
+	ucomisd	xmm0,qword ptr real_n_pi_p_0_542[rip]
+	addsd	xmm0,xmm1
+	att_ja		tan_real_0_5_n
+	att_jmp		tan_real_s_0_5_n
+
+tan_real_s_0_5_3:
+	ucomisd	xmm0,qword ptr real_pi_m_0_542[rip]
+	subsd	xmm0,xmm1
+	att_ja		tan_real_0_5_n
+
+tan_real_s_0_5_n:
+	movlpd	xmm1,qword ptr atan_sqrt_0_5[rip]
+	movlpd	xmm2,qword ptr atan_sqrt_0_5_l[rip]
+	movlpd	xmm10,qword ptr real_51_bits[rip]
+
+	addsd	xmm1,xmm0				# x_h+atan sqrt 0.5_h
+	subsd	xmm2,xmm12				# -n_x_l+atan sqrt 0.5_l
+	andpd	xmm10,xmm0				# x_h
+
+	movlpd	xmm11,qword ptr real_1_5[rip]
+	addsd	xmm2,xmm1				# y=x+atan sqrt 0.5
+	subsd	xmm0,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# 1.5x_h
+
+	movsd	xmm1,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	subsd	xmm0,xmm12				# x_h_l-n_x_l
+	subsd	xmm10,qword ptr sqrtn_0_5_m_1_5_atann_sqrt_0_5[rip] # 1.5x-(sqrt 0.5-1.5atan sqrt 0.5)=-sqrt 0.5+1.5y
+
+	movlpd	xmm12,qword ptr n_sqrtn_0_5_m_1_5_atann_sqrt_0_5_l[rip]
+	movlpd	xmm13,qword ptr real_1_5[rip]
+	movlpd	xmm14,qword ptr sqrt_0_5[rip]
+
+tan_real_0_5_a_s_0_5_3_pn:
+	movlpd	xmm7,qword ptr tan3_q_1[rip]
+	movlpd	xmm6,qword ptr tan3_q_2[rip]
+
+	movsd	xmm3,xmm2				# y2
+	mulsd	xmm2,xmm2				# y4
+
+	movlpd	xmm9,qword ptr tan3_q_3[rip]
+	mulsd	xmm7,xmm3				# q2
+
+	movsd	xmm4,xmm3				# y2
+	mulsd	xmm3,xmm2				# y6
+
+	movlpd	xmm8,qword ptr tan3_p_0[rip]
+
+	movsd	xmm5,xmm2				# y4
+	mulsd	xmm2,xmm2				# y8
+
+	mulsd	xmm9,xmm3				# q6
+	mulsd	xmm3,qword ptr tan3_p_2[rip]	# p6
+
+	mulsd	xmm6,xmm5				# q4
+	mulsd	xmm5,qword ptr tan3_p_1[rip]	# p4
+	
+	addsd	xmm2,xmm9				# y8+q6
+
+	mulsd	xmm8,xmm4				# p2
+
+	addsd	xmm2,xmm6				# y8+q6+q4
+
+	mulsd	xmm0,xmm11				# 1.25|1.5x_l
+
+	addsd	xmm3,xmm5				# p6+p4
+	addsd	xmm2,xmm7				# y8+q6+q4+q2
+
+	mulsd	xmm14,xmm1				# 0.5|sqrt 0.5*y
+
+	addsd	xmm3,xmm8				# p6+p4+p2
+	addsd	xmm2,qword ptr tan3_q_0[rip]	# y8+q6+q4+q2+q0
+
+	movsd	xmm4,xmm3				# p
+	addsd	xmm4,xmm2				# p+q
+
+	mulsd	xmm4,xmm14				# 0.5|sqrt 0.5*y*(p+q)
+
+	subsd	xmm3,xmm4				# p-sqrt 0.5*y*(p+q)
+	addsd	xmm2,xmm4				# q+sqrt 0.5*y*(p+q)
+
+	divsd	xmm3,xmm2				# (p-y*(p+q))/(q+y*(p+q))
+
+	addsd	xmm0,xmm12
+	
+	mulsd	xmm1,xmm3				# y * (p-y*(p+q))/(q+y*(p+q))
+
+	mulsd	xmm1,xmm13				# 1.25|1.5*y*..
+
+	addsd	xmm0,xmm1
+
+	addsd	xmm0,xmm10				# 0.5|sqrt 0.5+1.25|1.5*y+1.25|1.5*y * ..
+
+	ret
+
+tan_real_0_5_p:
+	movlpd	xmm1,qword ptr atan_0_5_53[rip]
+	movlpd	xmm2,qword ptr atan_0_5_53_l[rip]
+	movlpd	xmm10,qword ptr real_50_bits[rip]
+
+	subsd	xmm1,xmm0				# atan 0.5_h-x
+	addsd	xmm2,xmm12				# atan 0.5_l-n_x_l
+	andpd	xmm10,xmm0				# x_h
+
+	movlpd	xmm11,qword ptr real_1_25[rip]
+	addsd	xmm2,xmm1				# y=x+atan 0.5
+	subsd	xmm0,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# 1.25x_h
+
+	movsd	xmm1,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	subsd	xmm0,xmm12				# x_h_l-n_x_l
+	addsd	xmm10,qword ptr real_0_5_m_1_25_atan_0_5_49[rip] # 1.25x+(0.5-1.25atan 0.5)=0.5+1.25y
+
+	movlpd	xmm12,qword ptr real_0_5_m_1_25_atan_0_5_49_l[rip]
+	movlpd	xmm13,qword ptr real_m_1_25[rip]
+	movlpd	xmm14,qword ptr real_0_5[rip]
+	att_jmp		tan_real_0_5_a_s_0_5_3_pn
+
+tan_real_0_5_n:
+	movlpd	xmm1,qword ptr atan_0_5_53[rip]
+	movlpd	xmm2,qword ptr atan_0_5_53_l[rip]
+	movlpd	xmm10,qword ptr real_50_bits[rip]
+
+	addsd	xmm1,xmm0				# x_h+atan 0.5_h
+	subsd	xmm2,xmm12				# -n_x_l+atan 0.5_l
+	andpd	xmm10,xmm0				# x_h
+
+	movlpd	xmm11,qword ptr real_1_25[rip]
+	addsd	xmm2,xmm1				# y=x+atan 0.5
+	subsd	xmm0,xmm10				# x_h_l
+	mulsd	xmm10,xmm11				# 1.25x_h
+
+	movsd	xmm1,xmm2				# y
+	mulsd	xmm2,xmm2				# y2
+
+	subsd	xmm0,xmm12				# x_h_l-n_x_l
+	subsd	xmm10,qword ptr real_0_5_m_1_25_atan_0_5_49[rip] # 1.25x-(0.5-1.25atan 0.5)=-0.5+1.25y
+
+	movlpd	xmm12,qword ptr real_1_25_atan_0_5_m_0_5_49_l[rip]
+	movlpd	xmm13,qword ptr real_1_25[rip]
+	movlpd	xmm14,qword ptr real_0_5[rip]
+	att_jmp		tan_real_0_5_a_s_0_5_3_pn
+
+
+sin_cos_or_tan_real_too_large:
+sin_cos_or_tan_real_too_small_or_nan:
+	subsd	xmm0,xmm0
+	ret
+
+rem_36825084_pi:
+	movsd	xmm1,xmm0	
+	divsd	xmm0,qword ptr real_36825084_pi[rip]
+	
+	movq	xmm4,qword ptr round_c[rip]
+	movq	xmm2,xmm0
+	psrlq	xmm0,52
+	psubq	xmm4,xmm0
+	movq	xmm5,qword ptr mask_all_one[rip]
+	psllq	xmm5,xmm4
+	andpd	xmm5,xmm2
+	
+	movlpd	xmm2,qword ptr real_36825084_pi_27[rip]
+	movlpd	xmm3,qword ptr real_36825084_pi_27_27[rip]
+	movlpd	xmm4,qword ptr real_36825084_pi_54_l[rip]
+
+	movsd	xmm0,xmm1
+
+	mulsd	xmm2,xmm5
+	mulsd	xmm3,xmm5
+	mulsd	xmm4,xmm5
+
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+	movsd	xmm7,xmm0	
+	subsd	xmm0,xmm4
+
+	subsd	xmm7,xmm0
+	subsd	xmm7,xmm4
+
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+
+	ucomisd	xmm1,qword ptr real_3_0[rip]
+	jae		rem_36825084_pi_g3
+
+	ucomisd	xmm1,qword ptr real_1_0[rip]
+	jae		rem_36825084_pi_g1
+
+	movlpd	xmm1,qword ptr real_0_0[rip]
+	movsd	xmm5,xmm1
+	movsd	xmm6,xmm1
+	ret
+
+rem_36825084_pi_g1:
+	movlpd	xmm1,qword ptr real_2_0[rip]
+	movlpd	xmm5,qword ptr real_0_0[rip]
+	movsd	xmm6,xmm1
+	ret
+
+rem_36825084_pi_g3:
+	addsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi+1
+	
+	movq	xmm4,qword ptr round_even_c[rip]
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi+1)
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm5,xmm4
+	psllq	xmm6,xmm4
+	ret
+
+rem_n_36825084_pi:
+	movsd	xmm1,xmm0	
+	divsd	xmm0,qword ptr real_36825084_pi[rip]
+	
+	movq	xmm4,qword ptr m_round_c[rip]
+	movq	xmm2,xmm0
+	psrlq	xmm0,52
+	psubq	xmm4,xmm0
+	movq	xmm5,qword ptr mask_all_one[rip]
+	psllq	xmm5,xmm4
+	andpd	xmm5,xmm2
+	
+	movlpd	xmm2,qword ptr real_36825084_pi_27[rip]
+	movlpd	xmm3,qword ptr real_36825084_pi_27_27[rip]
+	movlpd	xmm4,qword ptr real_36825084_pi_54_l[rip]
+
+	movsd	xmm0,xmm1
+
+	mulsd	xmm2,xmm5
+	mulsd	xmm3,xmm5
+	mulsd	xmm4,xmm5
+
+	subsd	xmm0,xmm2
+	subsd	xmm0,xmm3
+	movsd	xmm7,xmm0	
+	subsd	xmm0,xmm4
+
+	subsd	xmm7,xmm0
+	subsd	xmm7,xmm4
+
+	movlpd	xmm1,qword ptr real_4_d_pi[rip]
+
+	mulsd	xmm1,xmm0				# x*4/pi
+
+	ucomisd	xmm1,qword ptr real_m_3_0[rip]
+	jbe		rem_n_36825084_pi_g3
+
+	ucomisd	xmm1,qword ptr real_m_1_0[rip]
+	jbe		rem_n_36825084_pi_g1
+
+	movlpd	xmm1,qword ptr real_0_0[rip]
+	movsd	xmm5,xmm1
+	movsd	xmm6,xmm1
+	ret
+
+rem_n_36825084_pi_g1:
+	movlpd	xmm1,qword ptr real_m_2_0[rip]
+	movlpd	xmm5,qword ptr real_0_0[rip]
+	movsd	xmm6,xmm1
+	ret
+	
+rem_n_36825084_pi_g3:
+	subsd	xmm1,qword ptr real_1_0[rip]	# x*4/pi-1
+	
+	movq	xmm4,qword ptr m_round_even_c[rip]
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2				# round_even (x*4/pi-1)
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	movq	xmm6,qword ptr mask_all_one_except_second_last[rip]
+	psllq	xmm5,xmm4
+	psllq	xmm6,xmm4
+	ret
+
+
+	.globl	asin_real
+
+asin_real:
+	ucomisd	xmm0,qword ptr real_0_4[rip]
+
+	lea		rcx,asin_c[rip]
+
+	ja		asin_real_2
+
+	ucomisd	xmm0,qword ptr real_m_0_4[rip]
+	jb		asin_real_3				# x<0.55 | NAN
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm6,qword ptr (asin_q_4-asin_c)[rcx]
+	movlpd	xmm7,qword ptr (asin_q_3-asin_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm8,qword ptr (asin_q_2-asin_c)[rcx]
+	movlpd	xmm9,qword ptr (asin_q_1-asin_c)[rcx]
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movlpd	xmm10,qword ptr (asin_p_1-asin_c)[rcx]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm3,xmm4				# x6
+
+	mulsd	xmm6,xmm0				# q8
+	movsd	xmm5,xmm0				# x8
+	mulsd	xmm0,xmm2				# x10
+
+	mulsd	xmm5,qword ptr (asin_p_4-asin_c)[rcx] # p8
+
+	mulsd	xmm7,xmm3				# q6
+	mulsd	xmm3,qword ptr (asin_p_3-asin_c)[rcx] # p6
+	
+	addsd	xmm0,xmm6				# x10+q8
+
+	mulsd	xmm8,xmm4				# q4
+	mulsd	xmm4,qword ptr (asin_p_2-asin_c)[rcx] # p4
+
+	addsd	xmm0,xmm7				# x10+q8+q6
+	addsd	xmm3,xmm5				# p8+p6
+
+	mulsd	xmm9,xmm2				# q2
+	mulsd	xmm10,xmm2				# p2
+
+	addsd	xmm0,xmm8				# x10+q8+q6+q4
+	addsd	xmm3,xmm4				# p8+p6+p4
+
+	addsd	xmm0,xmm9				# x10+q8+q6+q4+q2
+	addsd	xmm3,xmm10				# p8+p6+p4+p2
+
+	addsd	xmm0,qword ptr (asin_q_0-asin_c)[rcx] # x10+q8+q6+q4+q2+q0
+	addsd	xmm3,qword ptr (asin_p_0-asin_c)[rcx] # p8+p6+p4+p2+p0
+
+	divsd	xmm3,xmm0				# p/q
+
+	mulsd	xmm2,xmm1				# x3
+	movsd	xmm0,xmm1				# x
+
+	mulsd	xmm2,xmm3
+	addsd	xmm0,xmm2
+
+	ret
+
+asin_real_2:
+	ucomisd	xmm0,qword ptr real_0_675[rip]
+	jb		asin_real_0_54
+
+	movlpd	xmm1,qword ptr real_1_0[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm1,xmm0
+
+	jae		asin_real_1_or_e
+
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	
+	movsd	xmm0,xmm1					# x
+	mulsd	xmm1,xmm1					# x2
+
+	mulsd	xmm11,xmm0					# 2x
+
+	movlpd	xmm7,qword ptr (asin2_q_3-asin_c)[rcx]
+
+	movsd	xmm2,xmm1					# x2
+	mulsd	xmm1,xmm1					# x4
+
+	sqrtsd	xmm12,xmm11					# sqrt 2x
+
+	movlpd	xmm5,qword ptr (asin2_p_3-asin_c)[rcx]
+	movlpd	xmm8,qword ptr (asin2_q_2-asin_c)[rcx]
+
+	movsd	xmm4,xmm1					# x4
+
+	movlpd	xmm9,qword ptr (asin2_q_1-asin_c)[rcx]
+	movlpd	xmm10,qword ptr (asin2_p_1-asin_c)[rcx]
+
+	movsd	xmm3,xmm2					# x2
+	mulsd	xmm3,xmm0					# x3
+
+	mulsd	xmm4,qword ptr (asin2_p_4-asin_c)[rcx] # p4
+
+	mulsd	xmm7,xmm3					# q3
+	mulsd	xmm5,xmm3					# p3
+	
+	mulsd	xmm8,xmm2					# q2
+	mulsd	xmm2,qword ptr (asin2_p_2-asin_c)[rcx] # p2
+
+	addsd	xmm1,xmm7					# x4+q3
+	addsd	xmm4,xmm5					# p4+p3
+
+	mulsd	xmm9,xmm0					# q1
+	mulsd	xmm10,xmm0					# p1
+
+	addsd	xmm1,xmm8					# x5+q4+q3+q2
+	addsd	xmm4,xmm2					# p4+p3+p2
+
+	addsd	xmm1,xmm9					# x5+q4+q3+q2+q1
+	addsd	xmm4,xmm10					# p4+p3+p2+p1
+
+	movlpd	xmm13,qword ptr real_21_bits[rip]
+
+	addsd	xmm1,qword ptr (asin2_q_0-asin_c)[rcx] # x5+q4+q3+q2+q1+q0
+	addsd	xmm4,qword ptr (asin2_p_0-asin_c)[rcx] # p4+p3+p2+p1+p0
+
+	andpd	xmm13,xmm12					# s21
+
+	divsd	xmm4,xmm1					# p/q
+
+	movsd	xmm14,xmm13					# s21
+	movsd	xmm15,xmm12					# s
+	addsd	xmm12,xmm13					# s+s21
+
+	mulsd	xmm13,xmm13					# s21*s21
+	subsd	xmm11,xmm13					# 2x-s21*s21
+	divsd	xmm11,xmm12					# sl
+
+	addsd	xmm14,qword ptr real_m_pi_d_2[rip]	# s21-pi_d_2
+	mulsd	xmm4,xmm0					# x*p/q
+
+	movlpd	xmm0,qword ptr real_pi_d_2_l[rip]
+	subsd	xmm0,xmm11					# pi_d_2_l-sl
+
+	mulsd	xmm4,xmm15					# s*x*p/q
+
+	subsd	xmm0,xmm4					# pi_d_2_l-sl-s*x*p/q
+	subsd	xmm0,xmm14					# pi_d_2_l-sl-s*x*p/q-s21+pi_d_2
+
+	ret
+
+asin_real_3:
+	ucomisd	xmm0,qword ptr real_m_0_675[rip]
+	ja		asin_real_m_0_54
+
+	movlpd	xmm1,qword ptr real_m_1_0[rip]
+
+	ucomisd	xmm0,xmm1
+	jbe		asin_real_m_1_or_e			# x<=-1 | NAN
+
+	subsd	xmm0,xmm1
+
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	
+	movsd	xmm1,xmm0					# x
+	mulsd	xmm0,xmm0					# x2
+
+	mulsd	xmm11,xmm1					# 2x
+
+	movlpd	xmm7,qword ptr (asin2_q_3-asin_c)[rcx]
+
+	movsd	xmm2,xmm0					# x2
+	mulsd	xmm0,xmm0					# x4
+
+	sqrtsd	xmm12,xmm11					# sqrt 2x
+
+	movlpd	xmm5,qword ptr (asin2_p_3-asin_c)[rcx]
+	movlpd	xmm8,qword ptr (asin2_q_2-asin_c)[rcx]
+
+	movsd	xmm4,xmm0					# x4
+
+	movlpd	xmm9,qword ptr (asin2_q_1-asin_c)[rcx]
+	movlpd	xmm10,qword ptr (asin2_p_1-asin_c)[rcx]
+
+	movsd	xmm3,xmm2					# x2
+	mulsd	xmm3,xmm1					# x3
+
+	mulsd	xmm4,qword ptr (asin2_p_4-asin_c)[rcx] # p4
+
+	mulsd	xmm7,xmm3					# q3
+	mulsd	xmm5,xmm3					# p3
+	
+	mulsd	xmm8,xmm2					# q2
+	mulsd	xmm2,qword ptr (asin2_p_2-asin_c)[rcx] # p2
+
+	addsd	xmm0,xmm7					# x4+q3
+	addsd	xmm4,xmm5					# p4+p3
+
+	mulsd	xmm9,xmm1					# q1
+	mulsd	xmm10,xmm1					# p1
+
+	addsd	xmm0,xmm8					# x5+q4+q3+q2
+	addsd	xmm4,xmm2					# p4+p3+p2
+
+	addsd	xmm0,xmm9					# x5+q4+q3+q2+q1
+	addsd	xmm4,xmm10					# p4+p3+p2+p1
+
+	movlpd	xmm13,qword ptr real_21_bits[rip]
+
+	addsd	xmm0,qword ptr (asin2_q_0-asin_c)[rcx] # x5+q4+q3+q2+q1+q0
+	addsd	xmm4,qword ptr (asin2_p_0-asin_c)[rcx] # p4+p3+p2+p1+p0
+
+	andpd	xmm13,xmm12					# s21
+
+	divsd	xmm4,xmm0					# p/q
+
+	movsd	xmm14,xmm13					# s21
+	movsd	xmm15,xmm12					# s
+	addsd	xmm12,xmm13					# s+s21
+	mulsd	xmm13,xmm13					# s21*s21
+	subsd	xmm11,xmm13					# 2x-s21*s21
+	divsd	xmm11,xmm12					# sl
+
+	addsd	xmm14,qword ptr real_m_pi_d_2[rip]	# s21-pi_d_2
+	mulsd	xmm4,xmm1					# x*p/q
+
+	movlpd	xmm0,qword ptr real_m_pi_d_2_l[rip]
+	addsd	xmm0,xmm11					# -pi_d_2_l+sl
+
+	mulsd	xmm4,xmm15					# s*x*p/q
+
+	addsd	xmm0,xmm4					# -pi_d_2_l+sl+s*x*p/q
+	addsd	xmm0,xmm14					# -pi_d_2_l+sl+s*x*p/q+s21-pi_d_2
+
+	ret
+
+asin_real_m_0_54:
+	movlpd	xmm1,qword ptr real_m_0_54000000017867999524[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	lea		rcx,asin_p3_c[rip]
+	lea		rdx,asin_p4_c[rip]
+
+	cmova	rcx,rdx
+
+	jmp		asin_real_pm_0_54
+
+asin_real_0_54:
+	movlpd	xmm1,qword ptr real_0_54000000017867999524[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	lea		rcx,asin_p1_c[rip]
+	lea		rdx,asin_p2_c[rip]
+
+	cmovb	rcx,rdx
+
+asin_real_pm_0_54:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm7,qword ptr (asin_p1_4-asin_p1_c)[rcx]
+	movlpd	xmm8,qword ptr (asin_p1_3-asin_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm7,xmm2				# p4
+	mulsd	xmm8,xmm2				# p3
+
+	movlpd	xmm9,qword ptr (asin_p1_6-asin_p1_c)[rcx]
+	movlpd	xmm10,qword ptr (asin_p1_5-asin_p1_c)[rcx]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm2,xmm0				# x6
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+	
+	mulsd	xmm9,xmm4				# p6
+	mulsd	xmm10,xmm4				# p5
+
+	movlpd	xmm11,qword ptr (asin_p1_8-asin_p1_c)[rcx]
+	movlpd	xmm12,qword ptr (asin_p1_10-asin_p1_c)[rcx]
+
+	mulsd	xmm4,xmm2				# x10
+
+	movlpd	xmm13,qword ptr (asin_p1_12-asin_p1_c)[rcx]
+	movlpd	xmm14,qword ptr (asin_p1_14-asin_p1_c)[rcx]
+
+	movsd	xmm6,xmm2				# x6
+	mulsd	xmm2,xmm2				# x12
+
+	mulsd	xmm11,xmm6				# p8
+	mulsd	xmm6,qword ptr (asin_p1_7-asin_p1_c)[rcx] # p7
+
+	mulsd	xmm12,xmm0				# p10
+	mulsd	xmm0,qword ptr (asin_p1_9-asin_p1_c)[rcx] # p9
+
+	mulsd	xmm13,xmm4				# p12
+	mulsd	xmm4,qword ptr (asin_p1_11-asin_p1_c)[rcx] # p11
+
+	mulsd	xmm14,xmm2				# p14
+	mulsd	xmm2,qword ptr (asin_p1_13-asin_p1_c)[rcx] # p13
+
+	addsd	xmm13,xmm14				# p14+p12
+	addsd	xmm2,xmm4				# p13+p11
+
+	movlpd	xmm15,qword ptr (asin_p1_2-asin_p1_c)[rcx]
+	
+	addsd	xmm12,xmm13				# p14+p12+p10
+	addsd	xmm0,xmm2				# p13+p11+p9
+
+	mulsd	xmm15,xmm3				# p2
+	
+	addsd	xmm11,xmm12				# p14+p12+p10+p8
+	addsd	xmm0,xmm6				# p13+p11+p9+p7
+
+	movlpd	xmm2,qword ptr real_26_bits[rip]
+	movsd	xmm4,xmm1				# x
+	movlpd	xmm6,qword ptr (asin_p1_1l-asin_p1_c)[rcx]
+	movlpd	xmm5,qword ptr (asin_p1_1h-asin_p1_c)[rcx]
+
+	addsd	xmm9,xmm11				# p14+p12+p10+p8+p6
+	addsd	xmm0,xmm10				# p13+p11+p9+p7+p5
+
+	andpd	xmm2,xmm1				# x_h
+	mulsd	xmm6,xmm1				# x*c1l
+	movlpd	xmm13,qword ptr (asin_p1_0h-asin_p1_c)[rcx]
+
+	addsd	xmm7,xmm9				# p14+p12+p10+p8+p6+p4
+	addsd	xmm0,xmm8				# p13+p11+p9+p7+p5+p3
+
+	subsd	xmm4,xmm2				# x_l
+	mulsd	xmm2,xmm5				# x_h*c1h
+
+	mulsd	xmm7,xmm3				# (p14+p12+p10+p8+p6+p4)(*x2)
+	mulsd	xmm0,xmm1				# (p13+p11+p9+p7+p5+p3)(*x)
+
+	movsd	xmm14,xmm13
+	addsd	xmm13,xmm2				# x_h*c1h+c0h
+	
+	addsd	xmm0,xmm7				# p14+p13+..+p3
+
+	subsd	xmm14,xmm13				# c0h-(x_h*c1h+c0h)
+	mulsd	xmm4,xmm5				# x_l*c1h
+
+	addsd	xmm0,xmm15				# p14+p13+..+p3+p2
+#	addsd	xmm0,qword ptr (asin_p1_0l-asin_p1_c)[rcx]
+	
+	addsd	xmm14,xmm2				# (c0h-(x_h*c1h+c0h))+x_h*c1h
+	addsd	xmm4,xmm6				# x_l*c1h+x*c1l
+	
+	addsd	xmm4,xmm14				# (c0h-(x_h*c1h+c0h))+x_h*c1h+x_l*c1h+x*c1l
+
+	addsd	xmm0,xmm4
+	addsd	xmm0,xmm13
+	ret
+
+asin_real_1_or_e:
+	jne		asin_real_e
+
+	movlpd	xmm0,qword ptr real_pi_d_2[rip]
+	ret
+
+asin_real_m_1_or_e:
+	att_jb		asin_real_e					# NAN
+
+	movlpd	xmm0,qword ptr real_m_pi_d_2[rip]
+	ret
+
+asin_real_e:
+	subsd	xmm0,xmm0
+	divsd	xmm0,xmm0
+	ret
+
+
+	.globl	acos_real
+
+acos_real:
+#	ucomisd	xmm0,qword ptr real_0_58
+	ucomisd	xmm0,qword ptr real_0_5[rip]
+	ja		acos_real_2
+	ucomisd	xmm0,qword ptr real_m_0_58[rip]
+	jb		acos_real_3
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm6,qword ptr acos_q_4[rip]
+	movlpd	xmm7,qword ptr acos_q_3[rip]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm8,qword ptr acos_q_2[rip]
+	movlpd	xmm9,qword ptr acos_q_1[rip]
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movlpd	xmm10,qword ptr acos_p_1[rip]
+
+	movsd	xmm3,xmm2				# x2
+	mulsd	xmm3,xmm4				# x6
+
+	mulsd	xmm6,xmm0				# q8
+	movsd	xmm5,xmm0				# x8
+	mulsd	xmm0,xmm2				# x10
+
+	mulsd	xmm5,qword ptr acos_p_4[rip]	# p8
+
+	mulsd	xmm7,xmm3				# q6
+	mulsd	xmm3,qword ptr acos_p_3[rip] # p6
+	
+	addsd	xmm0,xmm6				# x10+q8
+
+	mulsd	xmm8,xmm4				# q4
+	mulsd	xmm4,qword ptr acos_p_2[rip] # p4
+
+	addsd	xmm0,xmm7				# x10+q8+q6
+	addsd	xmm3,xmm5				# p8+p6
+
+	mulsd	xmm9,xmm2				# q2
+	mulsd	xmm10,xmm2				# p2
+
+	addsd	xmm0,xmm8				# x10+q8+q6+q4
+	addsd	xmm3,xmm4				# p8+p6+p4
+
+	movlpd	xmm11,qword ptr real_21_bits[rip]
+
+	addsd	xmm0,xmm9				# x10+q8+q6+q4+q2
+	addsd	xmm3,xmm10				# p8+p6+p4+p2
+
+	andpd	xmm11,xmm1				# x21
+
+	addsd	xmm0,qword ptr acos_q_0[rip]	# x10+q8+q6+q4+q2+q0
+	addsd	xmm3,qword ptr acos_p_0[rip] # p8+p6+p4+p2+p0
+
+	mulsd	xmm2,xmm1				# x3
+
+	divsd	xmm3,xmm0				# p/q
+
+#	ucomisd	xmm1,qword ptr real_0_54
+#	ja		acos_real_1
+
+	movsd	xmm0,qword ptr real_pi_d_2[rip] # pi_d_2
+
+	subsd	xmm1,xmm11				# x21_l
+	subsd	xmm0,xmm11				# pi_d_2-x21
+
+	mulsd	xmm2,xmm3				# x3*(p/q)
+
+	addsd	xmm2,qword ptr real_m_pi_d_2_l[rip] # x3*(p/q)-pi_d_2_l
+	
+	addsd	xmm1,xmm2				# x21_l+x3*(p/q)-pi_d_2_l
+
+	subsd	xmm0,xmm1				# pi_d_2-x21-x21_l-x3*(p/q)+pi_d_2_l
+	
+	ret
+
+#acos_real_1:
+#	movlpd	xmm0,qword ptr real_pi_d_2_m_0_03125
+
+#	subsd	xmm1,xmm11				# x21_l
+#	subsd	xmm0,xmm11				# pi_d_2-0.03125-x21
+
+#	mulsd	xmm2,xmm3				# x3*(p/q)
+
+#	addsd	xmm1,qword ptr real_m_pi_d_2_l # x21_l-pi_d_2_l
+	
+#	subsd	xmm2,qword ptr real_0_03125 # x3*(p/q)-0.03125
+	
+#	addsd	xmm1,xmm2				# x21_l+x3*(p/q)-0.03125-pi_d_2_l
+
+#	subsd	xmm0,xmm1				# pi_d_2-x21-x21_l-x3*(p/q)+pi_d_2_l
+	
+#	ret
+
+acos_real_2:
+	ucomisd	xmm0,qword ptr real_0_75[rip]
+	jb		acos_real_0_65
+
+	movlpd	xmm1,qword ptr real_1_0[rip]
+
+	ucomisd	xmm0,xmm1
+	
+	subsd	xmm1,xmm0
+
+	jae		acos_real_1_or_e
+
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	
+	movsd	xmm0,xmm1					# x
+	mulsd	xmm1,xmm1					# x2
+
+	mulsd	xmm11,xmm0					# 2x
+
+	movlpd	xmm7,qword ptr acos2_q_3[rip]
+
+	movsd	xmm2,xmm1					# x2
+	mulsd	xmm1,xmm1					# x4
+
+	sqrtsd	xmm12,xmm11					# sqrt 2x
+
+	movlpd	xmm5,qword ptr acos2_p_3[rip]
+	movlpd	xmm8,qword ptr acos2_q_2[rip]
+
+	movsd	xmm4,xmm1					# x4
+
+	movlpd	xmm9,qword ptr acos2_q_1[rip]
+	movlpd	xmm10,qword ptr acos2_p_1[rip]
+
+	movsd	xmm3,xmm2					# x2
+	mulsd	xmm3,xmm0					# x3
+
+	mulsd	xmm1,qword ptr acos2_p_4[rip]	# p4
+
+	mulsd	xmm7,xmm3					# q3
+	mulsd	xmm5,xmm3					# p3
+
+	mulsd	xmm8,xmm2					# q2
+	mulsd	xmm2,qword ptr acos2_p_2[rip]	# p2
+
+	addsd	xmm4,xmm7					# x4+q3
+	addsd	xmm1,xmm5					# p4+p3
+
+	mulsd	xmm9,xmm0					# q1
+	mulsd	xmm10,xmm0					# p1
+
+	addsd	xmm4,xmm8					# x4+q3+q2
+	addsd	xmm1,xmm2					# p4+p3+p2
+
+	addsd	xmm4,xmm9					# x4+q3+q2+q1
+	addsd	xmm1,xmm10					# p4+p3+p2+p1
+
+	movlpd	xmm13,qword ptr real_21_bits[rip]
+
+	addsd	xmm4,qword ptr acos2_q_0[rip]	# x4+q3+q2+q1+q0
+	addsd	xmm1,qword ptr acos2_p_0[rip]	# p4+p3+p2+p1+p0
+
+	andpd	xmm13,xmm12					# s21
+
+	divsd	xmm1,xmm4					# p/q
+
+	movsd	xmm14,xmm13					# s21
+	movsd	xmm15,xmm12					# s
+	addsd	xmm12,xmm13					# s+s21
+	mulsd	xmm13,xmm13					# s21*s21
+	subsd	xmm11,xmm13					# 2x-s21*s21
+	divsd	xmm11,xmm12					# sl
+	
+	mulsd	xmm0,xmm1					# x*p/q
+
+	mulsd	xmm0,xmm15					# s
+	addsd	xmm0,xmm11					# s*x*p/q+sl
+	addsd	xmm0,xmm14					# s*x*p/q+sl+s21
+
+	ret
+
+acos_real_0_65:
+	movlpd	xmm1,qword ptr real_0_65000000004061742054[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	lea		rcx,acos_p1_c[rip]
+	lea		rdx,acos_p2_c[rip]
+
+	cmovb	rcx,rdx
+
+	att_jmp		asin_real_pm_0_54
+
+acos_real_3:
+	movlpd	xmm1,real_m_1_0[rip]
+
+	ucomisd	xmm0,xmm1
+	
+	subsd	xmm0,xmm1
+
+	jbe		acos_real_m_1_or_e
+
+	movlpd	xmm11,qword ptr real_2_0[rip]
+	
+	movsd	xmm1,xmm0					# x
+	mulsd	xmm0,xmm0					# x2
+
+	mulsd	xmm11,xmm1					# 2x
+
+	movlpd	xmm7,qword ptr acos2_q_3[rip]
+
+	movsd	xmm2,xmm0					# x2
+	mulsd	xmm0,xmm0					# x4
+
+	sqrtsd	xmm12,xmm11					# sqrt 2x
+
+	movlpd	xmm5,qword ptr acos2_p_3[rip]
+	movlpd	xmm8,qword ptr acos2_q_2[rip]
+
+	movsd	xmm4,xmm0					# x4
+
+	movlpd	xmm9,qword ptr acos2_q_1[rip]
+	movlpd	xmm10,qword ptr acos2_p_1[rip]
+
+	movsd	xmm3,xmm2					# x2
+	mulsd	xmm3,xmm1					# x3
+
+	mulsd	xmm0,qword ptr acos2_p_4[rip]	# p4
+
+	mulsd	xmm7,xmm3					# q3
+	mulsd	xmm5,xmm3					# p3
+
+	mulsd	xmm8,xmm2					# q2
+	mulsd	xmm2,qword ptr acos2_p_2[rip]	# p2
+
+	addsd	xmm4,xmm7					# x4+q3
+	addsd	xmm0,xmm5					# p4+p3
+
+	mulsd	xmm9,xmm1					# q1
+	mulsd	xmm10,xmm1					# p1
+
+	addsd	xmm4,xmm8					# x4+q3+q2
+	addsd	xmm0,xmm2					# p4+p3+p2
+
+	addsd	xmm4,xmm9					# x4+q3+q2+q1
+	addsd	xmm0,xmm10					# p4+p3+p2+p1
+
+	movlpd	xmm13,qword ptr real_21_bits[rip]
+
+	addsd	xmm4,qword ptr acos2_q_0[rip]	# x4+q3+q2+q1+q0
+	addsd	xmm0,qword ptr acos2_p_0[rip]	# p4+p3+p2+p1+p0
+
+	andpd	xmm13,xmm12					# s21
+
+	divsd	xmm0,xmm4					# p/q
+
+	movsd	xmm14,xmm13					# s21
+	movsd	xmm15,xmm12					# s
+	addsd	xmm12,xmm13					# s+s21
+	mulsd	xmm13,xmm13					# s21*s21
+	subsd	xmm11,xmm13					# 2x-s21*s21
+	divsd	xmm11,xmm12					# sl
+	
+	mulsd	xmm1,xmm0					# x*p/q
+
+	mulsd	xmm1,xmm15					# s
+	movlpd	xmm0,qword ptr real_pi[rip]
+
+	subsd	xmm11,qword ptr real_pi_l[rip]	# sl-pi_l
+	addsd	xmm1,xmm11					# s*x*p/q+sl-pi_l
+
+	subsd	xmm0,xmm14					# pi-s21
+	subsd	xmm0,xmm1					# pi-s21-(s*x*p/q+sl)
+
+#	addsd	xmm1,xmm14					# s*x*p/q+sl+s21-pi_l
+#	subsd	xmm0,xmm1					# pi-(s*x*p/q+sl+s21)
+
+	ret
+
+acos_real_1_or_e:
+	jne		acos_real_e
+
+	movlpd	xmm0,qword ptr real_0_0[rip]
+	ret
+
+acos_real_m_1_or_e:
+	att_jb		acos_real_e					# NAN
+
+	movlpd	xmm0,qword ptr real_pi[rip]
+	ret
+
+acos_real_e:
+	subsd	xmm0,xmm0
+	divsd	xmm0,xmm0
+	ret
+
+	.globl	atan_real
+
+atan_real:
+	ucomisd	xmm0,qword ptr real_m_0_26[rip]
+
+	movsd	xmm1,xmm0				# x
+
+	jb		atan_real_n				# x<-0.26 | NAN
+
+	ucomisd	xmm0,qword ptr real_1_0[rip]
+	ja		atan_real_g1			# x>1.0
+
+	ucomisd	xmm0,qword ptr real_0_70[rip]
+	ja		atan_real_2
+
+	ucomisd	xmm0,qword ptr real_0_26[rip]
+	ja		atan_real_1
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	je		atan_real_0_0			# atan -0.0 = -0.0
+
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm5,qword ptr atan_p_9_1[rip]
+	
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movsd	xmm14,xmm6				# x2
+	mulsd	xmm5,xmm6				# p2
+
+	mulsd	xmm6,xmm0				# x6
+
+	movlpd	xmm9,qword ptr atan_p_9_2[rip]
+
+	movsd	xmm10,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm7,qword ptr atan_p_9_3[rip]
+	movlpd	xmm12,qword ptr atan_p_9_4[rip]
+
+	mulsd	xmm10,xmm6				# x10
+
+	movsd	xmm2,xmm6				# x6
+	mulsd	xmm7,xmm6				# p6
+	mulsd	xmm6,xmm6				# x12
+
+	mulsd	xmm2,xmm0				# x14
+	mulsd	xmm12,xmm0				# p8
+	mulsd	xmm0,xmm0				# x16
+
+	mulsd	xmm10,qword ptr atan_p_9_5[rip] # p10
+	mulsd	xmm6,qword ptr atan_p_9_6[rip] # p12
+	mulsd	xmm2,qword ptr atan_p_9_7[rip] # p14
+	mulsd	xmm0,qword ptr atan_p_9_8[rip] # p16
+
+	addsd	xmm0,xmm2				# p16+p14
+	addsd	xmm0,xmm6				# p16+p14+p12
+	addsd	xmm0,xmm10				# p16+p14+p12+p10
+	addsd	xmm0,xmm12				# p16+p14+p12+p10+p8
+	addsd	xmm0,xmm7				# p16+p14+p12+p10+p8+p6	
+
+	mulsd	xmm14,xmm1				# x3
+
+	addsd	xmm0,xmm9				# p16+p14+p12+p10+p8+p6+p4
+	addsd	xmm0,xmm5				# p16+p14+p12+p10+p8+p6+p4+p2
+	addsd	xmm0,qword ptr atan_p_9_0[rip] # p16+p14+p12+p10+p8+p6+p4+p2+p0
+
+	mulsd	xmm0,xmm14				# x3*p
+
+	addsd	xmm0,xmm1				# x+x3*p
+
+atan_real_0_0:
+	ret
+
+atan_real_n:
+	ucomisd	xmm0,qword ptr real_m_1_0[rip]
+	jb		atan_real_l_m1			# x<-1.0 | NAN
+
+	ucomisd	xmm0,qword ptr real_m_0_70[rip]
+	jb		atan_real_2n
+
+	movlpd	xmm2,qword ptr real_m_2_0[rip]
+	movlpd	xmm3,qword ptr real_m_0_5[rip]
+
+	lea		rdx,n_atan_0_5_52[rip]
+	jmp		atan_real_1pn
+
+atan_real_1:
+	movlpd	xmm2,qword ptr real_2_0[rip]
+	movlpd	xmm3,qword ptr real_0_5[rip]
+	
+	lea		rdx,atan_0_5_52[rip]
+
+atan_real_1pn:
+	addsd	xmm0,xmm2				# -2+x
+	subsd	xmm3,xmm1				# -0.5-x
+
+	divsd	xmm2,xmm0				# 2/(2+x) = 1/(1+0.5x)
+
+	movlpd	xmm15,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm3				# 0.5-x
+
+	lea		rcx,atan_n_0_5_8_c[rip]
+	movsd	xmm11,xmm3				# (0.5-x)
+	andpd	xmm15,xmm3				# (0.5-x)_h
+
+	subsd	xmm11,xmm15				# (0.5-x)_l
+
+	mulsd	xmm0,xmm2				# y=(0.5-x)*(1/(1+0.5x))
+
+	movsd	xmm4,xmm0				# y
+	mulsd	xmm0,xmm0				# y2
+
+	movlpd	xmm5,qword ptr (atan_n_0_5_8_1-atan_n_0_5_8_c)[rcx]
+
+	movsd	xmm6,xmm0				# y2
+	mulsd	xmm0,xmm0				# y4
+
+	movsd	xmm14,xmm6				# y2
+	mulsd	xmm5,xmm6				# p2
+
+	movlpd	xmm8,qword ptr real_27_bits[rip]
+
+	mulsd	xmm6,xmm0				# y6
+
+	movlpd	xmm9,qword ptr (atan_n_0_5_8_2-atan_n_0_5_8_c)[rcx]
+
+	movsd	xmm10,xmm0				# y4
+	mulsd	xmm0,xmm0				# y8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm12,qword ptr (atan_n_0_5_8_4-atan_n_0_5_8_c)[rcx]
+
+	mulsd	xmm10,xmm6				# y10
+
+	movsd	xmm7,xmm6				# x6
+	mulsd	xmm6,xmm6				# x12
+
+	mulsd	xmm12,xmm0				# p8
+	mulsd	xmm0,xmm7				# x14
+
+	andpd	xmm8,xmm4				# y_h
+
+	mulsd	xmm7,qword ptr (atan_n_0_5_8_3-atan_n_0_5_8_c)[rcx] # p6
+	mulsd	xmm10,qword ptr (atan_n_0_5_8_5-atan_n_0_5_8_c)[rcx] # p10
+	mulsd	xmm6,qword ptr (atan_n_0_5_8_6-atan_n_0_5_8_c)[rcx] # p12
+	mulsd	xmm0,qword ptr (atan_n_0_5_8_7-atan_n_0_5_8_c)[rcx] # p14
+
+	movlpd	xmm13,qword ptr real_0_25[rip]
+	movlpd	xmm1,qword ptr (atan_0_5_real_0_5-atan_0_5_52)[rdx]
+	mulsd	xmm13,xmm8				# 0.25*y_h
+	mulsd	xmm1,xmm8				# 0.5*y_h
+	subsd	xmm3,xmm8				# (0.5-x)-y_h
+
+	addsd	xmm0,xmm6				# p14+p12
+
+	subsd	xmm3,xmm13				# (0.5-x)-1.25*y_h	
+	mulsd	xmm15,xmm1				# (0.5-x)_h*0.5*y_h
+	mulsd	xmm11,xmm1				# (0.5-x)_l*0.5*y_h
+
+	addsd	xmm0,xmm10				# p14+p12+p10
+
+	addsd	xmm3,xmm15				# (0.5-x)-1.25*y_h+(0.5-x)_h*0.5*y_h
+
+	addsd	xmm0,xmm12				# p14+p12+p10+p8
+
+	addsd	xmm3,xmm11				# (0.5-x)-1.25*y_h+(0.5-x)_h*0.5*y_h+(0.5-x)_l*0.5*y_h
+
+	addsd	xmm0,xmm7				# p14+p12+p10+p8+p6
+
+	mulsd	xmm14,xmm4				# y3
+
+	mulsd	xmm3,xmm2				# c=((0.5-x)-1.25*y_h-(0.5-x)_h*0.5*y_h-(0.5-x)_l*0.5*y_h)*(1/(1+0.5x))
+
+	addsd	xmm0,xmm9				# p14+p12+p10+p8+p6+p4
+
+	movsd	xmm11,xmm8				# y_h
+	addsd	xmm8,xmm3				# y_h+c
+
+	addsd	xmm0,xmm5				# p14+p12+p10+p8+p6+p4+p2
+
+	movlpd	xmm7,qword ptr real_40_bits[rip]
+	movlpd	xmm12,qword ptr real_abs_40_bits[rip]
+	andpd	xmm7,xmm8				# (y_h+c)_h
+	andpd	xmm12,xmm8
+
+	addsd	xmm0,qword ptr (atan_n_0_5_8_0-atan_n_0_5_8_c)[rcx] # p14+p12+p10+p8+p6+p4+p2+p0
+
+	subsd	xmm11,xmm7				# y_h-(y_h+c)_h
+	addsd	xmm3,xmm11				# (y_h-(y_h+c)_h)+c
+
+	mulsd	xmm0,xmm14				# y3*p/q
+
+	movlpd	xmm2,qword ptr (atan_0_5_52_l-atan_0_5_52)[rdx]
+	subsd	xmm2,xmm0				# atan_0_5_52_l-y3*p/q
+	movlpd	xmm0,qword ptr (atan_0_5_52-atan_0_5_52)[rdx]
+
+	ucomisd	xmm12,qword ptr real_2_p_m_13[rip]
+	jb		atan_real_1_s
+
+	subsd	xmm0,xmm7				# atan_0_5_52-(y_h+c)_h
+	subsd	xmm2,xmm3				# atan_0_5_52_l-y3*p/q-(y_h+c)_l
+	addsd	xmm0,xmm2				# atan 0_5-y3*p/q-(0.5-x)/(1+x)
+
+	ret
+
+atan_real_1_s:
+	subsd	xmm2,xmm3				# atan_0_5_52_l-y3*p/q-(y_h+c)_l
+	subsd	xmm2,xmm7				# atan_0_5_52_l-y3*p/q-(y_h+c)_l-(y_h+c)_h
+	addsd	xmm0,xmm2				# atan 0_5-y3*p/q-(1-x)/(1+x)
+	ret
+
+atan_real_2n:
+	lea		rcx,atan_n_8_c[rip]
+
+atan_real_2n_:
+	movlpd	xmm2,qword ptr real_m_1_0[rip]
+	movlpd	xmm3,qword ptr real_m_1_0[rip]
+
+	addsd	xmm0,xmm2				# 1+x
+	subsd	xmm3,xmm1				# 1-x
+
+	lea		rdx,n_atan_1_53[rip]
+
+	divsd	xmm2,xmm0				# 1/(1+x)
+
+	movlpd	xmm15,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm3				# 1-x
+
+	movsd	xmm11,xmm3				# (1-x)
+	andpd	xmm15,xmm3				# (1-x)_h
+
+	subsd	xmm11,xmm15				# (1-x)_l
+
+	mulsd	xmm0,xmm2				# y=(1-x)*(1/(1+x))
+
+	movsd	xmm4,xmm0				# y
+	mulsd	xmm0,xmm0				# y2
+	
+	movlpd	xmm5,qword ptr (atan_n_8_1-atan_n_8_c)[rcx]
+
+	movsd	xmm6,xmm0				# y2
+	mulsd	xmm0,xmm0				# y4
+
+	movsd	xmm14,xmm6				# y2
+	mulsd	xmm5,xmm6				# p2
+
+	movlpd	xmm8,qword ptr real_27_bits[rip]
+
+	mulsd	xmm6,xmm0				# y6
+
+	movlpd	xmm9,qword ptr (atan_n_8_2-atan_n_8_c)[rcx]
+
+	movsd	xmm10,xmm0				# y4
+	mulsd	xmm0,xmm0				# y8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm12,qword ptr (atan_n_8_4-atan_n_8_c)[rcx]
+
+	mulsd	xmm10,xmm6				# y10
+
+	movsd	xmm7,xmm6				# x6
+	mulsd	xmm6,xmm6				# x12
+
+	mulsd	xmm12,xmm0				# p8
+	mulsd	xmm0,xmm7				# x14
+
+	andpd	xmm8,xmm4				# y_h
+
+	mulsd	xmm7,qword ptr (atan_n_8_3-atan_n_8_c)[rcx] # p6
+	mulsd	xmm10,qword ptr (atan_n_8_5-atan_n_8_c)[rcx] # p10
+	mulsd	xmm6,qword ptr (atan_n_8_6-atan_n_8_c)[rcx] # p12
+	mulsd	xmm0,qword ptr (atan_n_8_7-atan_n_8_c)[rcx] # p14
+
+	movlpd	xmm13,qword ptr real_2_0[rip]
+	mulsd	xmm13,xmm8				# 2*y_h
+	mulsd	xmm15,xmm8				# (1-x)_h*y_h
+
+	addsd	xmm0,xmm6				# p14+p12
+
+	subsd	xmm3,xmm13				# (1-x)-2*y_h
+	mulsd	xmm11,xmm8				# (1-x)_l*y_h
+
+	addsd	xmm0,xmm10				# p14+p12+p10
+
+	subsd	xmm3,xmm15				# (1-x)-2*y_h-(1-x)_h*y_h
+
+	addsd	xmm0,xmm12				# p14+p12+p10+p8
+
+	subsd	xmm3,xmm11				# (1-x)-2*y_h-(1-x)_h*y_h-(1-x)_h*y_l
+	jmp		atan_real_2pn_
+
+atan_real_2:
+	lea		rcx,atan_n_8_c[rip]
+
+atan_real_2p_:
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	movlpd	xmm3,qword ptr real_1_0[rip]
+
+	addsd	xmm0,xmm2				# 1+x
+	subsd	xmm3,xmm1				# 1-x
+
+	lea		rdx,atan_1_53[rip]
+	
+	divsd	xmm2,xmm0				# 1/(1+x)
+
+	movlpd	xmm15,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm3				# 1-x
+	
+	movsd	xmm11,xmm3				# (1-x)
+	andpd	xmm15,xmm3				# (1-x)_h
+
+	subsd	xmm11,xmm15				# (1-x)_l
+
+	mulsd	xmm0,xmm2				# y=(1-x)*(1/(1+x))
+
+	movsd	xmm4,xmm0				# y
+	mulsd	xmm0,xmm0				# y2
+	
+	movlpd	xmm5,qword ptr (atan_n_8_1-atan_n_8_c)[rcx]
+
+	movsd	xmm6,xmm0				# y2
+	mulsd	xmm0,xmm0				# y4
+
+	movsd	xmm14,xmm6				# y2
+	mulsd	xmm5,xmm6				# p2
+
+	movlpd	xmm8,qword ptr real_27_bits[rip]
+
+	mulsd	xmm6,xmm0				# y6
+
+	movlpd	xmm9,qword ptr (atan_n_8_2-atan_n_8_c)[rcx]
+
+	movsd	xmm10,xmm0				# y4
+	mulsd	xmm0,xmm0				# y8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm12,qword ptr (atan_n_8_4-atan_n_8_c)[rcx]
+
+	mulsd	xmm10,xmm6				# y10
+
+	movsd	xmm7,xmm6				# x6
+	mulsd	xmm6,xmm6				# x12
+
+	mulsd	xmm12,xmm0				# p8
+	mulsd	xmm0,xmm7				# x14
+
+	andpd	xmm8,xmm4				# y_h
+
+	mulsd	xmm7,qword ptr (atan_n_8_3-atan_n_8_c)[rcx] # p6
+	mulsd	xmm10,qword ptr (atan_n_8_5-atan_n_8_c)[rcx] # p10
+	mulsd	xmm6,qword ptr (atan_n_8_6-atan_n_8_c)[rcx] # p12
+	mulsd	xmm0,qword ptr (atan_n_8_7-atan_n_8_c)[rcx] # p14
+
+	movlpd	xmm13,qword ptr real_2_0[rip]
+	mulsd	xmm13,xmm8				# 2*y_h
+	mulsd	xmm15,xmm8				# (1-x)_h*y_h
+
+	addsd	xmm0,xmm6				# p14+p12
+
+	subsd	xmm3,xmm13				# (1-x)-2*y_h
+	mulsd	xmm11,xmm8				# (1-x)_l*y_h
+
+	addsd	xmm0,xmm10				# p14+p12+p10
+
+	addsd	xmm3,xmm15				# (1-x)-2*y_h+(1-x)_h*y_h
+
+	addsd	xmm0,xmm12				# p14+p12+p10+p8
+
+	addsd	xmm3,xmm11				# (1-x)-2*y_h+(1-x)_h*y_h+(1-x)_h*y_l
+
+atan_real_2pn_:
+	addsd	xmm0,xmm7				# p14+p12+p10+p8+p6
+
+	mulsd	xmm14,xmm4				# y3
+	
+	mulsd	xmm3,xmm2				# c=((1-x)-2*y_h-(1-x)_h*y_h-(1-x)_h*y_l)*(1/(1+x))
+
+	addsd	xmm0,xmm9				# p14+p12+p10+p8+p6+p4
+
+	movsd	xmm11,xmm8				# y_h
+	addsd	xmm8,xmm3				# y_h+c
+
+	addsd	xmm0,xmm5				# p14+p12+p10+p8+p6+p4+p2
+
+	movlpd	xmm7,qword ptr (atan_n_8_real_40_bits-atan_n_8_c)[rcx]
+	movlpd	xmm12,qword ptr (atan_n_8_real_abs_40_bits-atan_n_8_c)[rcx]
+	andpd	xmm7,xmm8				# (y_h+c)_h
+	andpd	xmm12,xmm8
+
+	addsd	xmm0,qword ptr (atan_n_8_0-atan_n_8_c)[rcx] # p14+p12+p10+p8+p6+p4+p2+p0
+
+	subsd	xmm11,xmm7				# y_h-(y_h+c)_h
+	addsd	xmm3,xmm11				# (y_h-(y_h+c)_h)+c
+	
+	mulsd	xmm0,xmm14				# y3*p/q
+
+	movlpd	xmm2,qword ptr (atan_1_53_l-atan_1_53)[rdx]
+	subsd	xmm2,xmm0				# pi_d_4_l-y3*p/q
+	movlpd	xmm0,qword ptr (atan_1_53-atan_1_53)[rdx]
+
+	ucomisd	xmm12,qword ptr (atan_n_8_real_2_p_m_12-atan_n_8_c)[rcx]
+	jb		atan_real_2_s
+
+	subsd	xmm0,xmm7				# pi_d_4-(y_h+c)_h
+	subsd	xmm2,xmm3				# pi_d_4_l-y3*p/q-(y_h+c)_l
+	addsd	xmm0,xmm2				# pi/4-y3*p/q-(1-x)/(1+x)
+
+	ret
+
+atan_real_2_s:
+	subsd	xmm2,xmm3				# pi_d_4_l-y3*p/q-(y_h+c)_l
+	subsd	xmm2,xmm7				# pi_d_4_l-y3*p/q-(y_h+c)_l-(y_h+c)_h
+	addsd	xmm0,xmm2				# pi/4-y3*p/q-(1-x)/(1+x)
+	ret
+
+atan_real_l_m1:
+	ucomisd	xmm0,qword ptr real_m_4_0[rip]
+	jb		atan_real_4n			# x<4.0 | NAN
+
+	ucomisd	xmm0,qword ptr real_m_1_4[rip]
+	jb		atan_real_3n
+
+	lea		rcx,atan_p_8_c[rip]
+
+	att_jmp		atan_real_2n_
+
+atan_real_g1:
+	ucomisd	xmm0,qword ptr real_4_0[rip]
+	ja		atan_real_4				# x>4.0
+
+	ucomisd	xmm0,qword ptr real_1_4[rip]
+	ja		atan_real_3
+
+	lea		rcx,atan_p_8_c[rip]
+
+	att_jmp		atan_real_2p_
+
+atan_real_3n:
+	movlpd	xmm2,qword ptr real_m_0_5[rip]
+	movlpd	xmm3,qword ptr real_m_2_0[rip]
+
+	lea		rdx,n_atan_2_0[rip]
+	jmp		atan_real_3pn
+
+atan_real_3:
+	movlpd	xmm2,qword ptr real_0_5[rip]
+	movlpd	xmm3,qword ptr real_2_0[rip]
+
+	lea		rdx,atan_2_0[rip]
+
+atan_real_3pn:
+	addsd	xmm0,xmm2				# 0.5+x
+	subsd	xmm3,xmm1				# 2-x
+	
+	divsd	xmm2,xmm0				# 0.5/(0.5+x) = 1/(1+2*x)
+
+	movlpd	xmm15,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm3				# 2-x
+
+	movsd	xmm11,xmm3				# (2-x)
+	andpd	xmm15,xmm3				# (2-x)_h
+
+	lea		rcx,atan_n_2_8_c[rip]
+	subsd	xmm11,xmm15				# (2-x)_l
+
+	mulsd	xmm0,xmm2				# y=(2-x)*(1/(1+0.5x))
+
+	movsd	xmm4,xmm0				# y
+	mulsd	xmm0,xmm0				# y2
+
+	movlpd	xmm5,qword ptr (atan_n_2_8_1-atan_n_2_8_c)[rcx]
+
+	movsd	xmm6,xmm0				# y2
+	mulsd	xmm0,xmm0				# y4
+
+	movsd	xmm14,xmm6				# y2
+	mulsd	xmm5,xmm6				# p2
+
+	movlpd	xmm8,qword ptr real_27_bits[rip]
+
+	mulsd	xmm6,xmm0				# y6
+
+	movlpd	xmm9,qword ptr (atan_n_2_8_2-atan_n_2_8_c)[rcx]
+
+	movsd	xmm10,xmm0				# y4
+	mulsd	xmm0,xmm0				# y8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm12,qword ptr (atan_n_2_8_4-atan_n_2_8_c)[rcx]
+
+	mulsd	xmm10,xmm6				# y10
+
+	movsd	xmm7,xmm6				# x6
+	mulsd	xmm6,xmm6				# x12
+
+	mulsd	xmm12,xmm0				# p8
+	mulsd	xmm0,xmm7				# x14
+
+	andpd	xmm8,xmm4				# y_h
+
+	mulsd	xmm7,qword ptr (atan_n_2_8_3-atan_n_2_8_c)[rcx] # p6
+	mulsd	xmm10,qword ptr (atan_n_2_8_5-atan_n_2_8_c)[rcx] # p10
+	mulsd	xmm6,qword ptr (atan_n_2_8_6-atan_n_2_8_c)[rcx] # p12
+	mulsd	xmm0,qword ptr (atan_n_2_8_7-atan_n_2_8_c)[rcx] # p14
+
+	movlpd	xmm13,qword ptr real_4_0[rip]
+	movlpd	xmm1,qword ptr (atan_2_0_real_2_0-atan_2_0)[rdx]
+	mulsd	xmm13,xmm8				# 4*y_h
+	mulsd	xmm1,xmm8				# 2*y_h
+
+	addsd	xmm0,xmm6				# p14+p12
+
+	subsd	xmm3,xmm13				# (2-x)-4*y_h
+	mulsd	xmm15,xmm1				# (2-x)_h*2*y_h
+	mulsd	xmm11,xmm1				# (2-x)_l*2*y_h
+	subsd	xmm3,xmm8				# (2-x)-y_h
+
+	addsd	xmm0,xmm10				# p14+p12+p10
+
+	addsd	xmm3,xmm15				# (2-x)-5*y_h+(2-x)_h*2*y_h
+
+	addsd	xmm0,xmm12				# p14+p12+p10+p8
+
+	addsd	xmm3,xmm11				# (2-x)-5*y_h+(2-x)_h*2*y_h+(2-x)_l*2*y_h
+
+	addsd	xmm0,xmm7				# p14+p12+p10+p8+p6
+
+	mulsd	xmm14,xmm4				# y3
+
+	mulsd	xmm3,xmm2				# c=((2-x)-5*y_h-(2-x)_h*2*y_h-(2-x)_l*2*y_h)*(1/(1+0.5x))
+
+	addsd	xmm0,xmm9				# p14+p12+p10+p8+p6+p4
+
+	movsd	xmm11,xmm8				# y_h
+	addsd	xmm8,xmm3				# y_h+c
+
+	addsd	xmm0,xmm5				# p14+p12+p10+p8+p6+p4+p2
+
+	movlpd	xmm7,qword ptr real_39_bits[rip]
+	movlpd	xmm12,qword ptr real_abs_39_bits[rip]
+	andpd	xmm7,xmm8
+	andpd	xmm12,xmm8				# (y_h+c)_h
+
+	addsd	xmm0,qword ptr (atan_n_2_8_0-atan_n_2_8_c)[rcx] # p14+p12+p10+p8+p6+p4+p2+p0
+
+	subsd	xmm11,xmm7				# y_h-(y_h+c)_h
+	addsd	xmm3,xmm11				# (y_h-(y_h+c)_h)+c
+	
+	mulsd	xmm0,xmm14				# y3*p/q
+
+	movlpd	xmm2,qword ptr (atan_2_0_l-atan_2_0)[rdx]
+	subsd	xmm2,xmm0				# atan_2_0_l-y3*p/q
+	movlpd	xmm0,qword ptr (atan_2_0-atan_2_0)[rdx]
+
+	ucomisd	xmm12,qword ptr real_2_p_m_11[rip]
+	jb		atan_real_3_s
+
+	subsd	xmm0,xmm7				# atan_2_0-(y_h+c)_h
+	subsd	xmm2,xmm3				# atan_2_0_l-y3*p/q-(y_h+c)_l
+	addsd	xmm0,xmm2				# atan 2_0-y3*p/q-(2-x)/(1+x)
+
+	ret
+
+atan_real_3_s:
+	subsd	xmm2,xmm3				# atan_2_0_l-y3*p/q-(y_h+c)_l
+	subsd	xmm2,xmm7				# atan_2_0_l-y3*p/q-(y_h+c)_l-(y_h+c)_h
+	addsd	xmm0,xmm2				# atan 2_0-y3*p/q-(1-x)/(1+x)
+	ret
+
+atan_real_4n:
+	ucomisd	xmm0,qword ptr real_atan_m_large[rip]
+	jb		atan_real_m_large_or_nan # x<-5805358775541310.0840 | NAN
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	mulsd	xmm0,xmm0				# x2
+
+	lea		rdx,real_m_pi_d_2[rip]
+
+	jmp		atan_real_4pn
+
+atan_real_4:
+	ucomisd	xmm0,qword ptr real_atan_large[rip]
+	ja		atan_real_large			# x>5805358775541310.0840
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	mulsd	xmm0,xmm0				# x2
+
+	lea		rdx,real_pi_d_2[rip]
+
+atan_real_4pn:
+	divsd	xmm2,xmm1				# 1/x
+
+	lea		rcx,atan4_p_c[rip]
+
+	movlpd	xmm4,qword ptr (atan4_q_4-atan4_p_c)[rcx]
+	movlpd	xmm3,qword ptr (atan4_p_3-atan4_p_c)[rcx]
+
+	movsd	xmm5,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm4,xmm5				# q2
+	mulsd	xmm3,xmm5				# p2
+
+	movlpd	xmm9,qword ptr (atan4_q_3-atan4_p_c)[rcx]
+	movlpd	xmm8,qword ptr (atan4_p_2-atan4_p_c)[rcx]
+
+	movsd	xmm6,xmm0				# x4
+	mulsd	xmm0,xmm5				# x6
+
+	mulsd	xmm9,xmm6				# q4
+	mulsd	xmm8,xmm6				# p4
+
+	movlpd	xmm11,qword ptr (atan4_q_2-atan4_p_c)[rcx]
+	movlpd	xmm10,qword ptr (atan4_p_1-atan4_p_c)[rcx]
+
+	addsd	xmm4,qword ptr real_1_0[rip]	# 1+q2
+	addsd	xmm3,qword ptr (atan4_p_4-atan4_p_c)[rcx] # p0+p2
+	
+	movsd	xmm7,xmm6				# x4
+	mulsd	xmm6,xmm6				# x8
+
+	mulsd	xmm11,xmm0				# q6
+	mulsd	xmm10,xmm0				# p6
+
+	addsd	xmm4,xmm9				# 1+q2+q4
+	addsd	xmm3,xmm8				# p0+p2+p4
+
+	mulsd	xmm0,xmm7				# x10
+
+	movlpd	xmm9,qword ptr (atan4_q_1-atan4_p_c)[rcx]
+	movlpd	xmm8,qword ptr (atan4_p_0-atan4_p_c)[rcx]
+
+	addsd	xmm4,xmm11				# 1+q2+q4+q6
+	addsd	xmm3,xmm10				# p0+p2+p4+p6
+
+	mulsd	xmm9,xmm6				# q8
+	mulsd	xmm8,xmm6				# p8
+
+	mulsd	xmm0,qword ptr (atan4_q_0-atan4_p_c)[rcx]# q10
+
+	movlpd	xmm5,qword ptr real_26_bits[rip]
+	movsd	xmm6,xmm1				# x
+	movlpd	xmm7,qword ptr real_26_bits[rip]
+	andpd	xmm5,xmm1				# x_h
+	subsd	xmm6,xmm5				# x_l
+	andpd	xmm7,xmm2				# (1/x)_h
+
+	addsd	xmm4,xmm9				# 1+q2+q4+q6+q8
+	addsd	xmm3,xmm8				# p0+p2+p4+p6+p8
+
+	movlpd	xmm8,qword ptr real_1_0[rip]
+	mulsd	xmm5,xmm7				# x_h*(1/x)_h
+	mulsd	xmm6,xmm7				# x_l*(1/x)_h
+
+	addsd	xmm4,xmm0				# 1+q2+q4+q6+q8+q10
+
+	subsd	xmm8,xmm5				# 1-x_h*(1/x)_h
+	
+	subsd	xmm8,xmm6				# 1-x_h*(1/x)_h-x_l*(1/x)_h
+
+	mulsd	xmm8,xmm2				# (1-x_h*(1/x)_h-x_l*(1/x)_h)*(1/x)
+
+	divsd	xmm3,xmm4				# p/q
+
+	movlpd	xmm1,qword ptr (real_pi_d_2_l-real_pi_d_2)[rdx]
+	movlpd	xmm0,qword ptr (real_pi_d_2-real_pi_d_2)[rdx]
+
+	mulsd	xmm3,xmm2				# (p/q)*(1/x)
+
+	subsd	xmm1,xmm3				# pi_d_2_l-(p/q)*(1/x)
+	subsd	xmm0,xmm7				# pi_d_2-(1/x)_h
+	subsd	xmm1,xmm8				# pi_d_2_l-(p/q)*(1/x)-(1/x)_l
+	addsd	xmm0,xmm1
+	ret
+
+atan_real_4poly:
+	movlpd	xmm3,qword ptr real_1_0[rip]
+	divsd	xmm3,xmm0				# 1/x
+
+	movlpd	xmm5,qword ptr real_26_bits[rip]
+	movsd	xmm6,xmm0				# x
+	movlpd	xmm7,qword ptr real_26_bits[rip]
+	andpd	xmm5,xmm0				# x_h
+
+	subsd	xmm6,xmm5				# x_l
+
+	movsd	xmm15,xmm3				# x
+	mulsd	xmm3,xmm3				# x2
+
+	movlpd	xmm1,qword ptr atan_p_9_1[rip]
+	
+	movsd	xmm11,xmm3				# x2
+	mulsd	xmm3,xmm3				# x4
+
+	movsd	xmm14,xmm11				# x2
+	mulsd	xmm1,xmm11				# p2
+
+	mulsd	xmm11,xmm3				# x6
+
+	movlpd	xmm9,qword ptr atan_p_9_2[rip]
+
+	movsd	xmm10,xmm3				# x4
+	mulsd	xmm3,xmm3				# x8
+
+	mulsd	xmm9,xmm10				# p4
+
+	movlpd	xmm4,qword ptr atan_p_9_3[rip]
+	movlpd	xmm12,qword ptr atan_p_9_4[rip]
+
+	mulsd	xmm10,xmm11				# x10
+
+	movsd	xmm2,xmm11				# x6
+	mulsd	xmm4,xmm11				# p6
+	mulsd	xmm11,xmm11				# x12
+
+	mulsd	xmm2,xmm3				# x14
+	mulsd	xmm12,xmm3				# p8
+	mulsd	xmm3,xmm3				# x16
+
+	mulsd	xmm10,qword ptr atan_p_9_5[rip] # p10
+	mulsd	xmm11,qword ptr atan_p_9_6[rip] # p12
+	mulsd	xmm2,qword ptr atan_p_9_7[rip] # p14
+	mulsd	xmm3,qword ptr atan_p_9_8[rip] # p16
+
+	addsd	xmm3,xmm2				# p16+p14
+	addsd	xmm3,xmm11				# p16+p14+p12
+	addsd	xmm3,xmm10				# p16+p14+p12+p10
+
+	andpd	xmm7,xmm15				# (1/x)_h
+
+	addsd	xmm3,xmm12				# p16+p14+p12+p10+p8
+
+	movlpd	xmm8,qword ptr real_1_0[rip]
+	mulsd	xmm5,xmm7				# x_h*(1/x)_h
+	mulsd	xmm6,xmm7				# x_l*(1/x)_h
+
+	addsd	xmm3,xmm4				# p16+p14+p12+p10+p8+p6	
+
+	mulsd	xmm14,xmm15				# x3
+
+	subsd	xmm8,xmm5				# 1-x_h*(1/x)_h
+
+	addsd	xmm3,xmm9				# p16+p14+p12+p10+p8+p6+p4
+	
+	subsd	xmm8,xmm6				# 1-x_h*(1/x)_h-x_l*(1/x)_h
+
+	addsd	xmm3,xmm1				# p16+p14+p12+p10+p8+p6+p4+p2
+
+	mulsd	xmm8,xmm15				# (1-x_h*(1/x)_h-x_l*(1/x)_h)*(1/x)
+
+	addsd	xmm3,qword ptr atan_p_9_0[rip] # p16+p14+p12+p10+p8+p6+p4+p2+p0
+
+	movlpd	xmm1,qword ptr real_pi_d_2_l[rip]
+	movlpd	xmm0,qword ptr real_pi_d_2[rip]
+
+	mulsd	xmm3,xmm14				# x3*p
+
+	subsd	xmm1,xmm3				# pi_d_2_l-(p/q)*(1/x)
+	subsd	xmm0,xmm7				# pi_d_2-(1/x)_h
+	subsd	xmm1,xmm8				# pi_d_2_l-(p/q)*(1/x)-(1/x)_l
+	addsd	xmm0,xmm1
+	ret
+
+atan_real_m_large_or_nan:
+	jp		atan_real_nan
+
+	movlpd	xmm0,qword ptr real_m_pi_d_2[rip]
+	ret
+
+atan_real_nan:
+	addsd	xmm0,xmm0
+	ret
+
+atan_real_large:
+	movlpd	xmm0,qword ptr real_pi_d_2[rip]
+	ret
+
+	.globl	exp_real
+
+exp_real:
+	ucomisd	xmm0,qword ptr real_ln_2_t_0_5[rip]
+	ja	exp_real_2				# x>0.3.4657359027997265471
+
+	ucomisd	xmm0,qword ptr real_n_ln_2_t_0_5[rip]
+	jb	exp_real_3_n			# x<-0.34657359027997265471 | nan
+
+	movlpd	xmm1,qword ptr n_45_d_256[rip]
+	movlpd	xmm2,qword ptr n_m_45_d_256[rip]
+
+	ucomisd	xmm0,xmm1
+	ja	exp_real_p1
+
+	ucomisd	xmm0,xmm2
+	jb	exp_real_n1
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+
+	lea	rcx,exp_p0_c[rip]
+	lea	rdx,exp_m0_c[rip]
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	cmovb	rcx,rdx
+
+	movlpd	xmm6,qword ptr (exp_p0_2-exp_p0_c)[rcx]
+	movlpd	xmm7,qword ptr (exp_p0_4-exp_p0_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm6,xmm2				# p2
+
+	movsd	xmm3,xmm1				# x
+	mulsd	xmm1,xmm2				# x3
+
+	movlpd	xmm8,qword ptr (exp_p0_3-exp_p0_c)[rcx]
+	mulsd	xmm7,xmm0				# p4
+	
+	mulsd	xmm8,xmm1				# p3
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movsd	xmm5,xmm2				# x2
+	mulsd	xmm2,xmm1				# x5	
+
+	mulsd	xmm5,xmm4				# x6
+	mulsd	xmm1,xmm4				# x7
+	mulsd	xmm4,xmm2				# x9
+
+	mulsd	xmm0,qword ptr (exp_p0_8-exp_p0_c)[rcx] # p8
+	mulsd	xmm2,qword ptr (exp_p0_5-exp_p0_c)[rcx]	# p5
+	mulsd	xmm5,qword ptr (exp_p0_6-exp_p0_c)[rcx]	# p6
+	mulsd	xmm1,qword ptr (exp_p0_7-exp_p0_c)[rcx]	# p7
+	mulsd	xmm4,qword ptr (exp_p0_9-exp_p0_c)[rcx]	# p9
+
+	addsd	xmm0,xmm4				# p9+p8
+
+	movlpd	xmm9,qword ptr real_26_bits[rip]
+	movlpd	xmm10,qword ptr real_1_0[rip]
+
+	addsd	xmm0,xmm1				# p9+p8+p7
+	addsd	xmm0,xmm5				# p9+p8+p7+p6
+
+	andpd	xmm9,xmm3				# x_h
+
+	addsd	xmm0,xmm2				# p9+p8+p7+p6+p5
+
+	addsd	xmm9,xmm10				# 1+x_h
+
+	addsd	xmm0,xmm7				# p9+p8+p7+p6+p5+p4
+
+	subsd	xmm10,xmm9				# 1-(1+x_h)
+
+	addsd	xmm0,xmm8				# p9+p8+p7+p6+p5+p4+p3
+
+	addsd	xmm10,xmm3				# (1-(1+x_h))+x
+
+	addsd	xmm0,xmm6				# p9+p8+p7+p6+p5+p4+p3+p2
+
+	addsd	xmm0,xmm10
+	addsd	xmm0,xmm9
+	ret
+
+exp_real_n1:
+	subsd	xmm0,xmm2
+
+	lea		rcx,exp_m1_c[rip]
+	jmp		exp_real_np1
+
+exp_real_p1:
+	subsd	xmm0,xmm1
+
+	lea		rcx,exp_p1_c[rip]
+
+exp_real_np1:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm6,qword ptr (exp_p1_2-exp_p1_c)[rcx]
+	movlpd	xmm7,qword ptr (exp_p1_4-exp_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm6,xmm2				# p2
+
+	movsd	xmm3,xmm1				# x
+	mulsd	xmm1,xmm2				# x3
+
+	movlpd	xmm8,qword ptr (exp_p1_3-exp_p1_c)[rcx]
+	mulsd	xmm7,xmm0				# p4
+	
+	mulsd	xmm8,xmm1				# p3
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movsd	xmm5,xmm2				# x2
+	mulsd	xmm2,xmm1				# x5	
+
+	mulsd	xmm5,xmm4				# x6
+	mulsd	xmm1,xmm4				# x7
+	mulsd	xmm4,xmm2				# x9
+
+	mulsd	xmm0,qword ptr (exp_p1_8-exp_p1_c)[rcx] # p8
+	mulsd	xmm2,qword ptr (exp_p1_5-exp_p1_c)[rcx]	# p5
+	mulsd	xmm5,qword ptr (exp_p1_6-exp_p1_c)[rcx]	# p6
+	mulsd	xmm1,qword ptr (exp_p1_7-exp_p1_c)[rcx]	# p7
+	mulsd	xmm4,qword ptr (exp_p1_9-exp_p1_c)[rcx]	# p9
+
+	addsd	xmm0,xmm4				# p9+p8
+
+	movlpd	xmm9,qword ptr real_26_bits[rip]
+
+	addsd	xmm0,xmm1				# p9+p8+p7
+
+	movsd	xmm11,xmm3				# x
+	andpd	xmm9,xmm3				# x_h
+	mulsd	xmm3,qword ptr (exp_p1_1_l-exp_p1_c)[rcx] # x*c1_l
+
+	addsd	xmm0,xmm5				# p9+p8+p7+p6
+
+	subsd	xmm11,xmm9				# x_l
+	mulsd	xmm9,qword ptr (exp_p1_1_h-exp_p1_c)[rcx] # x_h*c1_h
+
+	addsd	xmm0,xmm2				# p9+p8+p7+p6+p5
+	
+	movlpd	xmm10,qword ptr (exp_p1_0-exp_p1_c)[rcx]
+	mulsd	xmm11,qword ptr (exp_p1_1_h-exp_p1_c)[rcx] # x_l*c1_h
+	addsd	xmm10,xmm9				# x_h*c1_h+c0
+	movlpd	xmm12,qword ptr (exp_p1_0-exp_p1_c)[rcx]
+
+	addsd	xmm0,xmm7				# p9+p8+p7+p6+p5+p4
+
+	addsd	xmm3,xmm11				# x*c1_l+x_l*c1_h
+	subsd	xmm12,xmm10				# c0-(x_h*c1_h+c0)
+
+	addsd	xmm0,xmm8				# p9+p8+p7+p6+p5+p4+p3
+
+	addsd	xmm12,xmm9				# (c0-(x_h*c1_h+c0))+x_h*c1_h
+
+	addsd	xmm0,xmm6				# p9+p8+p7+p6+p5+p4+p3+p2
+
+	addsd	xmm3,xmm12				# (c0-(x_h*c1_h+c0))+x_h*c1_h+x*c1_l+x_l*c1_h
+
+	addsd	xmm0,qword ptr (exp_p1_0_l-exp_p1_c)[rcx]
+
+	addsd	xmm0,xmm3
+	addsd	xmm0,xmm10
+	ret
+
+exp_real_2:
+	ucomisd	xmm0,qword ptr real_ln_2_t_1_5[rip]
+	ja		exp_real_3
+
+	subsd	xmm0,qword ptr real_ln2_42[rip]
+	movlpd	xmm14,qword ptr real_ln2_42_l[rip]
+
+	movsd	xmm13,xmm0
+	subsd	xmm0,xmm14
+
+	movlpd	xmm15,qword ptr real_2_0[rip]
+
+exp_real_:
+	movlpd	xmm1,qword ptr n_45_d_256[rip]
+	movlpd	xmm2,qword ptr n_m_45_d_256[rip]
+
+	ucomisd	xmm0,xmm1
+	ja		exp_real_p1_
+
+	ucomisd	xmm0,xmm2
+	jb		exp_real_n1_
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+
+	lea		rcx,exp_p0_c[rip]
+	lea		rdx,exp_m0_c[rip]
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	cmovb	rcx,rdx
+
+	movlpd	xmm6,qword ptr (exp_p0_2-exp_p0_c)[rcx]
+	movlpd	xmm7,qword ptr (exp_p0_4-exp_p0_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm6,xmm2				# p2
+
+	movsd	xmm3,xmm1				# x
+	mulsd	xmm1,xmm2				# x3
+
+	movlpd	xmm8,qword ptr (exp_p0_3-exp_p0_c)[rcx]
+	mulsd	xmm7,xmm0				# p4
+	
+	mulsd	xmm8,xmm1				# p3
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movsd	xmm5,xmm2				# x2
+	mulsd	xmm2,xmm1				# x5	
+
+	mulsd	xmm5,xmm4				# x6
+	mulsd	xmm1,xmm4				# x7
+	mulsd	xmm4,xmm2				# x9
+
+	mulsd	xmm0,qword ptr (exp_p0_8-exp_p0_c)[rcx] # p8
+	mulsd	xmm2,qword ptr (exp_p0_5-exp_p0_c)[rcx]	# p5
+	mulsd	xmm5,qword ptr (exp_p0_6-exp_p0_c)[rcx]	# p6
+	mulsd	xmm1,qword ptr (exp_p0_7-exp_p0_c)[rcx]	# p7
+	mulsd	xmm4,qword ptr (exp_p0_9-exp_p0_c)[rcx]	# p9
+
+	addsd	xmm0,xmm4				# p9+p8
+
+	movlpd	xmm9,qword ptr real_26_bits[rip]
+	movlpd	xmm10,qword ptr real_1_0[rip]
+
+	addsd	xmm0,xmm1				# p9+p8+p7
+
+	andpd	xmm9,xmm3				# x_h
+
+	addsd	xmm0,xmm5				# p9+p8+p7+p6
+
+	addsd	xmm9,xmm10				# 1+x_h
+
+	addsd	xmm0,xmm2				# p9+p8+p7+p6+p5
+
+	subsd	xmm10,xmm9				# 1-(1+x_h)
+
+	addsd	xmm0,xmm7				# p9+p8+p7+p6+p5+p4
+
+	addsd	xmm10,xmm13				# (1-(1+x_h))+xh
+
+	addsd	xmm0,xmm8				# p9+p8+p7+p6+p5+p4+p3
+
+	subsd	xmm10,xmm14				# (1-(1+x_h))+xh-nxl
+
+	addsd	xmm0,xmm6				# p9+p8+p7+p6+p5+p4+p3+p2
+
+	addsd	xmm0,xmm10
+	addsd	xmm0,xmm9
+
+	mulsd	xmm0,xmm15
+	ret
+
+exp_real_n1_:
+	subsd	xmm0,xmm2
+	subsd	xmm13,xmm2
+
+	lea		rcx,exp_m1_c[rip]
+	jmp		exp_real_np1_
+
+exp_real_p1_:
+	subsd	xmm0,xmm1
+	subsd	xmm13,xmm1
+
+	lea		rcx,exp_p1_c[rip]
+
+exp_real_np1_:
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm6,qword ptr (exp_p1_2-exp_p1_c)[rcx]
+	movlpd	xmm7,qword ptr (exp_p1_4-exp_p1_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm6,xmm2				# p2
+
+	movsd	xmm3,xmm1				# x
+	mulsd	xmm1,xmm2				# x3
+
+	movlpd	xmm8,qword ptr (exp_p1_3-exp_p1_c)[rcx]
+	mulsd	xmm7,xmm0				# p4
+	
+	mulsd	xmm8,xmm1				# p3
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movsd	xmm5,xmm2				# x2
+	mulsd	xmm2,xmm1				# x5	
+
+	mulsd	xmm5,xmm4				# x6
+	mulsd	xmm1,xmm4				# x7
+	mulsd	xmm4,xmm2				# x9
+
+	mulsd	xmm0,qword ptr (exp_p1_8-exp_p1_c)[rcx] # p8
+	mulsd	xmm2,qword ptr (exp_p1_5-exp_p1_c)[rcx]	# p5
+	mulsd	xmm5,qword ptr (exp_p1_6-exp_p1_c)[rcx]	# p6
+	mulsd	xmm1,qword ptr (exp_p1_7-exp_p1_c)[rcx]	# p7
+	mulsd	xmm4,qword ptr (exp_p1_9-exp_p1_c)[rcx]	# p9
+
+	movlpd	xmm9,qword ptr real_26_bits[rip]
+
+	addsd	xmm0,xmm4				# p9+p8
+
+	movsd	xmm11,xmm13				# xh
+	andpd	xmm9,xmm3				# x_h
+	mulsd	xmm3,qword ptr (exp_p1_1_l-exp_p1_c)[rcx] # x*c1_l
+
+	addsd	xmm0,xmm1				# p9+p8+p7
+
+	subsd	xmm11,xmm9				# xh-x_h
+
+	addsd	xmm0,xmm5				# p9+p8+p7+p6
+
+	subsd	xmm11,xmm14				# x_l=(xh-x_h)-nxl
+
+	mulsd	xmm9,qword ptr (exp_p1_1_h-exp_p1_c)[rcx] # x_h*c1_h
+
+	addsd	xmm0,xmm2				# p9+p8+p7+p6+p5
+	
+	movlpd	xmm10,qword ptr (exp_p1_0-exp_p1_c)[rcx]
+	mulsd	xmm11,qword ptr (exp_p1_1_h-exp_p1_c)[rcx] # x_l*c1_h
+	addsd	xmm10,xmm9				# x_h*c1_h+c0
+	movlpd	xmm12,qword ptr (exp_p1_0-exp_p1_c)[rcx]
+
+	addsd	xmm0,xmm7				# p9+p8+p7+p6+p5+p4
+
+	addsd	xmm3,xmm11				# x*c1_l+x_l*c1_h
+	subsd	xmm12,xmm10				# c0-(x_h*c1_h+c0)
+
+	addsd	xmm0,xmm8				# p9+p8+p7+p6+p5+p4+p3
+
+	addsd	xmm12,xmm9				# (c0-(x_h*c1_h+c0))+x_h*c1_h
+
+	addsd	xmm0,xmm6				# p9+p8+p7+p6+p5+p4+p3+p2
+
+	addsd	xmm3,xmm12				# (c0-(x_h*c1_h+c0))+x_h*c1_h+x*c1_l+x_l*c1_h
+
+	addsd	xmm0,qword ptr (exp_p1_0_l-exp_p1_c)[rcx]
+
+	addsd	xmm0,xmm3
+	addsd	xmm0,xmm10
+
+	mulsd	xmm0,xmm15
+	ret
+
+exp_real_3_n:
+	movlpd	xmm1,qword ptr real_1_d_ln2[rip]
+	mulsd	xmm1,xmm0
+	movq	xmm4,qword ptr m_round_c[rip]
+	subsd	xmm1,qword ptr real_0_5[rip]
+
+	ucomisd	xmm1,qword ptr real_m_1022[rip]
+	jae		exp_real_3_pn
+
+	ucomisd	xmm1,qword ptr real_m_1076[rip]
+	jb		exp_real_3_n_large
+
+	movq	xmm9,qword ptr qw_1078[rip]
+
+	call	exp_real_3_pn_
+
+	mulsd	xmm0,qword ptr real_2_p_m_55[rip]
+	ret
+
+exp_real_3:
+	movlpd	xmm1,qword ptr real_1_d_ln2[rip]
+	mulsd	xmm1,xmm0
+	movq	xmm4,qword ptr round_c[rip]
+	addsd	xmm1,qword ptr real_0_5[rip]
+
+	ucomisd	xmm1,qword ptr real_1024[rip]
+	jae	exp_real_3_large
+
+exp_real_3_pn:
+	movq	xmm9,qword ptr qw_1023[rip]
+
+exp_real_3_pn_:
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2
+
+	movlpd	xmm13,qword ptr real_ln2_42[rip]
+	movlpd	xmm14,qword ptr real_ln2_42_l[rip]
+
+	mulsd	xmm13,xmm1
+	mulsd	xmm14,xmm1
+
+	cvtpd2dq xmm15,xmm1
+	subsd	xmm0,xmm13
+
+	movsd	xmm13,xmm0
+	subsd	xmm0,xmm14
+
+	paddq	xmm15,xmm9
+	psllq	xmm15,52
+	att_jmp		exp_real_
+
+exp_real_3_large:
+	ucomisd	xmm1,qword ptr real_1025[rip]
+	jae		exp_real_3_larger
+
+	movq	xmm9,qword ptr qw_1022[rip]
+
+	att_call	exp_real_3_pn_
+
+	mulsd	xmm0,qword ptr real_2_0[rip]
+	ret
+
+exp_real_3_larger:
+	ucomisd	xmm0,qword ptr real_max[rip]
+	ja		exp_real_3_inf_or_nan
+	movlpd	xmm0,qword ptr qword ptr real_2_p_1023[rip]
+	mulsd	xmm0,xmm0
+exp_real_3_inf_or_nan:
+	ret
+
+exp_real_3_n_large:
+	att_jp		exp_real_3_inf_or_nan	# nan
+	movlpd	xmm0,qword ptr real_0_0[rip]
+	ret
+
+	.globl	ln_real
+
+ln_real:
+	ucomisd	xmm0,qword ptr real_0_8243[rip]
+	jb		ln_real_small			# x<0.82436063535006407342 | nan
+	ucomisd	xmm0,qword ptr real_1_6487[rip]
+	ja		ln_real_large			# x>1.6487212707001281468
+	movlpd	xmm2,qword ptr real_1_0[rip]
+
+	ucomisd	xmm0,qword ptr real_7_d_6[rip]
+	ja		ln_real_a_7_d_6
+
+	ucomisd	xmm0,xmm2
+
+	subsd	xmm0,xmm2
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	jae		ln_real_0
+	jmp		ln_real_s
+
+ln_real_small:
+	ucomisd	xmm0,qword ptr real_2_p_m_1022[rip]
+	jae		ln_real_
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	jbe		ln_real_less_or_equal_0_or_nan
+
+# denormalized number
+
+	mulsd	xmm0,qword ptr real_2_p_55[rip]
+
+	movlpd	xmm1,qword ptr mask_significand[rip]
+	andpd	xmm1,xmm0
+	psrlq	xmm0,52
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	movlpd	xmm3,qword ptr real_0_5[rip]
+	orpd	xmm2,xmm1
+	orpd	xmm3,xmm1
+
+	ucomisd	xmm2,qword ptr real_1_6487[rip]
+	jbe		ln_real_dn_g1
+
+	subsd	xmm3,qword ptr real_1_0[rip]
+	psubq	xmm0,xmmword ptr qw_1077[rip]
+	jmp		ln_real_l1_
+
+ln_real_dn_g1:
+	ucomisd	xmm2,qword ptr real_7_d_6[rip]
+
+	subsd	xmm2,qword ptr real_1_0[rip]
+	psubq	xmm0,xmmword ptr qw_1078[rip]
+
+	ja		ln_real_g_7_d_6
+	jmp		ln_real_g1_
+
+ln_real_large:
+	ucomisd	xmm0,qword ptr real_max[rip]
+	ja		ln_real_inf_or_nan
+
+ln_real_:
+	movlpd	xmm1,qword ptr mask_significand[rip]
+	andpd	xmm1,xmm0
+	psrlq	xmm0,52
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	movlpd	xmm3,qword ptr real_0_5[rip]
+	orpd	xmm2,xmm1
+	orpd	xmm3,xmm1
+
+	ucomisd	xmm2,qword ptr real_1_6487[rip]
+	jbe		ln_real_g1
+	jmp		ln_real_l1
+
+ln_real_0:
+	movlpd	xmm3,qword ptr ln_b_q_1[rip]
+	movlpd	xmm2,qword ptr ln_b_p_1[rip]
+
+	mulsd	xmm3,xmm1				# q1
+	mulsd	xmm2,xmm1				# p1
+	movsd	xmm7,xmm1				# x
+	movlpd	xmm5,qword ptr ln_b_q_2[rip]
+	movlpd	xmm4,qword ptr ln_b_p_2[rip]
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movsd	xmm11,xmm1				# x
+	mulsd	xmm7,xmm6				# x3
+
+	movlpd	xmm9,qword ptr ln_b_q_3[rip]
+	movlpd	xmm8,qword ptr ln_b_p_3[rip]
+
+	mulsd	xmm11,xmm0				# x5
+
+	mulsd	xmm5,xmm6				# q2
+	mulsd	xmm4,xmm6				# p2
+
+	movlpd	xmm10,qword ptr ln_b_q_4[rip]
+
+	mulsd	xmm9,xmm7				# q3
+	mulsd	xmm8,xmm7				# p3
+	
+	mulsd	xmm10,xmm0				# q4
+	mulsd	xmm0,qword ptr ln_b_p_4[rip]	# p4
+
+	addsd	xmm10,xmm11				# q4+x5
+
+	addsd	xmm0,xmm8				# p3+p4
+	addsd	xmm9,xmm10				# q3+q4+x5
+
+	movlpd	xmm6,qword ptr real_17_bits[rip]
+
+	addsd	xmm0,xmm4				# p2+p3+p4	
+	addsd	xmm5,xmm9				# q2+q3+q4+x5
+
+	andpd	xmm6,xmm1				# x_17_h
+	movsd	xmm8,xmm1				# x
+	movlpd	xmm4,qword ptr real_0_5[rip]
+
+	addsd	xmm0,xmm2				# p1+p2+p3+p4
+	addsd	xmm3,xmm5				# q1+q2+q3+q4+x5
+
+	subsd	xmm8,xmm6				# x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h
+
+	addsd	xmm0,qword ptr ln_b_p_0[rip]	# p0+p1+p2+p3+p4
+	addsd	xmm3,qword ptr ln_b_q_0[rip]	# q0+q1+q2+q3+q4+x5
+
+	mulsd	xmm8,qword ptr real_0_5[rip]	# 0.5*x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h*x_17_h
+	addsd	xmm6,xmm1				# x+x_17_h
+
+	ucomisd	xmm1,qword ptr real_2_p_m_16[rip]
+	jb		ln_real_0_s
+
+	divsd	xmm0,xmm3				# p/q
+
+	subsd	xmm1,xmm4				# x-0.5*x_17_h*x_17_h
+	mulsd	xmm6,xmm8				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+	subsd	xmm0,xmm6				# (p/q)*x3-0.5x2_l
+	addsd	xmm0,xmm1				# (p/q)*x3-0.5x2+x
+	ret
+
+ln_real_0_s:
+	divsd	xmm0,xmm3				# p/q
+
+	mulsd	xmm6,xmm8				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+	subsd	xmm0,xmm6				# (p/q)*x3-0.5x2_l
+	subsd	xmm0,xmm4				# (p/q)*x3-0.5x2
+	addsd	xmm0,xmm1				# (p/q)*x3-0.5x2+x
+	ret
+
+ln_real_a_7_d_6:
+	subsd	xmm0,xmm2
+
+	mulsd	xmm0,qword ptr real_0_75[rip]
+	movlpd	xmm1,qword ptr real_0_25[rip]
+
+	lea		rcx,ln_b_c[rip]
+	lea		rdx,ln_s_c[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	cmovb	rcx,rdx
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm3,qword ptr (ln_b_q_1-ln_b_c)[rcx]
+	movlpd	xmm2,qword ptr (ln_b_p_1-ln_b_c)[rcx]
+
+	mulsd	xmm3,xmm1				# q1
+	mulsd	xmm2,xmm1				# p1
+	movsd	xmm7,xmm1				# x
+	movlpd	xmm5,qword ptr (ln_b_q_2-ln_b_c)[rcx]
+	movlpd	xmm4,qword ptr (ln_b_p_2-ln_b_c)[rcx]
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movsd	xmm11,xmm1				# x
+	mulsd	xmm7,xmm6				# x3
+
+	movlpd	xmm9,qword ptr (ln_b_q_3-ln_b_c)[rcx]
+	movlpd	xmm8,qword ptr (ln_b_p_3-ln_b_c)[rcx]
+
+	mulsd	xmm11,xmm0				# x5
+
+	mulsd	xmm5,xmm6				# q2
+	mulsd	xmm4,xmm6				# p2
+
+	movlpd	xmm6,qword ptr real_17_bits[rip]
+	movlpd	xmm10,qword ptr (ln_b_q_4-ln_b_c)[rcx]
+
+	mulsd	xmm9,xmm7				# q3
+	mulsd	xmm8,xmm7				# p3
+	
+	andpd	xmm6,xmm1				# x_17_h
+
+	mulsd	xmm10,xmm0				# q4
+	mulsd	xmm0,qword ptr (ln_b_p_4-ln_b_c)[rcx] # p4
+
+	addsd	xmm10,xmm11				# q4+x5
+
+	addsd	xmm0,xmm8				# p3+p4
+	addsd	xmm9,xmm10				# q3+q4+x5
+
+	movsd	xmm8,xmm1				# x
+
+	addsd	xmm0,xmm4				# p2+p3+p4	
+	addsd	xmm5,xmm9				# q2+q3+q4+x5
+
+	movlpd	xmm4,qword ptr real_0_5[rip]
+	subsd	xmm8,xmm6				# x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h
+
+	addsd	xmm0,xmm2				# p1+p2+p3+p4
+	addsd	xmm3,xmm5				# q1+q2+q3+q4+x5
+
+	mulsd	xmm8,qword ptr real_0_5[rip]	# 0.5*x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h*x_17_h
+	addsd	xmm6,xmm1				# x+x_17_h
+
+	addsd	xmm0,qword ptr (ln_b_p_0-ln_b_c)[rcx] # p0+p1+p2+p3+p4
+	addsd	xmm3,qword ptr (ln_b_q_0-ln_b_c)[rcx] # q0+q1+q2+q3+q4+x5
+
+	subsd	xmm1,xmm4				# x-0.5*x_17_h*x_17_h
+	mulsd	xmm6,xmm8				# 0.5*x_17_l*(x+x_17_h)
+
+	divsd	xmm0,xmm3				# p/q
+
+	movlpd	xmm2,qword ptr ln_4_d_3_53[rip]
+	movlpd	xmm4,qword ptr ln_4_d_3_53_l[rip]
+	addsd	xmm2,xmm1
+	subsd	xmm4,xmm6
+
+	movlpd	xmm3,qword ptr ln_4_d_3_53[rip]
+	subsd	xmm3,xmm2
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+
+	addsd	xmm3,xmm1
+	addsd	xmm3,xmm4
+
+	addsd	xmm0,xmm3
+	addsd	xmm0,xmm2
+	ret
+
+ln_real_g1:
+	ucomisd	xmm2,qword ptr real_7_d_6[rip]
+
+	subsd	xmm2,qword ptr real_1_0[rip]
+
+	psubq	xmm0,xmmword ptr qw_1023[rip]
+
+	att_ja		ln_real_g_7_d_6
+
+ln_real_g1_:
+	movsd	xmm1,xmm2				# x
+	mulsd	xmm2,xmm2				# x2
+
+	movlpd	xmm4,qword ptr ln_b_q_1[rip]
+	movlpd	xmm3,qword ptr ln_b_p_1[rip]
+	mulsd	xmm4,xmm1				# q1
+	mulsd	xmm3,xmm1				# p1
+	movsd	xmm8,xmm1				# x
+
+	cvtdq2pd xmm0,xmm0
+
+	movlpd	xmm6,qword ptr ln_b_q_2[rip]
+	movlpd	xmm5,qword ptr ln_b_p_2[rip]
+	movsd	xmm7,xmm2				# x2
+	mulsd	xmm2,xmm2				# x4
+
+	movsd	xmm12,xmm1				# x
+	mulsd	xmm8,xmm7				# x3
+
+	movlpd	xmm10,qword ptr ln_b_q_3[rip]
+	movlpd	xmm9,qword ptr ln_b_p_3[rip]
+
+	mulsd	xmm12,xmm2				# x5
+
+	mulsd	xmm6,xmm7				# q2
+	mulsd	xmm5,xmm7				# p2
+
+	movlpd	xmm11,qword ptr ln_b_p_4[rip]
+
+	mulsd	xmm10,xmm8				# q3
+	mulsd	xmm9,xmm8				# p3
+	
+	mulsd	xmm11,xmm2				# p4
+	mulsd	xmm2,qword ptr ln_b_q_4[rip] # q4
+	
+	addsd	xmm2,xmm12				# q4+x5
+
+	addsd	xmm2,xmm10				# q3+q4+x5
+	addsd	xmm9,xmm11				# p3+p4
+
+	movlpd	xmm7,qword ptr real_17_bits[rip]
+
+	addsd	xmm2,xmm6				# q2+q3+q4+x5
+	addsd	xmm5,xmm9				# p2+p3+p4
+	
+	andpd	xmm7,xmm1				# x_17_h
+	movsd	xmm14,xmm1				# x
+	movlpd	xmm13,qword ptr real_0_5[rip]
+
+	movlpd	xmm6,qword ptr real_ln2_42_l[rip]
+	
+	addsd	xmm2,xmm4				# q1+q2+q3+q4+x5
+	addsd	xmm3,xmm5				# p1+p2+p3+p4
+	
+	subsd	xmm14,xmm7				# x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h
+
+	mulsd	xmm6,xmm0
+	mulsd	xmm0,qword ptr real_ln2_42[rip]
+	
+	addsd	xmm2,qword ptr ln_b_q_0[rip]	# q0+q1+q2+q3+q4+x5
+	addsd	xmm3,qword ptr ln_b_p_0[rip]	# p0+p1+p2+p3+p4
+
+	mulsd	xmm14,qword ptr real_0_5[rip] # 0.5*x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h*x_17_h
+	addsd	xmm7,xmm1				# x+x_17_h
+
+	ucomisd	xmm1,qword ptr real_2_p_m_16[rip]
+	jb		ln_real_g1_s
+
+ln_real_g1_l:
+	divsd	xmm3,xmm2				# p/q
+
+	subsd	xmm1,xmm13				# x-0.5*x_17_h*x_17_h
+	mulsd	xmm7,xmm14				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm3,xmm8				# (p/q)*x3
+	subsd	xmm3,xmm7				# (p/q)*x3-0.5x2_l
+	addsd	xmm3,xmm6
+
+
+	movsd	xmm2,xmm0
+	addsd	xmm0,xmm1
+	subsd	xmm2,xmm0
+	addsd	xmm2,xmm1
+	addsd	xmm2,xmm3
+	addsd	xmm0,xmm2
+
+
+#	addsd	xmm1,xmm3				# (p/q)*x3-0.5x2+x
+#	addsd	xmm0,xmm1
+	ret
+
+ln_real_g1_s:
+	divsd	xmm3,xmm2				# p/q
+
+	mulsd	xmm7,xmm14				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm3,xmm8				# (p/q)*x3
+	subsd	xmm3,xmm7				# (p/q)*x3-0.5x2_l
+	addsd	xmm3,xmm6
+	subsd	xmm3,xmm13				# (p/q)*x3-0.5x2
+	addsd	xmm1,xmm3				# (p/q)*x3-0.5x2_l+x
+	addsd	xmm0,xmm1
+	ret
+
+ln_real_g_7_d_6:
+	mulsd	xmm2,qword ptr real_0_75[rip]
+	movlpd	xmm1,qword ptr real_0_25[rip]
+
+	lea		rcx,ln_b_c[rip]
+	lea		rdx,ln_s_c[rip]
+
+	ucomisd	xmm2,xmm1
+	subsd	xmm2,xmm1
+
+	cmovb	rcx,rdx
+
+	movsd	xmm1,xmm2				# x
+	mulsd	xmm2,xmm2				# x2
+
+	movlpd	xmm4,qword ptr (ln_b_q_1-ln_b_c)[rcx]
+	movlpd	xmm3,qword ptr (ln_b_p_1-ln_b_c)[rcx]
+	mulsd	xmm4,xmm1				# q1
+	mulsd	xmm3,xmm1				# p1
+	movsd	xmm8,xmm1				# x
+
+	cvtdq2pd xmm0,xmm0
+
+	movlpd	xmm6,qword ptr (ln_b_q_2-ln_b_c)[rcx]
+	movlpd	xmm5,qword ptr (ln_b_p_2-ln_b_c)[rcx]
+	movsd	xmm7,xmm2				# x2
+	mulsd	xmm2,xmm2				# x4
+
+	movsd	xmm12,xmm1				# x
+	mulsd	xmm8,xmm7				# x3
+
+	movlpd	xmm10,qword ptr (ln_b_q_3-ln_b_c)[rcx]
+	movlpd	xmm9,qword ptr (ln_b_p_3-ln_b_c)[rcx]
+
+	mulsd	xmm12,xmm2				# x5
+
+	mulsd	xmm6,xmm7				# q2
+	mulsd	xmm5,xmm7				# p2
+
+	movlpd	xmm11,qword ptr (ln_b_p_4-ln_b_c)[rcx]
+
+	mulsd	xmm10,xmm8				# q3
+	mulsd	xmm9,xmm8				# p3
+	
+	mulsd	xmm11,xmm2				# p4
+	mulsd	xmm2,qword ptr (ln_b_q_4-ln_b_c)[rcx] # q4
+	
+	addsd	xmm2,xmm12				# q4+x5
+
+	addsd	xmm2,xmm10				# q3+q4+x5
+	addsd	xmm9,xmm11				# p3+p4
+
+	movlpd	xmm7,qword ptr real_17_bits[rip]
+
+	addsd	xmm2,xmm6				# q2+q3+q4+x5
+	addsd	xmm5,xmm9				# p2+p3+p4
+	
+	andpd	xmm7,xmm1				# x_17_h
+	movsd	xmm14,xmm1				# x
+	movlpd	xmm13,qword ptr real_0_5[rip]
+
+	movlpd	xmm6,qword ptr real_ln2_42_l[rip]
+	
+	addsd	xmm2,xmm4				# q1+q2+q3+q4+x5
+	addsd	xmm3,xmm5				# p1+p2+p3+p4
+	
+	subsd	xmm14,xmm7				# x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h
+
+	mulsd	xmm6,xmm0
+	mulsd	xmm0,qword ptr real_ln2_42[rip]
+	
+	addsd	xmm2,qword ptr (ln_b_q_0-ln_b_c)[rcx] # q0+q1+q2+q3+q4+x5
+	addsd	xmm3,qword ptr (ln_b_p_0-ln_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	mulsd	xmm14,qword ptr real_0_5[rip] # 0.5*x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h*x_17_h
+	addsd	xmm7,xmm1				# x+x_17_h
+
+	addsd	xmm0,qword ptr ln_4_d_3_42[rip]
+	addsd	xmm6,qword ptr ln_4_d_3_42_l[rip]
+	att_jmp		ln_real_g1_l
+
+ln_real_s:
+	movlpd	xmm3,qword ptr ln_s_q_1[rip]
+	movlpd	xmm2,qword ptr ln_s_p_1[rip]
+
+	mulsd	xmm3,xmm1				# q1
+	mulsd	xmm2,xmm1				# p1
+	movsd	xmm7,xmm1				# x
+	movlpd	xmm5,qword ptr ln_s_q_2[rip]
+	movlpd	xmm4,qword ptr ln_s_p_2[rip]
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movsd	xmm11,xmm1				# x
+	mulsd	xmm7,xmm6				# x3
+
+	movlpd	xmm9,qword ptr ln_s_q_3[rip]
+	movlpd	xmm8,qword ptr ln_s_p_3[rip]
+
+	mulsd	xmm11,xmm0				# x5
+
+	mulsd	xmm5,xmm6				# q2
+	mulsd	xmm4,xmm6				# p2
+
+	movlpd	xmm10,qword ptr ln_s_q_4[rip]
+
+	mulsd	xmm9,xmm7				# q3
+	mulsd	xmm8,xmm7				# p3
+	
+	mulsd	xmm10,xmm0				# q4
+	mulsd	xmm0,qword ptr ln_s_p_4[rip]	# p4
+
+	addsd	xmm10,xmm11				# q4+x5
+
+	addsd	xmm0,xmm8				# p3+p4
+	addsd	xmm9,xmm10				# q3+q4+x5
+
+	movlpd	xmm6,qword ptr real_17_bits[rip]
+
+	addsd	xmm0,xmm4				# p2+p3+p4	
+	addsd	xmm5,xmm9				# q2+q3+q4+x5
+
+	andpd	xmm6,xmm1				# x_17_h
+	movsd	xmm8,xmm1				# x
+	movlpd	xmm4,qword ptr real_0_5[rip]
+
+	addsd	xmm0,xmm2				# p1+p2+p3+p4
+	addsd	xmm3,xmm5				# q1+q2+q3+q4+x5
+
+	subsd	xmm8,xmm6				# x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h
+
+	addsd	xmm0,qword ptr ln_s_p_0[rip]	# p0+p1+p2+p3+p4
+	addsd	xmm3,qword ptr ln_s_q_0[rip]	# q0+q1+q2+q3+q4+x5
+
+	mulsd	xmm8,qword ptr real_0_5[rip] # 0.5*x_17_l
+	mulsd	xmm4,xmm6				# 0.5*x_17_h*x_17_h
+	addsd	xmm6,xmm1				# x+x_17_h
+
+	ucomisd	xmm1,qword ptr real_m_2_p_m_16[rip]
+	ja		ln_real_s_s
+
+	divsd	xmm0,xmm3				# p/q
+
+	subsd	xmm1,xmm4				# x-0.5*x_17_h*x_17_h
+	mulsd	xmm6,xmm8				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+	subsd	xmm0,xmm6				# (p/q)*x3-0.5x2_l
+	addsd	xmm0,xmm1				# (p/q)*x3-0.5x2+x
+	ret
+
+ln_real_s_s:
+	divsd	xmm0,xmm3				# p/q
+
+	mulsd	xmm6,xmm8				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+	subsd	xmm0,xmm6				# (p/q)*x3-0.5x2_l
+	subsd	xmm0,xmm4				# (p/q)*x3-0.5x2
+	addsd	xmm0,xmm1				# (p/q)*x3-0.5x2+x
+	ret
+	
+
+ln_real_l1:
+	subsd	xmm3,qword ptr real_1_0[rip]
+
+	psubq	xmm0,xmmword ptr qw_1022[rip]
+ln_real_l1_:
+	movsd	xmm1,xmm3				# x
+	mulsd	xmm3,xmm3				# x2
+
+	movlpd	xmm4,qword ptr ln_s_q_1[rip]
+	movlpd	xmm2,qword ptr ln_s_p_1[rip]
+	mulsd	xmm4,xmm1				# q1
+	mulsd	xmm2,xmm1				# p1
+	movsd	xmm8,xmm1				# x
+
+	cvtdq2pd xmm0,xmm0
+
+	movlpd	xmm6,qword ptr ln_s_q_2[rip]
+	movlpd	xmm5,qword ptr ln_s_p_2[rip]
+	movsd	xmm7,xmm3				# x2
+	mulsd	xmm3,xmm3				# x4
+
+	movsd	xmm12,xmm1				# x
+	mulsd	xmm8,xmm7				# x3
+
+	movlpd	xmm10,qword ptr ln_s_q_3[rip]
+	movlpd	xmm9,qword ptr ln_s_p_3[rip]
+
+	mulsd	xmm12,xmm3				# x5
+
+	mulsd	xmm6,xmm7				# q2
+	mulsd	xmm5,xmm7				# p2
+
+	movlpd	xmm11,qword ptr ln_s_p_4[rip]
+
+	mulsd	xmm10,xmm8				# q3
+	mulsd	xmm9,xmm8				# p3
+	
+	mulsd	xmm11,xmm3				# p4
+	mulsd	xmm3,qword ptr ln_s_q_4[rip] # q4
+
+	addsd	xmm3,xmm12				# q4+x5
+
+	addsd	xmm3,xmm10				# q3+q4+x5
+	addsd	xmm9,xmm11				# p3+p4
+
+	movlpd	xmm7,qword ptr real_17_bits[rip]
+
+	addsd	xmm3,xmm6				# q2+q3+q4+x5
+	addsd	xmm5,xmm9				# p2+p3+p4	
+
+	andpd	xmm7,xmm1				# x_17_h
+	movsd	xmm14,xmm1				# x
+	movlpd	xmm13,qword ptr real_0_5[rip]
+
+	movlpd	xmm6,qword ptr real_ln2_42_l[rip]
+
+	addsd	xmm3,xmm4				# q1+q2+q3+q4+x5
+	addsd	xmm2,xmm5				# p1+p2+p3+p4
+
+	subsd	xmm14,xmm7				# x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h
+
+	mulsd	xmm6,xmm0
+	mulsd	xmm0,qword ptr real_ln2_42[rip]
+
+	addsd	xmm3,qword ptr ln_s_q_0[rip]	# q0+q1+q2+q3+q4+x5
+	addsd	xmm2,qword ptr ln_s_p_0[rip]	# p0+p1+p2+p3+p4
+
+	mulsd	xmm14,qword ptr real_0_5[rip] # 0.5*x_17_l
+	mulsd	xmm13,xmm7				# 0.5*x_17_h*x_17_h
+	addsd	xmm7,xmm1				# x+x_17_h
+
+	ucomisd	xmm1,qword ptr real_m_2_p_m_16[rip]
+	ja		ln_real_l1_s
+
+	divsd	xmm2,xmm3				# p/q
+
+	subsd	xmm1,xmm13				# x-0.5*x_17_h*x_17_h
+	mulsd	xmm7,xmm14				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm2,xmm8				# (p/q)*x3
+	subsd	xmm2,xmm7				# (p/q)*x3-0.5x2_l
+	addsd	xmm2,xmm6
+
+
+	movsd	xmm3,xmm0
+	addsd	xmm0,xmm1
+	subsd	xmm3,xmm0
+	addsd	xmm3,xmm1
+	addsd	xmm3,xmm2
+	addsd	xmm0,xmm3
+
+#	addsd	xmm1,xmm2				# (p/q)*x3-0.5x2+x
+#	addsd	xmm0,xmm1
+	ret
+
+ln_real_l1_s:
+	divsd	xmm2,xmm3				# p/q
+
+	mulsd	xmm7,xmm14				# 0.5*x_17_l*(x+x_17_h)
+
+	mulsd	xmm2,xmm8				# (p/q)*x3
+	subsd	xmm2,xmm7				# (p/q)*x3-0.5x2_l
+	addsd	xmm2,xmm6
+	subsd	xmm2,xmm13				# (p/q)*x3-0.5x2
+	addsd	xmm1,xmm2				# (p/q)*x3-0.5x2+x
+	addsd	xmm0,xmm1
+	ret
+
+ln_real_less_or_equal_0_or_nan:
+	jb		ln_real_less_0_or_nan	# x<0.0 | nan
+
+	movsd	xmm0,qword ptr real_m_1_0[rip]
+	divsd	xmm0,qword ptr real_0_0[rip]	# yield -inf
+	ret
+
+ln_real_less_0_or_nan:
+	att_jp		ln_real_inf_or_nan		# nan
+
+	movsd	xmm0,qword ptr real_0_0[rip]	# yield nan
+	divsd	xmm0,xmm0
+
+ln_real_inf_or_nan:
+	ret
+
+	.globl	log10_real
+
+log10_real:
+#	ucomisd	xmm0,qword ptr real_0_7025
+	ucomisd	xmm0,qword ptr real_0_833[rip]
+#	jb		log10_real_small			# x<0.7025 | nan
+	jb		log10_real_small			# x<0.833 | nan
+	ucomisd	xmm0,qword ptr real_1_666[rip]
+	ja		log10_real_large			# x>1.666
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+
+	ucomisd	xmm0,qword ptr real_7_d_6[rip]
+	ja		log10_real_a_7_d_6
+
+	ucomisd	xmm0,xmm2
+
+	subsd	xmm0,xmm2
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	jae		log10_real_0
+	jmp		log10_real_s
+
+log10_real_small:
+	ucomisd	xmm0,qword ptr real_2_p_m_1022[rip]
+	jae		log10_real_
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	jbe		log10_real_less_or_equal_0_or_nan
+
+# denormalized number
+
+	mulsd	xmm0,qword ptr real_2_p_55[rip]
+
+	movlpd	xmm1,qword ptr mask_significand[rip]
+	andpd	xmm1,xmm0
+	psrlq	xmm0,52
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	movlpd	xmm3,qword ptr real_0_5[rip]
+	orpd	xmm2,xmm1
+	orpd	xmm3,xmm1
+
+#	ucomisd	xmm2,qword ptr real_1_405
+	ucomisd	xmm2,qword ptr real_1_666[rip]
+	jbe		log10_real_dn_g1
+
+	subsd	xmm3,qword ptr real_1_0[rip]
+	psubq	xmm0,xmmword ptr qw_1077[rip]
+	jmp		log10_real_l1_
+
+log10_real_dn_g1:
+	ucomisd	xmm2,qword ptr real_7_d_6[rip]
+
+	subsd	xmm2,qword ptr real_1_0[rip]
+	psubq	xmm0,xmmword ptr qw_1078[rip]
+
+	ja		log10_real_g_7_d_6
+	jmp		log10_real_g1_
+
+log10_real_large:
+	ucomisd	xmm0,qword ptr real_max[rip]
+	ja		log10_real_inf_or_nan
+
+log10_real_:
+	movlpd	xmm1,qword ptr mask_significand[rip]
+	andpd	xmm1,xmm0
+	psrlq	xmm0,52
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	movlpd	xmm3,qword ptr real_0_5[rip]
+	orpd	xmm2,xmm1
+	orpd	xmm3,xmm1
+
+#	ucomisd	xmm2,qword ptr real_1_405
+	ucomisd	xmm2,qword ptr real_1_666[rip]
+	jbe		log10_real_g1
+	jmp		log10_real_l1
+
+log10_real_0:
+	movlpd	xmm2,qword ptr log10_b_p_1[rip]
+	movlpd	xmm3,qword ptr log10_b_q_1[rip]
+	mulsd	xmm2,xmm1				# p1
+	mulsd	xmm3,xmm1				# q1
+	movsd	xmm7,xmm1				# x
+
+	movlpd	xmm4,qword ptr log10_b_p_2[rip]
+	movlpd	xmm5,qword ptr log10_b_q_2[rip]
+	mulsd	xmm7,xmm0				# x3
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm8,qword ptr log10_b_p_3[rip]
+	movlpd	xmm9,qword ptr log10_b_q_3[rip]
+
+	mulsd	xmm4,xmm6				# p2
+	mulsd	xmm5,xmm6				# q2
+
+	movlpd	xmm10,qword ptr log10_b_p_4[rip]
+
+	mulsd	xmm8,xmm7				# p3
+	mulsd	xmm9,xmm7				# q3
+
+	movsd	xmm11,xmm7				# x3
+	mulsd	xmm7,xmm6				# x5
+	
+	mulsd	xmm10,xmm0				# p4
+	mulsd	xmm0,qword ptr log10_b_q_4[rip] # q4
+
+	addsd	xmm0,xmm7				# q4+x5
+
+	addsd	xmm9,xmm0				# q3+q4+x5
+	addsd	xmm8,xmm10				# p3+p4
+
+	mulsd	xmm6,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm9,xmm5				# q2+q3+q4+x5
+	addsd	xmm4,xmm8				# p2+p3+p4	
+
+	addsd	xmm9,xmm3				# q1+q2+q3+q4+x5
+	addsd	xmm2,xmm4				# p1+p2+p3+p4
+
+	movlpd	xmm4,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm1				# x
+
+	addsd	xmm9,qword ptr log10_b_q_0[rip] # q0+q1+q2+q3+q4+x5
+	addsd	xmm2,qword ptr log10_b_p_0[rip] # p0+p1+p2+p3+p4
+
+	subsd	xmm1,xmm6				# x-0.5x2
+
+	divsd	xmm2,xmm9				# p/q
+
+	andpd	xmm4,xmm1				# (x-0.5x2)_h
+	subsd	xmm0,xmm4				# x-(x-0.5x2)_h
+	subsd	xmm0,xmm6				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm11				# (p/q)*x3
+
+	mulsd	xmm1,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm0,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+	mulsd	xmm4,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm0,xmm1				# (x-0.5x2) * (1/ln 10)_l+(x-0.5x2)_l * (1/ln 10)_h
+	addsd	xmm0,xmm2
+	addsd	xmm0,xmm4				# (x-0.5x2)*(1/ln 10) + (p/q)*x3
+
+	ret
+
+#log10_real_0_:
+#	movlpd	xmm2,qword ptr log10_b_p_1
+#	movlpd	xmm10,qword ptr log10_b_q_1
+#	mulsd	xmm2,xmm1				# p1
+#	mulsd	xmm10,xmm1				# q1
+#	movsd	xmm7,xmm1				# x
+#
+#	movlpd	xmm11,qword ptr log10_b_p_2
+#	movlpd	xmm5,qword ptr log10_b_q_2
+#	mulsd	xmm7,xmm0				# x3
+#	movsd	xmm6,xmm0				# x2
+#	mulsd	xmm0,xmm0				# x4
+#
+#	movlpd	xmm8,qword ptr log10_b_p_3
+#	movlpd	xmm9,qword ptr log10_b_q_3
+#
+##	mulsd	xmm11,xmm6				# p2
+#	mulsd	xmm5,xmm6				# q2
+#
+#	movlpd	xmm3,qword ptr log10_b_p_4
+#
+#	mulsd	xmm8,xmm7				# p3
+#	mulsd	xmm9,xmm7				# q3
+#
+#	mulsd	xmm6,xmm7				# x5
+#
+#	mulsd	xmm3,xmm0				# p4
+#	mulsd	xmm0,qword ptr log10_b_q_4 # q4
+#	
+#	movsd	xmm4,xmm1				# x
+#	addsd	xmm0,xmm6				# q4+x5
+#	movlpd	xmm6,qword ptr real_m_0_5
+#
+#	addsd	xmm9,xmm0				# q3+q4+x5
+#	addsd	xmm8,xmm3				# p3+p4
+#
+#	movlpd	xmm0,qword ptr real_14_bits
+#
+#	andpd	xmm0,xmm1				# x_h
+#	mulsd	xmm6,xmm1 				# -0.5x
+#
+#	addsd	xmm9,xmm5				# q2+q3+q4+x5
+#	addsd	xmm11,xmm8				# p2+p3+p4	
+#
+#	movsd	xmm3,xmm1				# x
+#	subsd	xmm4,xmm0				# x_l
+#	addsd	xmm6,qword ptr real_1_0	# 1-0.5x
+#	movlpd	xmm5,qword ptr real_13_bits
+#
+#	addsd	xmm9,xmm10				# q1+q2+q3+q4+x5
+#	addsd	xmm2,xmm11				# p1+p2+p3+p4
+#	
+#	andpd	xmm5,xmm6				# (1-0.5x)_h
+#	mulsd	xmm3,xmm6				# (1-0.5x)*x = x-0.5x2
+#
+#	addsd	xmm9,qword ptr log10_b_q_0 # q0+q1+q2+q3+q4+x5
+#	addsd	xmm2,qword ptr log10_b_p_0 # p0+p1+p2+p3+p4
+#
+#	subsd	xmm6,xmm5				# (1-0.5x)_l
+#	mulsd	xmm0,xmm5				# (1-0.5x)_h*x_h = (x-0.5x2)_h
+#	mulsd	xmm4,xmm5				# (1-0.5x)_h*x_l
+#
+#	divsd	xmm2,xmm9				# p/q
+#
+#	mulsd	xmm3,qword ptr d1ln10_26_l # (x-0.5x2) * (1/ln 10)_l
+#	mulsd	xmm6,xmm1				# (1-0.5x)_l*x
+#	mulsd	xmm0,qword ptr d1ln10_26 # (x-0.5x2)_h * (1/ln 10)_h
+#
+#	addsd	xmm6,xmm4				# (x-0.5x2)_l
+#
+#	mulsd	xmm6,qword ptr d1ln10_26 # (x-0.5x2)_l * (1/ln 10)_h
+#
+#	mulsd	xmm2,xmm7				# (p/q)*x3
+#
+#	addsd	xmm6,xmm3
+#	
+#	addsd	xmm2,xmm6
+#	
+#	addsd	xmm0,xmm2
+#
+#	ret
+
+log10_real_a_7_d_6:
+	subsd	xmm0,xmm2
+
+	mulsd	xmm0,qword ptr real_0_75[rip]
+	movlpd	xmm1,qword ptr real_0_25[rip]
+
+	lea		rcx,log10_b_c[rip]
+	lea		rdx,log10_s_c[rip]
+
+	ucomisd	xmm0,xmm1
+	subsd	xmm0,xmm1
+
+	cmovb	rcx,rdx
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	movlpd	xmm2,qword ptr (log10_b_p_1-log10_b_c)[rcx]
+	movlpd	xmm3,qword ptr (log10_b_q_1-log10_b_c)[rcx]
+	mulsd	xmm2,xmm1				# p1
+	mulsd	xmm3,xmm1				# q1
+	movsd	xmm7,xmm1				# x
+
+	movlpd	xmm4,qword ptr (log10_b_p_2-log10_b_c)[rcx]
+	movlpd	xmm5,qword ptr (log10_b_q_2-log10_b_c)[rcx]
+	mulsd	xmm7,xmm0				# x3
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm8,qword ptr (log10_b_p_3-log10_b_c)[rcx]
+	movlpd	xmm9,qword ptr (log10_b_q_3-log10_b_c)[rcx]
+
+	mulsd	xmm4,xmm6				# p2
+	mulsd	xmm5,xmm6				# q2
+
+	movlpd	xmm10,qword ptr (log10_b_p_4-log10_b_c)[rcx]
+
+	mulsd	xmm8,xmm7				# p3
+	mulsd	xmm9,xmm7				# q3
+
+	movsd	xmm11,xmm7				# x3
+	mulsd	xmm7,xmm6				# x5
+	
+	mulsd	xmm10,xmm0				# p4
+	mulsd	xmm0,qword ptr (log10_b_q_4-log10_b_c)[rcx] # q4
+
+	addsd	xmm0,xmm7				# q4+x5
+
+	addsd	xmm9,xmm0				# q3+q4+x5
+	addsd	xmm8,xmm10				# p3+p4
+
+	mulsd	xmm6,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm9,xmm5				# q2+q3+q4+x5
+	addsd	xmm4,xmm8				# p2+p3+p4	
+
+	addsd	xmm9,xmm3				# q1+q2+q3+q4+x5
+	addsd	xmm2,xmm4				# p1+p2+p3+p4
+
+	movlpd	xmm4,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm1				# x
+
+	addsd	xmm9,qword ptr (log10_b_q_0-log10_b_c)[rcx] # q0+q1+q2+q3+q4+x5
+	addsd	xmm2,qword ptr (log10_b_p_0-log10_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	subsd	xmm1,xmm6				# x-0.5x2
+
+	divsd	xmm2,xmm9				# p/q
+
+	andpd	xmm4,xmm1				# (x-0.5x2)_h
+	subsd	xmm0,xmm4				# x-(x-0.5x2)_h
+	subsd	xmm0,xmm6				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm11				# (p/q)*x3
+
+	mulsd	xmm1,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm0,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+	mulsd	xmm4,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm0,xmm1				# (x-0.5x2) * (1/ln 10)_l+(x-0.5x2)_l * (1/ln 10)_h
+
+	movlpd	xmm3,qword ptr log10_4_d_3_53[rip]
+	movlpd	xmm5,qword ptr log10_4_d_3_53[rip]
+
+	addsd	xmm3,xmm4
+	addsd	xmm0,qword ptr log10_4_d_3_53_l[rip]
+
+	subsd	xmm5,xmm3
+
+	addsd	xmm5,xmm4
+	
+	addsd	xmm0,xmm5
+
+	addsd	xmm0,xmm2
+	addsd	xmm0,xmm3				# (x-0.5x2)*(1/ln 10) + (p/q)*x3
+	ret
+
+log10_real_g1:
+	ucomisd	xmm2,qword ptr real_7_d_6[rip]
+
+	subsd	xmm2,qword ptr real_1_0[rip]
+
+	psubq	xmm0,xmmword ptr qw_1023[rip]
+
+	att_ja		log10_real_g_7_d_6
+
+log10_real_g1_:
+	movsd	xmm1,xmm2				# x
+	mulsd	xmm2,xmm2				# x2
+
+	movlpd	xmm11,qword ptr log10_b_p_1[rip]
+	movlpd	xmm10,qword ptr log10_b_q_1[rip]
+	mulsd	xmm11,xmm1				# p1
+	mulsd	xmm10,xmm1				# q1
+	movsd	xmm7,xmm1				# x
+
+	cvtdq2pd xmm12,xmm0
+
+	movlpd	xmm0,qword ptr log10_b_p_2[rip]
+	movlpd	xmm5,qword ptr log10_b_q_2[rip]
+	mulsd	xmm7,xmm2				# x3
+	movsd	xmm6,xmm2				# x2
+	mulsd	xmm2,xmm2				# x4
+
+	movlpd	xmm8,qword ptr log10_b_p_3[rip]
+	movlpd	xmm9,qword ptr log10_b_q_3[rip]
+
+	mulsd	xmm0,xmm6				# p2
+	mulsd	xmm5,xmm6				# q2
+
+	movlpd	xmm3,qword ptr log10_b_p_4[rip]
+
+	mulsd	xmm8,xmm7				# p3
+	mulsd	xmm9,xmm7				# q3
+
+	mulsd	xmm6,xmm7				# x5
+
+	mulsd	xmm3,xmm2				# p4
+	mulsd	xmm2,qword ptr log10_b_q_4[rip] # q4
+	
+	movsd	xmm4,xmm1				# x
+	addsd	xmm2,xmm6				# q4+x5	
+	movlpd	xmm6,qword ptr real_m_0_5[rip]
+
+	addsd	xmm9,xmm2				# q3+q4+x5
+	addsd	xmm8,xmm3				# p3+p4
+
+	movlpd	xmm13,qword ptr real_log2_10_42_l[rip]
+	movlpd	xmm2,qword ptr real_14_bits[rip]
+
+	andpd	xmm2,xmm1				# x_h
+	mulsd	xmm6,xmm1 				# -0.5x
+
+	addsd	xmm9,xmm5				# q2+q3+q4+x5
+	addsd	xmm0,xmm8				# p2+p3+p4	
+
+	mulsd	xmm13,xmm12
+	mulsd	xmm12,qword ptr real_log2_10_42[rip]
+
+	movsd	xmm3,xmm1				# x
+	subsd	xmm4,xmm2				# x_l
+	addsd	xmm6,qword ptr real_1_0[rip]	# 1-0.5x
+	movlpd	xmm5,qword ptr real_13_bits[rip]
+
+	addsd	xmm9,xmm10				# q1+q2+q3+q4+x5
+	addsd	xmm0,xmm11				# p1+p2+p3+p4
+	
+	andpd	xmm5,xmm6				# (1-0.5x)_h
+	mulsd	xmm3,xmm6				# (1-0.5x)*x = x-0.5x2
+
+	addsd	xmm9,qword ptr log10_b_q_0[rip] # q0+q1+q2+q3+q4+x5
+	addsd	xmm0,qword ptr log10_b_p_0[rip] # p0+p1+p2+p3+p4
+
+	subsd	xmm6,xmm5				# (1-0.5x)_l
+	mulsd	xmm2,xmm5				# (1-0.5x)_h*x_h = (x-0.5x2)_h
+	mulsd	xmm4,xmm5				# (1-0.5x)_h*x_l
+
+	divsd	xmm0,xmm9				# p/q
+
+	mulsd	xmm3,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm6,xmm1				# (1-0.5x)_l*x
+	mulsd	xmm2,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm6,xmm4				# (x-0.5x2)_l
+
+	mulsd	xmm6,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+
+	mulsd	xmm0,xmm7				# (p/q)*x3
+
+	addsd	xmm6,xmm3
+	
+	addsd	xmm0,xmm6
+	
+	addsd	xmm0,xmm13
+
+	movsd	xmm3,xmm2
+	addsd	xmm2,xmm12
+	subsd	xmm12,xmm2
+	addsd	xmm12,xmm3
+	addsd	xmm0,xmm12
+	addsd	xmm0,xmm2
+
+#	addsd	xmm0,xmm2
+#	addsd	xmm0,xmm12
+
+	ret
+
+log10_real_g_7_d_6:
+	mulsd	xmm2,qword ptr real_0_75[rip]
+	movlpd	xmm1,qword ptr real_0_25[rip]
+
+	lea		rcx,log10_b_c[rip]
+	lea		rdx,log10_s_c[rip]
+
+	ucomisd	xmm2,xmm1
+	subsd	xmm2,xmm1
+
+	cmovb	rcx,rdx
+
+	movsd	xmm1,xmm2				# x
+	mulsd	xmm2,xmm2				# x2
+
+	movlpd	xmm3,qword ptr (log10_b_p_1-log10_b_c)[rcx]
+	movlpd	xmm4,qword ptr (log10_b_q_1-log10_b_c)[rcx]
+	mulsd	xmm3,xmm1				# p1
+	mulsd	xmm4,xmm1				# q1
+	movsd	xmm8,xmm1				# x
+
+	cvtdq2pd xmm0,xmm0
+
+	movlpd	xmm5,qword ptr (log10_b_p_2-log10_b_c)[rcx]
+	movlpd	xmm6,qword ptr (log10_b_q_2-log10_b_c)[rcx]
+	mulsd	xmm8,xmm2				# x3
+	movsd	xmm7,xmm2				# x2
+	mulsd	xmm2,xmm2				# x4
+
+	movlpd	xmm9,qword ptr (log10_b_p_3-log10_b_c)[rcx]
+	movlpd	xmm10,qword ptr (log10_b_q_3-log10_b_c)[rcx]
+
+	mulsd	xmm5,xmm7				# p2
+	mulsd	xmm6,xmm7				# q2
+
+	movlpd	xmm11,qword ptr (log10_b_p_4-log10_b_c)[rcx]
+
+	mulsd	xmm9,xmm8				# p3
+	mulsd	xmm10,xmm8				# q3
+
+	movsd	xmm12,xmm8				# x3
+	mulsd	xmm8,xmm7				# x5
+
+	mulsd	xmm11,xmm2				# p4
+	mulsd	xmm2,qword ptr (log10_b_q_4-log10_b_c)[rcx] # q4
+	
+	addsd	xmm2,xmm8				# q4+x5
+	
+	addsd	xmm10,xmm2				# q3+q4+x5
+	addsd	xmm9,xmm11				# p3+p4
+
+	mulsd	xmm7,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm10,xmm6				# q2+q3+q4+x5
+	addsd	xmm5,xmm9				# p2+p3+p4
+
+	movlpd	xmm6,qword ptr real_log2_10_42_l[rip]
+
+	addsd	xmm10,xmm4				# q1+q2+q3+q4+x5
+	addsd	xmm3,xmm5				# p1+p2+p3+p4
+
+	movlpd	xmm5,qword ptr real_26_bits[rip]
+	movsd	xmm2,xmm1				# x
+
+	mulsd	xmm6,xmm0
+	mulsd	xmm0,qword ptr real_log2_10_42[rip]
+
+	addsd	xmm10,qword ptr (log10_b_q_0-log10_b_c)[rcx] # q0+q1+q2+q3+q4+x5
+	addsd	xmm3,qword ptr (log10_b_p_0-log10_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	subsd	xmm1,xmm7				# x-0.5x2
+
+	addsd	xmm6,qword ptr log10_4_d_3_42_l[rip]
+	addsd	xmm0,qword ptr log10_4_d_3_42[rip]
+
+	divsd	xmm3,xmm10				# p/q
+
+	andpd	xmm5,xmm1				# (x-0.5x2)_h
+	subsd	xmm2,xmm5				# x-(x-0.5x2)_h
+	subsd	xmm2,xmm7				# (x-0.5x2)_l
+
+	mulsd	xmm3,xmm12				# (p/q)*x3
+
+	mulsd	xmm1,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm2,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+	mulsd	xmm5,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm2,xmm1				# (x-0.5x2) * (1/ln 10)_l+(x-0.5x2)_l * (1/ln 10)_h
+	addsd	xmm2,xmm3
+	addsd	xmm2,xmm6
+
+	movsd	xmm3,xmm0
+	addsd	xmm0,xmm5
+	subsd	xmm3,xmm0
+	addsd	xmm3,xmm5
+	addsd	xmm3,xmm2
+	addsd	xmm0,xmm3
+
+#	addsd	xmm2,xmm5				# (x-0.5x2)*(1/ln 10) + (p/q)*x3
+#	addsd	xmm0,xmm2
+	ret
+
+log10_real_s:
+	movlpd	xmm2,qword ptr log10_s_p_1[rip]
+	movlpd	xmm3,qword ptr log10_s_q_1[rip]
+	mulsd	xmm2,xmm1				# p1
+	mulsd	xmm3,xmm1				# q1
+	movsd	xmm7,xmm1				# x
+
+	movlpd	xmm4,qword ptr log10_s_p_2[rip]
+	movlpd	xmm5,qword ptr log10_s_q_2[rip]
+	mulsd	xmm7,xmm0				# x3
+	movsd	xmm6,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	movlpd	xmm8,qword ptr log10_s_p_3[rip]
+	movlpd	xmm9,qword ptr log10_s_q_3[rip]
+
+	mulsd	xmm4,xmm6				# p2
+	mulsd	xmm5,xmm6				# q2
+
+	movlpd	xmm10,qword ptr log10_s_p_4[rip]
+
+	mulsd	xmm8,xmm7				# p3
+	mulsd	xmm9,xmm7				# q3
+
+	movsd	xmm11,xmm7				# x3
+	mulsd	xmm7,xmm6				# x5
+	
+	mulsd	xmm10,xmm0				# p4
+	mulsd	xmm0,qword ptr log10_s_q_4[rip] # q4
+
+	addsd	xmm0,xmm7				# q4+x5
+
+	addsd	xmm9,xmm0				# q3+q4+x5
+	addsd	xmm8,xmm10				# p3+p4
+
+	mulsd	xmm6,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm9,xmm5				# q2+q3+q4+x5
+	addsd	xmm4,xmm8				# p2+p3+p4	
+
+	addsd	xmm9,xmm3				# q1+q2+q3+q4+x5
+	addsd	xmm2,xmm4				# p1+p2+p3+p4
+
+	movlpd	xmm4,qword ptr real_26_bits[rip]
+	movsd	xmm0,xmm1				# x
+
+	addsd	xmm9,qword ptr log10_s_q_0[rip] # q0+q1+q2+q3+q4+x5
+	addsd	xmm2,qword ptr log10_s_p_0[rip] # p0+p1+p2+p3+p4
+
+	subsd	xmm1,xmm6				# x-0.5x2
+
+	divsd	xmm2,xmm9				# p/q
+
+	andpd	xmm4,xmm1				# (x-0.5x2)_h
+	subsd	xmm0,xmm4				# x-(x-0.5x2)_h
+	subsd	xmm0,xmm6				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm11				# (p/q)*x3
+
+	mulsd	xmm1,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm0,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+	mulsd	xmm4,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm0,xmm1				# (x-0.5x2) * (1/ln 10)_l+(x-0.5x2)_l * (1/ln 10)_h
+	addsd	xmm0,xmm2
+	addsd	xmm0,xmm4				# (x-0.5x2)*(1/ln 10) + (p/q)*x3
+
+	ret
+
+log10_real_l1:
+	subsd	xmm3,qword ptr real_1_0[rip]
+
+	psubq	xmm0,xmmword ptr qw_1022[rip]
+log10_real_l1_:
+	movsd	xmm1,xmm3				# x
+	mulsd	xmm3,xmm3				# x2
+
+	movlpd	xmm2,qword ptr log10_s_p_1[rip]
+	movlpd	xmm4,qword ptr log10_s_q_1[rip]
+	mulsd	xmm2,xmm1				# p1
+	mulsd	xmm4,xmm1				# q1
+	movsd	xmm8,xmm1				# x
+
+	cvtdq2pd xmm0,xmm0
+
+	movlpd	xmm5,qword ptr log10_s_p_2[rip]
+	movlpd	xmm6,qword ptr log10_s_q_2[rip]
+	mulsd	xmm8,xmm3				# x3
+	movsd	xmm7,xmm3				# x2
+	mulsd	xmm3,xmm3				# x4
+
+	movlpd	xmm9,qword ptr log10_s_p_3[rip]
+	movlpd	xmm10,qword ptr log10_s_q_3[rip]
+
+	mulsd	xmm5,xmm7				# p2
+	mulsd	xmm6,xmm7				# q2
+
+	movlpd	xmm11,qword ptr log10_s_p_4[rip]
+
+	mulsd	xmm9,xmm8				# p3
+	mulsd	xmm10,xmm8				# q3
+
+	movsd	xmm12,xmm8				# x3
+	mulsd	xmm8,xmm7				# x5
+
+	mulsd	xmm11,xmm3				# p4
+	mulsd	xmm3,qword ptr log10_s_q_4[rip] # q4
+	
+	addsd	xmm3,xmm8				# q4+x5
+	
+	addsd	xmm10,xmm3				# q3+q4+x5
+	addsd	xmm9,xmm11				# p3+p4
+
+	mulsd	xmm7,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm10,xmm6				# q2+q3+q4+x5
+	addsd	xmm5,xmm9				# p2+p3+p4
+
+	movlpd	xmm6,qword ptr real_log2_10_42_l[rip]
+
+	addsd	xmm10,xmm4				# q1+q2+q3+q4+x5
+	addsd	xmm2,xmm5				# p1+p2+p3+p4
+
+	movlpd	xmm5,qword ptr real_26_bits[rip]
+	movsd	xmm3,xmm1				# x
+
+	mulsd	xmm6,xmm0
+	mulsd	xmm0,qword ptr real_log2_10_42[rip]
+
+	addsd	xmm10,qword ptr log10_s_q_0[rip] # q0+q1+q2+q3+q4+x5
+	addsd	xmm2,qword ptr log10_s_p_0[rip] # p0+p1+p2+p3+p4
+
+	subsd	xmm1,xmm7				# x-0.5x2
+
+	divsd	xmm2,xmm10				# p/q
+
+	andpd	xmm5,xmm1				# (x-0.5x2)_h
+	subsd	xmm3,xmm5				# x-(x-0.5x2)_h
+	subsd	xmm3,xmm7				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm12				# (p/q)*x3
+
+	mulsd	xmm1,qword ptr d1ln10_26_l[rip] # (x-0.5x2) * (1/ln 10)_l
+	mulsd	xmm3,qword ptr d1ln10_26[rip] # (x-0.5x2)_l * (1/ln 10)_h
+	mulsd	xmm5,qword ptr d1ln10_26[rip] # (x-0.5x2)_h * (1/ln 10)_h
+
+	addsd	xmm3,xmm1				# (x-0.5x2) * (1/ln 10)_l+(x-0.5x2)_l * (1/ln 10)_h
+	addsd	xmm3,xmm2
+	addsd	xmm3,xmm6
+
+	movsd	xmm2,xmm0
+	addsd	xmm0,xmm5
+	subsd	xmm2,xmm0
+	addsd	xmm2,xmm5
+	addsd	xmm2,xmm3
+	addsd	xmm0,xmm2
+
+#	addsd	xmm3,xmm5				# (x-0.5x2)*(1/ln 10) + (p/q)*x3
+#	addsd	xmm0,xmm3
+	ret
+
+log10_real_less_or_equal_0_or_nan:
+	jb		log10_real_less_0_or_nan # x<0.0 | nan
+
+	movsd	xmm0,qword ptr real_m_1_0[rip]
+	divsd	xmm0,qword ptr real_0_0[rip]	# yield -inf
+	ret
+
+log10_real_less_0_or_nan:
+	att_jp		log10_real_inf_or_nan	# nan
+
+	movsd	xmm0,qword ptr real_0_0[rip]	# yield nan
+	divsd	xmm0,xmm0
+
+log10_real_inf_or_nan:
+	ret
+
+
+	.globl	pow_real
+
+pow_real:
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	je		pow_real_to_zero_or_nan	# y==0.0 | nan
+
+	ucomisd	xmm1,qword ptr real_0_83[rip]
+	jb		pow_real_small			# x<0.83 | nan
+pow_real_not_small:
+	ucomisd	xmm1,qword ptr real_1_66[rip]
+	ja		pow_real_large
+
+	ucomisd	xmm1,qword ptr real_7_d_6[rip]
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+
+	ja		pow_real_a_7_d_6
+
+pow_real_between_0_83_and_7_d_6:
+	ucomisd	xmm0,qword ptr real_power_exp_too_large[rip]
+	ja		pow_real_exp_too_large
+
+	ucomisd	xmm0,qword ptr real_power_exp_too_small[rip]
+	jb		pow_real_exp_too_large_m
+
+	ucomisd	xmm1,xmm2
+	
+	subsd	xmm1,xmm2
+
+	lea		rcx,log2_b_c[rip]
+	lea		rdx,log2_s_c[rip]
+	cmovb	rcx,rdx
+
+	movsd	xmm5,xmm1				# x
+	mulsd	xmm1,xmm1				# x2
+
+	movlpd	xmm2,qword ptr (log2_b_p_1-log2_b_c)[rcx]
+	movlpd	xmm4,qword ptr (log2_b_q_1-log2_b_c)[rcx]
+	mulsd	xmm2,xmm5				# p1
+	mulsd	xmm4,xmm5				# q1
+	movsd	xmm8,xmm5				# x
+	movlpd	xmm3,qword ptr (log2_b_p_2-log2_b_c)[rcx]
+	movlpd	xmm6,qword ptr (log2_b_q_2-log2_b_c)[rcx]
+	mulsd	xmm8,xmm1				# x3
+	movsd	xmm7,xmm1				# x2
+	mulsd	xmm1,xmm1				# x4
+
+	movlpd	xmm9,qword ptr (log2_b_p_3-log2_b_c)[rcx]
+	movlpd	xmm10,qword ptr (log2_b_q_3-log2_b_c)[rcx]
+
+	mulsd	xmm3,xmm7				# p2
+	mulsd	xmm6,xmm7				# q2
+
+	movlpd	xmm11,qword ptr (log2_b_p_4-log2_b_c)[rcx]
+
+	mulsd	xmm9,xmm8				# p3
+	mulsd	xmm10,xmm8				# q3
+	
+	mulsd	xmm11,xmm1				# p4
+	
+	addsd	xmm10,xmm1				# q3+x4
+	addsd	xmm9,xmm11				# p3+p4
+
+	mulsd	xmm7,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm10,xmm6				# q2+q3+x4
+	addsd	xmm3,xmm9				# p2+p3+p4	
+	addsd	xmm10,xmm4				# q1+q2+q3+x4
+	addsd	xmm2,xmm3				# p1+p2+p3+p4
+
+	movlpd	xmm3,qword ptr real_26_bits[rip]
+	movsd	xmm1,xmm5				# x
+
+	addsd	xmm10,qword ptr (log2_b_q_0-log2_b_c)[rcx] # q0+q1+q2+q3+x4
+	addsd	xmm2,qword ptr (log2_b_p_0-log2_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	subsd	xmm5,xmm7				# x-0.5x2
+
+	divsd	xmm2,xmm10				# p/q
+
+	andpd	xmm3,xmm5				# (x-0.5x2)_h
+	subsd	xmm1,xmm3				# x-(x-0.5x2)_h
+	subsd	xmm1,xmm7				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm8				# (p/q)*x3
+
+	mulsd	xmm5,qword ptr d1ln2_26_l[rip] # (x-0.5x2) * (1/ln 2)_l
+	mulsd	xmm1,qword ptr d1ln2_26[rip] # (x-0.5x2)_l * (1/ln 2)_h
+	mulsd	xmm3,qword ptr d1ln2_26[rip] # (x-0.5x2)_h * (1/ln 2)_h
+
+	addsd	xmm1,xmm5				# (x-0.5x2) * (1/ln 2)_l+(x-0.5x2)_l * (1/ln 2)_h
+
+
+	movsd	xmm5,xmm1				# xl
+	addsd	xmm1,xmm2				# xl+r
+	movlpd	xmm6,qword ptr real_26_bits[rip]
+
+	addsd	xmm1,xmm3				# xh+xl+r
+
+	andpd	xmm1,xmm6				# (log2 x)_h
+	andpd	xmm6,xmm0				# y_h
+
+	subsd	xmm3,xmm1				# xh-(log2 x)_h
+	jmp		pow_real_2
+
+pow_real_a_7_d_6:
+	ucomisd	xmm0,qword ptr real_4605_0[rip]
+	jae		pow_real_overflow		# because 1.1666666666666667407^4605>2^1024
+
+	ucomisd	xmm0,qword ptr real_m_4834_0[rip]
+	jbe		pow_real_underflow		# because 1.1666666666666667407^-4834<2^-1075
+
+	subsd	xmm1,xmm2
+
+	mulsd	xmm1,qword ptr real_0_75[rip]
+	lea		rcx,log2_b_c[rip]
+	lea		rdx,log2_s_c[rip]
+
+	movlpd	xmm2,qword ptr real_0_25[rip]
+
+	ucomisd	xmm1,xmm2
+	subsd	xmm1,xmm2
+
+	cmovb	rcx,rdx
+
+	movsd	xmm5,xmm1				# x
+	mulsd	xmm1,xmm1				# x2
+
+	movlpd	xmm2,qword ptr (log2_b_p_1-log2_b_c)[rcx]
+	movlpd	xmm4,qword ptr (log2_b_q_1-log2_b_c)[rcx]
+	mulsd	xmm2,xmm5				# p1
+	mulsd	xmm4,xmm5				# q1
+	movsd	xmm8,xmm5				# x
+	movlpd	xmm3,qword ptr (log2_b_p_2-log2_b_c)[rcx]
+	movlpd	xmm6,qword ptr (log2_b_q_2-log2_b_c)[rcx]
+	mulsd	xmm8,xmm1				# x3
+	movsd	xmm7,xmm1				# x2
+	mulsd	xmm1,xmm1				# x4
+
+	movlpd	xmm9,qword ptr (log2_b_p_3-log2_b_c)[rcx]
+	movlpd	xmm10,qword ptr (log2_b_q_3-log2_b_c)[rcx]
+
+	mulsd	xmm3,xmm7				# p2
+	mulsd	xmm6,xmm7				# q2
+
+	movlpd	xmm11,qword ptr (log2_b_p_4-log2_b_c)[rcx]
+
+	mulsd	xmm9,xmm8				# p3
+	mulsd	xmm10,xmm8				# q3
+	
+	mulsd	xmm11,xmm1				# p4
+	
+	addsd	xmm10,xmm1				# q3+x4
+	addsd	xmm9,xmm11				# p3+p4
+
+	mulsd	xmm7,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm10,xmm6				# q2+q3+x4
+	addsd	xmm3,xmm9				# p2+p3+p4	
+	addsd	xmm10,xmm4				# q1+q2+q3+x4
+	addsd	xmm2,xmm3				# p1+p2+p3+p4
+
+	movlpd	xmm3,qword ptr real_26_bits[rip]
+	movsd	xmm1,xmm5				# x
+
+	addsd	xmm10,qword ptr (log2_b_q_0-log2_b_c)[rcx] # q0+q1+q2+q3+x4
+	addsd	xmm2,qword ptr (log2_b_p_0-log2_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	subsd	xmm5,xmm7				# x-0.5x2
+
+	divsd	xmm2,xmm10				# p/q
+
+	andpd	xmm3,xmm5				# (x-0.5x2)_h
+	subsd	xmm1,xmm3				# x-(x-0.5x2)_h
+	subsd	xmm1,xmm7				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm8				# (p/q)*x3
+
+	mulsd	xmm5,qword ptr d1ln2_26_l[rip] # (x-0.5x2) * (1/ln 2)_l
+	mulsd	xmm1,qword ptr d1ln2_26[rip] # (x-0.5x2)_l * (1/ln 2)_h
+	mulsd	xmm3,qword ptr d1ln2_26[rip] # (x-0.5x2)_h * (1/ln 2)_h
+
+
+	movlpd	xmm6,qword ptr log2_4_d_3_53[rip]
+	movsd	xmm7,xmm3				# xh
+	addsd	xmm3,xmm6				# xh+(log2 4/3)_h
+
+	addsd	xmm1,xmm5				# (x-0.5x2) * (1/ln 2)_l+(x-0.5x2)_l * (1/ln 2)_h
+
+	subsd	xmm6,xmm3				# (log2 4/3)_h-(xh+(log2 4/3)_h)
+
+	addsd	xmm6,xmm7				# ((log2 4/3)_h-(xh+(log2 4/3)_h))+xh
+	addsd	xmm1,qword ptr log2_4_d_3_53_l[rip] # xl+(log2 4/3)_l
+
+	addsd	xmm1,xmm6				# (x+log2 4/3)_l
+
+	movsd	xmm5,xmm1				# xl
+	addsd	xmm1,xmm2				# xl+r
+	movlpd	xmm6,qword ptr real_26_bits[rip]
+
+	addsd	xmm1,xmm3				# xh+xl+r
+
+	andpd	xmm1,xmm6				# (log2 x)_h
+	andpd	xmm6,xmm0				# y_h
+
+	subsd	xmm3,xmm1				# xh-(log2 x)_h
+	att_jmp		pow_real_2
+
+pow_real_small:
+	ucomisd	xmm1,qword ptr real_2_p_m_1022[rip]
+	jb		pow_real_negative_zero_denormalized_or_nan
+
+pow_real_small_:
+	ucomisd	xmm0,qword ptr real_4000_0[rip]
+	att_jae		pow_real_underflow		# because 0.83^4000<2^-1075
+
+	ucomisd	xmm0,qword ptr real_m_3810_0[rip]
+	ja		log2_real_
+	att_jmp		pow_real_overflow		# because 0.83^-3810>2^1024
+
+pow_real_large:
+	ucomisd	xmm1,qword ptr real_max[rip]
+	ja		pow_real_inf
+
+	ucomisd	xmm0,qword ptr real_1401_0[rip]
+	att_jae		pow_real_overflow		# because 1.66^1401>2^1024
+
+	ucomisd	xmm0,qword ptr real_m_1471_0[rip]
+	att_jbe		pow_real_underflow		# because 1.66^-1471<2^-1075
+
+log2_real_:
+	movlpd	xmm2,qword ptr mask_significand[rip]
+	andpd	xmm2,xmm1
+	psrlq	xmm1,52
+
+log2_real__:
+	movlpd	xmm3,qword ptr real_1_0[rip]
+	movlpd	xmm4,qword ptr real_0_5[rip]
+	orpd	xmm3,xmm2
+	orpd	xmm4,xmm2
+
+	ucomisd	xmm3,qword ptr real_1_66[rip]
+
+	movlpd	xmm5,qword ptr real_1_0[rip]
+
+	ja		log2_real_l1
+
+	ucomisd	xmm3,qword ptr real_7_d_6[rip]
+	ja		pow_real_g_7_d_6
+
+log2_real_g1:
+	subsd	xmm3,xmm5
+
+	psubq	xmm1,xmmword ptr qw_1023[rip]
+
+	movsd	xmm6,xmm3				# x
+	mulsd	xmm3,xmm3				# x2
+
+	movlpd	xmm2,qword ptr log2_b_p_1[rip]
+	movlpd	xmm5,qword ptr log2_b_q_1[rip]
+	mulsd	xmm2,xmm6				# p1
+	mulsd	xmm5,xmm6				# q1
+	movsd	xmm9,xmm6				# x
+
+	cvtdq2pd xmm4,xmm1
+
+	movlpd	xmm1,qword ptr log2_b_p_2[rip]
+	movlpd	xmm7,qword ptr log2_b_q_2[rip]
+	mulsd	xmm9,xmm3				# x3
+	movsd	xmm8,xmm3				# x2
+	mulsd	xmm3,xmm3				# x4
+
+	movlpd	xmm10,qword ptr log2_b_p_3[rip]
+	movlpd	xmm11,qword ptr log2_b_q_3[rip]
+
+	mulsd	xmm1,xmm8				# p2
+	mulsd	xmm7,xmm8				# q2
+
+	movlpd	xmm12,qword ptr log2_b_p_4[rip]
+
+	mulsd	xmm10,xmm9				# p3
+	mulsd	xmm11,xmm9				# q3
+	
+	mulsd	xmm12,xmm3				# p4
+	
+	addsd	xmm11,xmm3				# q3+x4
+	addsd	xmm10,xmm12				# p3+p4
+
+	mulsd	xmm8,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm11,xmm7				# q2+q3+x4
+	addsd	xmm1,xmm10				# p2+p3+p4	
+	addsd	xmm11,xmm5				# q1+q2+q3+x4
+	addsd	xmm2,xmm1				# p1+p2+p3+p4
+
+	movlpd	xmm3,qword ptr real_26_bits[rip]
+	movsd	xmm1,xmm6				# x
+
+	addsd	xmm11,qword ptr log2_b_q_0[rip] # q0+q1+q2+q3+x4
+	addsd	xmm2,qword ptr log2_b_p_0[rip] # p0+p1+p2+p3+p4
+
+log2_real_gl:
+	subsd	xmm6,xmm8				# x-0.5x2
+
+	divsd	xmm2,xmm11				# p/q
+
+	andpd	xmm3,xmm6				# (x-0.5x2)_h
+	subsd	xmm1,xmm3				# x-(x-0.5x2)_h
+	subsd	xmm1,xmm8				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm9				# (p/q)*x3
+
+	mulsd	xmm6,qword ptr d1ln2_26_l[rip] # (x-0.5x2) * (1/ln 2)_l
+	mulsd	xmm1,qword ptr d1ln2_26[rip] # (x-0.5x2)_l * (1/ln 2)_h
+	mulsd	xmm3,qword ptr d1ln2_26[rip] # (x-0.5x2)_h * (1/ln 2)_h
+
+	addsd	xmm1,xmm6				# (x-0.5x2) * (1/ln 2)_l+(x-0.5x2)_l * (1/ln 2)_h
+
+	movsd	xmm5,xmm1				# xl
+	addsd	xmm1,xmm2				# xl+r
+	movlpd	xmm6,qword ptr real_26_bits[rip]
+
+	addsd	xmm1,xmm3				# xh+xl+r
+	
+	addsd	xmm1,xmm4				# e+xh+xl+r
+
+	andpd	xmm1,xmm6				# (log2 x)_h
+	andpd	xmm6,xmm0				# y_h
+	
+	subsd	xmm4,xmm1				# e-(log2 x)_h
+
+	addsd	xmm3,xmm4				# e+xh-(log2 x)_h
+	att_jmp		pow_real_2
+
+log2_real_l1:
+	subsd	xmm4,xmm5
+
+	psubq	xmm1,xmmword ptr qw_1022[rip]
+
+	movsd	xmm6,xmm4				# x
+	mulsd	xmm4,xmm4				# x2
+
+	movlpd	xmm2,qword ptr log2_s_p_1[rip]
+	movlpd	xmm5,qword ptr log2_s_q_1[rip]
+	mulsd	xmm2,xmm6				# p1
+	mulsd	xmm5,xmm6				# q1
+	movsd	xmm9,xmm6				# x
+
+	movlpd	xmm3,qword ptr log2_s_p_2[rip]
+	movlpd	xmm7,qword ptr log2_s_q_2[rip]
+	mulsd	xmm9,xmm4				# x3
+	movsd	xmm8,xmm4				# x2
+	mulsd	xmm4,xmm4				# x4
+
+	movlpd	xmm10,qword ptr log2_s_p_3[rip]
+	movlpd	xmm11,qword ptr log2_s_q_3[rip]
+
+	mulsd	xmm3,xmm8				# p2
+	mulsd	xmm7,xmm8				# q2
+
+	movlpd	xmm12,qword ptr log2_s_p_4[rip]
+
+	mulsd	xmm10,xmm9				# p3
+	mulsd	xmm11,xmm9				# q3
+	
+	mulsd	xmm12,xmm4				# p4
+
+	addsd	xmm11,xmm4				# q3+x4
+	addsd	xmm10,xmm12				# p3+p4
+
+	mulsd	xmm8,qword ptr real_0_5[rip] # 0.5x2
+
+	cvtdq2pd xmm4,xmm1
+
+	addsd	xmm11,xmm7				# q2+q3+x4
+	addsd	xmm3,xmm10				# p2+p3+p4	
+	addsd	xmm11,xmm5				# q1+q2+q3+x4
+	addsd	xmm2,xmm3				# p1+p2+p3+p4
+
+	movlpd	xmm3,qword ptr real_26_bits[rip]
+	movsd	xmm1,xmm6				# x
+
+	addsd	xmm11,qword ptr log2_s_q_0[rip] # q0+q1+q2+q3+x4
+	addsd	xmm2,qword ptr log2_s_p_0[rip] # p0+p1+p2+p3+p4
+
+	att_jmp		log2_real_gl
+
+pow_real_to_zero_or_nan:
+	jp		pow_real_to_nan
+	movlpd	xmm0,qword ptr real_1_0[rip]
+	ret
+
+pow_real_negative_zero_denormalized_or_nan:
+	ucomisd	xmm1,qword ptr real_0_0[rip]
+	jbe		pow_real_zero_negative_or_nan
+	
+	# denormalized > 0.0
+
+pow_real_denormalized:
+	ucomisd	xmm0,qword ptr real_1_052[rip]
+	att_jae		pow_real_underflow		# because (2^-1022)^1.052<2^-1075
+
+	ucomisd	xmm0,qword ptr real_m_1_052[rip]
+	att_jbe		pow_real_overflow		# because (2^-1022)^-1.052>2^1075
+
+	mulsd	xmm1,qword ptr real_2_p_55[rip]
+
+	movlpd	xmm2,qword ptr mask_significand[rip]
+	andpd	xmm2,xmm1
+
+	movq	xmm5,qword ptr qw_55[rip]
+	psrlq	xmm1,52
+
+	psubq	xmm1,xmm5
+	att_jmp		log2_real__
+
+pow_real_zero_negative_or_nan:
+	je		pow_real_zero_or_nan
+
+	movlpd	xmm2,qword ptr mask_all_except_sign[rip]
+	andpd	xmm2,xmm0				# |y|
+
+	ucomisd	xmm2,qword ptr real_2_p_53[rip]
+	jae		pow_real_negative_to_large_or_inf
+
+	ucomisd	xmm2,qword ptr real_1_0[rip]
+	jb		pow_real_negative_to_non_int
+
+	movq	xmm3,qword ptr round_c[rip]
+	psrlq	xmm2,52
+	psubq	xmm3,xmm2
+
+	movq	xmm4,qword ptr mask_all_one[rip]
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm4,xmm3
+	psllq	xmm5,xmm3
+	andpd	xmm4,xmm0
+	andpd	xmm5,xmm0				# 1 -> 2 and -1 -> -2
+
+	ucomisd	xmm4,xmm0
+	att_jne		pow_real_negative_to_non_int
+
+	ucomisd	xmm5,xmm0
+	je		pow_real_negative_to_even_int
+	
+pow_real_negative_to_odd_int:
+	att_call	pow_real_negative_to_even_int
+	movlpd	xmm1,qword ptr mask_sign[rip]
+	xorpd	xmm0,xmm1
+	ret
+
+pow_real_negative_to_even_int:
+	movlpd	xmm2,qword ptr mask_all_except_sign[rip]
+	andpd	xmm1,xmm2				# |x|
+	ucomisd	xmm1,qword ptr real_0_83[rip] # x>=0.83
+	att_jae		pow_real_not_small
+
+	ucomisd	xmm1,qword ptr real_2_p_m_1022[rip]
+	att_jae		pow_real_small_
+	att_jmp		pow_real_denormalized
+
+pow_real_negative_to_large_or_inf:
+	movlpd	xmm2,qword ptr mask_all_except_sign[rip]
+	andpd	xmm1,xmm2				# |x|
+	ucomisd	xmm1,qword ptr real_0_83[rip]
+	jb		pow_real_small_to_large_or_inf	# because 0.83^(2^53) too large or small
+	ucomisd	xmm1,qword ptr real_7_d_6[rip]
+
+	movlpd	xmm2,qword ptr real_1_0[rip]
+
+	ja		pow_real_large_to_large_or_inf	# because 1.16^(2^53) too large or small
+
+	att_jmp		pow_real_between_0_83_and_7_d_6
+
+pow_real_negative_to_non_int:
+	ucomisd	xmm1,qword ptr real_m_max[rip]
+	jb		pow_real_m_inf_to_non_int		# x==-inf
+	movlpd	xmm0,qword ptr real_0_0[rip]
+
+	divsd	xmm0,xmm0
+	ret
+
+pow_real_zero_or_nan:
+	jp		pow_real_nan_to_non_zero
+pow_real_zero:
+	movlpd	xmm2,qword ptr real_1_0[rip]
+	orpd	xmm2,xmm1						# 0.0 -> 1.0 and -0.0 -> -1.0
+	ucomisd	xmm2,qword ptr real_0_0[rip]
+	jb		pow_real_m_zero
+
+pow_real_small_to_large_or_inf:
+pow_real_m_zero_to_non_odd_int:
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	att_jb		pow_real_overflow
+
+pow_real_inf_to_negative:
+pow_real_underflow:
+	movlpd	xmm0,qword ptr real_0_0[rip]
+	ret
+
+pow_real_m_inf_to_non_int:
+pow_real_large_to_large_or_inf:
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	att_jb		pow_real_underflow
+
+pow_real_overflow:
+	movlpd	xmm0,qword ptr real_1_0[rip]
+	divsd	xmm0,qword ptr real_0_0[rip]	# inf
+	ret
+
+pow_real_m_zero:
+	movlpd	xmm2,qword ptr mask_all_except_sign[rip]
+	andpd	xmm2,xmm0				# |y|
+
+	ucomisd	xmm2,qword ptr real_2_p_53[rip]
+	att_jae		pow_real_m_zero_to_non_odd_int
+
+	ucomisd	xmm2,qword ptr real_m_1_0[rip]
+	att_jb		pow_real_m_zero_to_non_odd_int
+
+	movq	xmm4,qword ptr round_c[rip]
+	psrlq	xmm2,52
+	psubq	xmm4,xmm2
+
+	movq	xmm5,qword ptr mask_all_one_except_last[rip]
+	psllq	xmm5,xmm4
+	andpd	xmm5,xmm0				# 1 -> 2 and -1 -> -2
+
+	ucomisd	xmm5,xmm0
+	att_je		pow_real_m_zero_to_non_odd_int
+
+	ucomisd	xmm0,xmm3
+	jb		pow_real_m_overflow
+
+	movsd	xmm0,xmm1				# -0.0
+	ret
+
+pow_real_m_overflow:
+	movlpd	xmm0,qword ptr real_m_1_0[rip]
+	divsd	xmm0,qword ptr real_0_0[rip]	# -inf
+	ret
+
+pow_real_nan_to_non_zero:
+pow_real_to_nan:
+	addsd	xmm0,xmm1
+	ret
+
+pow_real_inf:
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	att_jb		pow_real_inf_to_negative
+	movsd	xmm0,xmm1
+	ret
+
+pow_real_exp_too_large:
+	ucomisd	xmm1,xmm2
+	je		power_one_to_large_or_inf
+	att_ja		pow_real_overflow
+	att_jmp		pow_real_underflow
+
+pow_real_exp_too_large_m:
+	ucomisd	xmm1,xmm2
+	je		power_one_to_large_or_inf_m
+	att_jb		pow_real_overflow
+	att_jmp		pow_real_underflow
+
+power_one_to_large_or_inf:
+	ucomisd	xmm0,qword ptr real_max[rip]
+	ja		power_one_to_inf
+	movsd	xmm0,xmm1
+	ret
+
+power_one_to_large_or_inf_m:
+	ucomisd	xmm0,qword ptr real_m_max[rip]
+	jb		power_one_to_inf_m
+	movsd	xmm0,xmm1
+	ret
+
+power_one_to_inf_m:
+power_one_to_inf:
+	subsd	xmm0,xmm0
+	ret
+
+
+pow_real_g_7_d_6:
+	subsd	xmm3,xmm5
+
+	mulsd	xmm3,qword ptr real_0_75[rip]
+	lea		rcx,log2_b_c[rip]
+	lea		rdx,log2_s_c[rip]
+
+	movlpd	xmm4,qword ptr real_0_25[rip]
+
+	ucomisd	xmm3,xmm4
+	subsd	xmm3,xmm4
+
+	cmovb	rcx,rdx
+	psubq	xmm1,xmmword ptr qw_1023[rip]
+
+	movsd	xmm6,xmm3				# x
+	mulsd	xmm3,xmm3				# x2
+
+	movlpd	xmm2,qword ptr (log2_b_p_1-log2_b_c)[rcx]
+	movlpd	xmm5,qword ptr (log2_b_q_1-log2_b_c)[rcx]
+	mulsd	xmm2,xmm6				# p1
+	mulsd	xmm5,xmm6				# q1
+	movsd	xmm9,xmm6				# x
+
+	cvtdq2pd xmm4,xmm1
+
+	movlpd	xmm1,qword ptr (log2_b_p_2-log2_b_c)[rcx]
+	movlpd	xmm7,qword ptr (log2_b_q_2-log2_b_c)[rcx]
+	mulsd	xmm9,xmm3				# x3
+	movsd	xmm8,xmm3				# x2
+	mulsd	xmm3,xmm3				# x4
+
+	movlpd	xmm10,qword ptr (log2_b_p_3-log2_b_c)[rcx]
+	movlpd	xmm11,qword ptr (log2_b_q_3-log2_b_c)[rcx]
+
+	mulsd	xmm1,xmm8				# p2
+	mulsd	xmm7,xmm8				# q2
+
+	movlpd	xmm12,qword ptr (log2_b_p_4-log2_b_c)[rcx]
+
+	mulsd	xmm10,xmm9				# p3
+	mulsd	xmm11,xmm9				# q3
+	
+	mulsd	xmm12,xmm3				# p4
+	
+	addsd	xmm11,xmm3				# q3+x4
+	addsd	xmm10,xmm12				# p3+p4
+
+	mulsd	xmm8,qword ptr real_0_5[rip] # 0.5x2
+
+	addsd	xmm11,xmm7				# q2+q3+x4
+	addsd	xmm1,xmm10				# p2+p3+p4	
+	addsd	xmm11,xmm5				# q1+q2+q3+x4
+	addsd	xmm2,xmm1				# p1+p2+p3+p4
+
+	movlpd	xmm3,qword ptr real_26_bits[rip]
+	movsd	xmm1,xmm6				# x
+
+	addsd	xmm11,qword ptr (log2_b_q_0-log2_b_c)[rcx] # q0+q1+q2+q3+x4
+	addsd	xmm2,qword ptr (log2_b_p_0-log2_b_c)[rcx] # p0+p1+p2+p3+p4
+
+	subsd	xmm6,xmm8				# x-0.5x2
+
+	addsd	xmm4,qword ptr log2_4_d_3_42[rip] # e+(log2 4/3)_h
+
+	divsd	xmm2,xmm11				# p/q
+
+	andpd	xmm3,xmm6				# (x-0.5x2)_h
+	subsd	xmm1,xmm3				# x-(x-0.5x2)_h
+	subsd	xmm1,xmm8				# (x-0.5x2)_l
+
+	mulsd	xmm2,xmm9				# (p/q)*x3
+
+	mulsd	xmm6,qword ptr d1ln2_26_l[rip] # (x-0.5x2) * (1/ln 2)_l
+	mulsd	xmm1,qword ptr d1ln2_26[rip] # (x-0.5x2)_l * (1/ln 2)_h
+	mulsd	xmm3,qword ptr d1ln2_26[rip] # (x-0.5x2)_h * (1/ln 2)_h
+
+	movsd	xmm7,xmm3				# xh
+	addsd	xmm3,xmm4				# xh+e+(log2 4/3)_h
+
+	addsd	xmm1,xmm6				# (x-0.5x2) * (1/ln 2)_l+(x-0.5x2)_l * (1/ln 2)_h
+
+	subsd	xmm4,xmm3				# e+(log2 4/3)_h-(xh+e+(log2 4/3)_h)
+
+	addsd	xmm4,xmm7				# (e+(log2 4/3)_h-(xh+e+(log2 4/3)_h))+xh
+	addsd	xmm1,qword ptr log2_4_d_3_42_l[rip] # xl+(log2 4/3)_l
+
+	addsd	xmm1,xmm4				# (x+log2 4/3)_l
+
+	movsd	xmm5,xmm1				# xl
+	addsd	xmm1,xmm2				# xl+r
+	movlpd	xmm6,qword ptr real_26_bits[rip]
+
+	addsd	xmm1,xmm3				# e+xh+xl+r
+	
+	andpd	xmm1,xmm6				# (log2 x)_h
+	andpd	xmm6,xmm0				# y_h
+	
+	subsd	xmm4,xmm1				# e-(log2 x)_h
+
+	addsd	xmm3,xmm4				# e+xh-(log2 x)_h
+#	jmp		pow_real_2
+
+
+pow_real_2:
+# xmm0 = y
+# xmm1 = (log2 x)_h
+# xmm2 = r
+# xmm3 = (e+)xh-(log2 x)_h
+# xmm5 = xl
+# xmm6 = y_h
+	movsd	xmm7,xmm0				# y
+	subsd	xmm0,xmm6				# y_l
+	mulsd	xmm6,xmm1				# p_h = y_h*(log2 x)_h
+
+	addsd	xmm3,xmm2				# xh+r-(log2 x)_h
+	mulsd	xmm0,xmm1				# y_l*(log2 x)_h
+
+	addsd	xmm3,xmm5				# (log2 x)_l
+
+	mulsd	xmm3,xmm7				# y*(log2 x)_l
+
+	addsd	xmm0,xmm3				# p_l
+
+	movsd	xmm7,xmm0				# p_l
+	addsd	xmm0,xmm6				# p
+
+exp2:
+	ucomisd	xmm0,qword ptr real_m_0_5[rip]
+	jb		exp2_l_n
+	ucomisd	xmm0,qword ptr real_0_5[rip]
+	ja		exp2_l
+exp2_l_:
+	ucomisd	xmm0,qword ptr real_m_0_25[rip]
+	jbe		exp2_l_m_0_25
+	ucomisd	xmm0,qword ptr real_0_25[rip]
+	jae		exp2_g_0_25
+
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+
+	lea		rcx,exp2_p0_c[rip]
+	lea		rdx,exp2_m0_c[rip]
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+
+	cmovb	rcx,rdx
+exp2_:
+	movlpd	xmm13,qword ptr (exp2_p0_2-exp2_p0_c)[rcx]
+	movlpd	xmm14,qword ptr (exp2_p0_4-exp2_p0_c)[rcx]
+
+	movsd	xmm2,xmm0				# x2
+	mulsd	xmm0,xmm0				# x4
+
+	mulsd	xmm13,xmm2				# p2
+
+	movsd	xmm3,xmm1				# x
+	mulsd	xmm1,xmm2				# x3
+
+	movlpd	xmm15,qword ptr (exp2_p0_3-exp2_p0_c)[rcx]
+	mulsd	xmm14,xmm0				# p4
+	
+	mulsd	xmm15,xmm1				# p3
+
+	movsd	xmm4,xmm0				# x4
+	mulsd	xmm0,xmm0				# x8
+
+	movsd	xmm5,xmm2				# x2
+	mulsd	xmm2,xmm1				# x5	
+
+	mulsd	xmm5,xmm4				# x6
+	mulsd	xmm1,xmm4				# x7
+	mulsd	xmm4,xmm2				# x9
+
+	mulsd	xmm0,qword ptr (exp2_p0_8-exp2_p0_c)[rcx] # p8
+	mulsd	xmm2,qword ptr (exp2_p0_5-exp2_p0_c)[rcx] # p5
+	mulsd	xmm5,qword ptr (exp2_p0_6-exp2_p0_c)[rcx] # p6
+	mulsd	xmm1,qword ptr (exp2_p0_7-exp2_p0_c)[rcx] # p7
+	mulsd	xmm4,qword ptr (exp2_p0_9-exp2_p0_c)[rcx] # p9
+
+	addsd	xmm0,xmm4				# p9+p8
+
+	movlpd	xmm9,qword ptr real_26_bits[rip]
+	andpd	xmm9,xmm6				# xh_h
+	mulsd	xmm3,qword ptr (exp2_p0_1l-exp2_p0_c)[rcx] # x*c1l
+
+	addsd	xmm0,xmm1				# p9+p8+p7
+
+	subsd	xmm6,xmm9				# xh_l
+	mulsd	xmm9,qword ptr (exp2_p0_1h-exp2_p0_c)[rcx] # p1h=xh_h*c1h
+
+	addsd	xmm0,xmm5				# p9+p8+p7+p6
+
+	movlpd	xmm1,qword ptr (exp2_p0_0h-exp2_p0_c)[rcx]
+	addsd	xmm6,xmm7				# xh_l+x_l
+	movsd	xmm4,xmm9				# p1h
+	addsd	xmm9,xmm1				# p1h+c0h
+
+	addsd	xmm0,xmm2				# p9+p8+p7+p6+p5
+
+	mulsd	xmm6,qword ptr (exp2_p0_1h-exp2_p0_c)[rcx] # (xh_l+x_l)*c1h
+	subsd	xmm1,xmm9				# c0h-(p1h+c0h)
+
+	addsd	xmm0,xmm14				# p9+p8+p7+p6+p5+p4
+
+	addsd	xmm3,xmm6				# (xh_l+x_l)*c1h+x*c1l
+	addsd	xmm1,xmm4				# (c0h-(p1h+c0h))+p1h
+
+	addsd	xmm0,xmm15				# p9+p8+p7+p6+p5+p4+p3
+
+	addsd	xmm1,xmm3
+
+	addsd	xmm0,xmm13				# p9+p8+p7+p6+p5+p4+p3+p2
+
+	addsd	xmm0,qword ptr (exp2_p0_0l-exp2_p0_c)[rcx]
+	
+	addsd	xmm0,xmm1
+	addsd	xmm0,xmm9
+	ret
+
+exp2_l_m_0_25:
+	movlpd	xmm1,qword ptr real_0_25[rip]
+	addsd	xmm0,xmm1
+	addsd	xmm6,xmm1
+
+	lea		rcx,exp2_m0_25_c[rip]
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+	att_jmp		exp2_
+
+exp2_g_0_25:
+	movlpd	xmm1,qword ptr real_0_25[rip]
+	subsd	xmm0,xmm1
+	subsd	xmm6,xmm1
+
+	lea		rcx,exp2_p0_25_c[rip]
+
+	movsd	xmm1,xmm0				# x
+	mulsd	xmm0,xmm0				# x2
+	att_jmp		exp2_
+
+exp2_l_n:
+	movlpd	xmm1,qword ptr real_m_0_5[rip]
+	movq	xmm4,qword ptr m_round_c[rip]
+	jmp		exp2_l_pn
+
+exp2_l:
+	movlpd	xmm1,qword ptr real_0_5[rip]
+	movq	xmm4,qword ptr round_c[rip]
+
+exp2_l_pn:
+	addsd	xmm1,xmm0
+
+	movq	xmm2,xmm1
+	psrlq	xmm1,52
+	psubq	xmm4,xmm1
+	movq	xmm1,qword ptr mask_all_one[rip]
+	psllq	xmm1,xmm4
+	andpd	xmm1,xmm2
+
+	subsd	xmm6,xmm1
+	movsd	xmm0,xmm7
+
+	cvtpd2dq xmm8,xmm1
+
+	ucomisd	xmm1,qword ptr qword ptr real_1023[rip]
+	ja		exp2_overflow0
+	ucomisd	xmm1,qword ptr qword ptr real_m_1022[rip]
+	jb		exp2_underflow0
+
+	movq	xmm9,qword ptr qw_1023[rip]
+
+	addsd	xmm0,xmm6
+
+	paddq	xmm8,xmm9
+
+	psllq	xmm8,52
+
+	att_call	exp2_l_
+
+	mulsd	xmm0,xmm8
+	ret
+
+exp2_overflow0:
+	ucomisd	xmm1,qword ptr qword ptr real_1025[rip]
+	ja		exp2_overflow
+
+	movq	xmm9,qword ptr qw_1021[rip]
+
+	addsd	xmm0,xmm6
+
+	paddq	xmm8,xmm9
+
+	psllq	xmm8,52
+
+	att_call	exp2_l_
+
+	mulsd	xmm0,xmm8
+	mulsd	xmm0,qword ptr real_4_0[rip]
+	ret
+
+exp2_overflow:
+	movlpd	xmm0,qword ptr qword ptr real_2_p_1023[rip]
+	mulsd	xmm0,xmm0
+	ret
+
+exp2_underflow0:
+	ucomisd	xmm1,qword ptr qword ptr real_m_1076[rip]
+	jbe		exp2_underflow
+
+	movq	xmm9,qword ptr qw_1078[rip]
+
+	addsd	xmm0,xmm6
+
+	paddq	xmm8,xmm9
+
+	psllq	xmm8,52
+
+	att_call	exp2_l_
+
+	mulsd	xmm0,xmm8
+	mulsd	xmm0,qword ptr real_2_p_m_55[rip]
+	ret
+
+exp2_underflow:
+	movlpd	xmm0,qword ptr qword ptr real_0_0[rip]
+	ret
+
+	.data
+
+real_13_bits:
+	.quad	0xffffff0000000000
+real_14_bits:
+	.quad	0xffffff8000000000
+real_17_bits:
+	.quad	0xfffffff000000000
+real_18_bits:
+	.quad	0xfffffff800000000
+real_21_bits:
+	.quad	0xffffffff00000000
+real_25_bits:
+	.quad	0xfffffffff0000000
+real_26_bits:
+	.quad	0xfffffffff8000000
+real_27_bits:
+	.quad	0xfffffffffc000000
+real_34_bits:
+	.quad	0xfffffffffff80000
+real_39_bits:
+	.quad	0xffffffffffffc000
+real_40_bits:
+	.quad	0xffffffffffffe000
+real_47_bits:
+	.quad	0xffffffffffffffc0
+real_48_bits:
+	.quad	0xffffffffffffffe0
+real_49_bits:
+	.quad	0xfffffffffffffff0
+real_50_bits:
+	.quad	0xfffffffffffffff8
+real_51_bits:
+	.quad	0xfffffffffffffffc
+
+real_abs_40_bits:
+	.quad	0x7fffffffffffe000
+real_abs_39_bits:
+	.quad	0x7fffffffffffc000
+
+real_0_0:
+	.double	0r0.0
+real_0_25:
+	.double	0r0.25
+real_m_0_25:
+	.double	0r-0.25
+real_0_5:
+	.double	0r0.5
+real_m_0_5:
+	.double	0r-0.5
+real_1_0:
+	.double	0r1.0
+real_m_1_0:
+	.double	0r-1.0
+real_1_25:
+	.double	0r1.25
+real_m_1_25:
+	.double	0r-1.25
+real_1_5:
+	.double	0r1.5
+real_m_1_5:
+	.double	0r-1.5
+real_2_0:
+	.double	0r2.0
+real_m_2_0:
+	.double	0r-2.0
+real_3_0:
+	.double	0r3.0
+real_m_3_0:
+	.double	0r-3.0
+real_4_0:
+	.double	0r4.0
+real_m_4_0:
+	.double	0r-4.0
+real_5_0:
+	.double	0r5.0
+real_m_5_0:
+	.double	0r-5.0
+
+real_2_p_m_11:
+	.double	0r4.8828125E-4
+real_2_p_m_12:
+	.double	0r2.44140625E-4
+real_m_2_p_m_12:
+	.double	0r-2.44140625E-4
+real_2_p_m_13:
+	.double	0r1.220703125E-4
+real_2_p_m_16:
+	.double	0r1.52587890625E-5
+real_m_2_p_m_16:
+	.double	0r-1.52587890625E-5
+real_2_p_m_18:
+	.double	0r3.814697265625E-6
+real_2_p_m_31:
+	.double	0r4.656612873077392578125E-10
+
+real_0_4:
+	.double	0r0.4
+real_m_0_4:
+	.double	0r-0.4
+real_0_125:
+	.double	0r0.125
+
+real_0_675:
+	.double	0r0.675
+real_m_0_675:
+	.double	0r-0.675
+real_0_9920:
+	.double	0r0.9920
+real_0_54000000017867999524:
+	.double	0r0.54000000017867999524
+real_m_0_54000000017867999524:
+	.double	0r-0.54000000017867999524
+real_0_03125:
+	.double	0r0.03125
+real_pi_d_2_m_0_03125:
+	.double	0r1.5395463267948965579
+
+real_0_58:
+	.double	0r0.58
+real_m_0_58:
+	.double	0r-0.58
+real_0_54:
+	.double	0r0.54
+real_0_76:
+	.double	0r0.76
+real_0_65000000004061742054:
+	.double	0r0.65000000004061742054
+
+real_0_26:
+	.double	0r0.26
+real_m_0_26:
+	.double	0r-0.26
+real_0_70:
+	.double	0r0.70
+real_m_0_70:
+	.double	0r-0.70
+real_1_4:
+	.double	0r1.4
+real_m_1_4:
+	.double	0r-1.4
+
+real_0_7025:
+	.double	0r0.7025
+real_1_405:
+	.double	0r1.405
+real_0_8243:
+	.double	0r0.82436063535006407342
+real_1_6487:
+	.double	0r1.6487212707001281468
+real_0_833:
+	.double	0r0.833
+real_1_666:
+	.double	0r1.666
+real_7_d_6:
+	.double	0r1.1666666666666667407
+real_0_75:
+	.double	0r0.75
+
+ln_4_d_3_53:
+	.double	0r2.8768207245178090137E-1
+ln_4_d_3_53_l:
+	.double	0r2.6071606164425639761E-17
+ln_4_d_3_42:
+	.double	0r2.8768207245173016417E-1
+ln_4_d_3_42_l:
+	.double	0r5.0763263831534079404E-14
+
+log10_4_d_3_53:
+	.double	0r1.2493873660829994676E-1
+log10_4_d_3_53_l:
+	.double	0r6.3704258509422005489E-18
+log10_4_d_3_42:
+	.double	0r1.2493873660829990513E-1
+log10_4_d_3_42_l:
+	.double	0r4.8003789274385573126E-17
+
+log2_4_d_3_53:
+	.double	0r4.1503749927884381332E-1
+log2_4_d_3_53_l:
+	.double	0r5.2244900613901090572E-18
+log2_4_d_3_42:
+	.double	0r4.1503749927881017356E-1
+log2_4_d_3_42_l:
+	.double	0r3.3644982136203633607E-14
+
+mask_significand:
+	.quad	0x0000fffffffffffff
+
+real_1_66:
+	.double	0r1.66
+real_0_83:
+	.double	0r0.83
+
+real_pi:
+	.double	0r3.1415926535897931160
+real_m_pi:
+	.double	0r-3.1415926535897931160
+real_pi_l:
+	.double	0r1.2246467991473532072E-16
+real_m_pi_l:
+	.double	0r-1.2246467991473532072E-16
+
+real_pi_d_2:
+	.double	0r1.5707963267948965579
+real_pi_d_2_l:
+	.double	0r6.1232339957367660359E-17
+
+real_m_pi_d_2:
+	.double	0r-1.5707963267948965579
+real_m_pi_d_2_l:
+	.double	0r-6.1232339957367660359E-17
+
+real_pi_m_1_d_2:
+	.double	0r1.07079632679489655800
+real_pi_d_4:
+	.double	0r7.8539816339744827900E-1
+real_m_pi_d_4:
+	.double	0r-7.8539816339744827900E-1
+real_pi_d_4_l:
+	.double	0r3.0616169978683830179E-17
+real_m_pi_d_4_l:
+	.double	0r-3.0616169978683830179E-17
+real_pi_p_0_375_d_4:
+	.double	0r8.7914816339744830962E-1
+real_pi_p_0_875_d_4:
+	.double	0r1.0041481633974482790
+real_pi_p_1_5_d_4:
+	.double	0r1.1603981633974483096
+real_3_pi_d_4:
+	.double	0r2.3561944901923449288
+real_m_3_pi_d_4:
+	.double	0r-2.3561944901923449288
+real_5_pi_d_4:
+	.double	0r3.9269908169872413950
+real_m_5_pi_d_4:
+	.double	0r-3.9269908169872413950
+real_7_pi_d_4:
+	.double	0r5.4977871437821379530
+real_m_7_pi_d_4:
+	.double	0r-5.4977871437821379530
+real_9_pi_d_4:
+	.double	0r7.0685834705770345110
+real_m_9_pi_d_4:
+	.double	0r-7.0685834705770345110
+real_3_pi_d_2:
+	.double	0r4.7123889803846896740
+real_m_3_pi_d_2:
+	.double	0r-4.7123889803846896740
+real_3_pi_d_2_l:
+	.double	0r1.8369701987210296875E-16
+real_m_3_pi_d_2_l:
+	.double	0r-1.8369701987210296875E-16
+real_2_pi:
+	.double	0r6.2831853071795862320
+real_2_pi_l:
+	.double	0r2.4492935982947064143E-16
+real_m_2_pi_l:
+	.double	0r-2.4492935982947064143E-16
+
+real_4_d_pi:
+	.double	0r1.2732395447351627649
+real_pi_d_4_26:
+	.double	0r7.8539815545082092285E-1
+real_pi_d_4_26_26:
+	.double	0r7.9466273561479283671E-9
+real_pi_d_4_52_l:
+	.double	0r3.0616169978683830179E-17
+
+real_36825084_pi:
+	.double	0r115689413.36222703755
+real_m_36825084_pi:
+	.double	0r-115689413.36222703755
+real_36825084_pi_27:
+	.double	0r115689413.0
+real_36825084_pi_27_27:
+	.double	0r3.6222703754901885986E-1
+real_36825084_pi_54_l:
+	.double	0r1.3588030639188803060E-17
+
+real_pi_d_2_m_1_0_52:
+	.double	0r5.7079632679489655800E-1
+real_1_0_m_pi_d_2_52:
+	.double	0r-5.7079632679489655800E-1
+real_pi_d_2_m_1_0_52_l:
+	.double	0r6.1232339957367660359E-17
+real_m_pi_d_2_m_1_0_52_l:
+	.double	0r-6.1232339957367660359E-17
+real_pi_d_2_p_1_0:
+	.double	0r2.57079632679489655800
+real_pi_d_2_p_1_0_l:
+	.double	0r6.1232339957367660359E-17
+real_m_pi_d_2_p_1_0_l:
+	.double	0r-6.1232339957367660359E-17
+
+real_0_338:
+	.double	0r0.338
+real_m_0_338:
+	.double	0r-0.338
+real_0_542:
+	.double	0r0.542
+real_m_0_542:
+	.double	0r-0.542
+real_0_699:
+	.double	0r0.699
+real_m_0_699:
+	.double	0r-0.699
+real_0_89:
+	.double	0r0.89
+real_m_0_89:
+	.double	0r-0.89
+real_1_04:
+	.double	0r1.04
+real_m_1_04:
+	.double	0r-1.04
+real_1_18:
+	.double	0r1.18
+real_m_1_18:
+	.double	0r-1.18
+
+real_0_5_pi_m_1_18:
+	.double	0r3.9079632679489667568E-1
+real_1_18_m_0_5_pi:
+	.double	0r-3.9079632679489667568E-1
+real_0_5_pi_m_1_04:
+	.double	0r5.3079632679489663349E-1
+real_1_04_m_0_5_pi:
+	.double	0r-5.3079632679489663349E-1
+real_0_5_pi_m_0_89:
+	.double	0r6.8079632679489665570E-1
+real_0_89_m_0_5_pi:
+	.double	0r-6.8079632679489665570E-1
+
+real_pi_m_1_18:
+	.double	0r1.9615926535897934002
+real_n_pi_m_1_18:
+	.double	0r-1.9615926535897934002
+real_pi_m_1_04:
+	.double	0r2.1015926535897930805
+real_n_pi_m_1_04:
+	.double	0r-2.1015926535897930805
+real_pi_m_0_89:
+	.double	0r2.2515926535897934357
+real_n_pi_m_0_89:
+	.double	0r-2.2515926535897934357
+real_pi_m_0_699:
+	.double	0r2.4425926535897932723
+real_n_pi_m_0_699:
+	.double	0r-2.4425926535897932723
+real_pi_m_0_542:
+	.double	0r2.5995926535897933007
+real_n_pi_m_0_542:
+	.double	0r-2.5995926535897933007
+real_pi_m_0_338:
+	.double	0r2.8035926535897930378
+real_n_pi_m_0_338:
+	.double	0r-2.8035926535897930378
+
+real_pi_p_0_338:
+	.double	0r3.4795926535897931942
+real_n_pi_p_0_338:
+	.double	0r-3.4795926535897931942
+real_pi_p_0_542:
+	.double	0r3.6835926535897933753
+real_n_pi_p_0_542:
+	.double	0r-3.6835926535897933753
+real_pi_p_0_699:
+	.double	0r3.8405926535897934038
+real_n_pi_p_0_699:
+	.double	0r-3.8405926535897934038
+real_pi_p_0_89:
+	.double	0r4.0315926535897936844
+real_n_pi_p_0_89:
+	.double	0r-4.0315926535897936844
+real_pi_p_1_04:
+	.double	0r4.1815926535897931515
+real_n_pi_p_1_04:
+	.double	0r-4.1815926535897931515
+real_pi_p_1_18:
+	.double	0r4.3215926535897928318
+real_n_pi_p_1_18:
+	.double	0r-4.3215926535897928318
+
+real_2pi_m_1_18:
+	.double	0r5.1031853071795865162
+real_n_2pi_m_1_18:
+	.double	0r-5.1031853071795865162
+real_2pi_m_1_04:
+	.double	0r5.2431853071795861965
+real_n_2pi_m_1_04:
+	.double	0r-5.2431853071795861965
+real_2pi_m_0_89:
+	.double	0r5.3931853071795865517
+real_n_2pi_m_0_89:
+	.double	0r-5.3931853071795865517
+real_2pi_m_0_699:
+	.double	0r5.5841853071795863883
+real_n_2pi_m_0_699:
+	.double	0r-5.5841853071795863883
+real_2pi_m_0_542:
+	.double	0r5.7411853071795864167
+real_n_2pi_m_0_542:
+	.double	0r-5.7411853071795864167
+real_2pi_m_0_338:
+	.double	0r5.9451853071795861538
+real_n_2pi_m_0_338:
+	.double	0r-5.9451853071795861538
+
+real_2pi_p_0_338:
+	.double	0r6.6211853071795863102
+real_n_2pi_p_0_338:
+	.double	0r-6.6211853071795863102
+real_2pi_p_0_542:
+	.double	0r6.8251853071795869354
+real_n_2pi_p_0_542:
+	.double	0r-6.8251853071795869354
+real_2pi_p_0_699:
+	.double	0r6.9821853071795860757
+real_n_2pi_p_0_699:
+	.double	0r-6.9821853071795860757
+real_2pi_p_0_89:
+	.double	0r7.1731853071795868004
+real_n_2pi_p_0_89:
+	.double	0r-7.1731853071795868004
+
+atan_0_5_53:
+	.double	0r4.6364760900080609352E-1
+m_atan_0_5_53:
+	.double	0r-4.6364760900080609352E-1
+atan_0_5_53_l:
+	.double	0r2.2698777452961687092E-17
+
+atan_sqrt_0_5:
+	.double	0r6.1547970867038737097E-1
+m_atan_sqrt_0_5:
+	.double	0r-6.1547970867038737097E-1
+atan_sqrt_0_5_l:
+	.double	0r-2.9904856561351187768E-17
+
+real_h_pi_m_atan_s_2_53:
+	.double	0r6.1547970867038737097E-1
+real_h_pi_m_atan_s_2_53_l:
+	.double	0r-2.9904856561351187768E-17
+
+real_sqrt_2_p_1_5_pi_m_3_atan_sqrt_2_53:
+	.double	0r3.2606526883842570363
+real_sqrt_2_p_1_5_pi_m_3_atan_sqrt_2_53_l:
+	.double	0r3.5657102106448616432E-17
+real_3_atan_sqrt_2_m_sqrt_2_m_1_5_pi_m_53_l:
+	.double	0r-3.5657102106448616432E-17
+
+atan_sqrt_2:
+	.double	0r9.5531661812450929805E-1
+m_atan_sqrt_2:
+	.double	0r-9.5531661812450929805E-1
+atan_sqrt_2_l:
+	.double	0r-1.9885105943796805915E-17
+
+real_h_pi_m_atan_2_52:
+	.double	0r4.6364760900080603800E-1
+real_h_pi_m_atan_2_52_l:
+	.double	0r7.8209928684219507951E-17
+real_2_p_2_5_pi_m_5_atan_2_53:
+	.double	0r4.3182380450040307451
+real_2_p_2_5_pi_m_5_atan_2_53_l:
+	.double	0r-1.6406186889148070581E-16
+real_5_atan_2_m_2_m_2_5_pi_53_l:
+	.double	0r1.6406186889148070581E-16
+
+atan_2_53:
+	.double	0r1.1071487177940904090
+m_atan_2_53:
+	.double	0r-1.1071487177940904090
+atan_2_53_l:
+	.double	0r9.4044713735663794124E-17
+real_n_2_m_5_atan_2_53:
+	.double	0r-3.5357435889704524890
+real_n_2_m_5_atan_2_53_l:
+	.double	0r-2.6134358828256379104E-17
+
+real_5_atan_2_m_2_53:
+	.double	0r3.5357435889704524890
+real_5_atan_2_m_2_53_l:
+	.double	0r2.6134358828256379104E-17
+
+real_ln_2_t_0_5:
+	.double	0r3.4657359027997265471E-1
+real_n_ln_2_t_0_5:
+	.double	0r-3.4657359027997265471E-1
+real_ln_2_t_1_5:
+	.double	0r1.0397207708399179641
+real_ln2_42:
+	.double	0r0.69314718055989033019
+real_ln2_42_l:
+	.double	0r5.4979230187083711552E-14
+real_1_d_ln2:
+	.double	0r1.4426950408889634074
+
+real_log2_10_42:
+	.double	0r0.30102999566395283182
+real_log2_10_42_l:
+	.double	0r2.8363394551044964154E-14
+
+#d1ln10:
+#	.double	0r4.3429448190325181667E-1
+d1ln10_26:
+	.double	0r4.3429447710514068603E-1
+d1ln10_26_l:
+	.double	0r4.7981111416159728067E-9
+
+#d1ln2:
+#	.double	0r1.4426950408889633870
+d1ln2_26:
+	.double	0r1.4426950216293334961
+d1ln2_26_l:
+	.double	0r1.9259629911266174689E-8
+
+	.align	4
+round_c:
+	.quad		0x3ff+52
+	.quad		0x3ff+52
+m_round_c:
+	.quad		0x0bff+52
+	.quad		0x0bff+52
+round_even_c:
+	.quad		0x3ff+53
+	.quad		0x3ff+53
+m_round_even_c:
+	.quad		0x0bff+53
+	.quad		0x0bff+53
+mask_all_one:
+	.quad		-1
+	.quad		-1
+mask_all_one_except_last:
+	.quad		-2
+	.quad		-2
+mask_all_one_except_second_last:
+	.quad		-3
+	.quad		-3
+qw_1023:
+	.quad		0x3ff
+	.quad		0x3ff
+qw_1022:
+	.quad		0x3fe
+	.quad		0x3fe
+qw_1021:
+	.quad		0x3fd
+	.quad		0x3fd
+qw_1077:
+	.quad		0x3ff+54
+	.quad		0x3ff+54
+qw_1078:
+	.quad		0x3ff+55
+	.quad		0x3ff+55
+qw_55:
+	.quad		55
+	.quad		55
+mask_sign:
+	.quad		0x08000000000000000
+	.quad		0x08000000000000000
+mask_all_except_sign:
+	.quad		0x07fffffffffffffff
+	.quad		0x07fffffffffffffff
+
+real_0_43540000008249979402:
+	.double	0r0.43540000008249979402
+real_m_0_43540000008249979402:
+	.double	0r-0.43540000008249979402
+real_0_600000000082499762577:
+	.double	0r0.600000000082499762577
+real_m_0_600000000082499762577:
+	.double	0r-0.600000000082499762577
+
+#sin_p_0:
+#	.double	0r-1.6666666666666629659E-1
+#sin_p_1:
+#	.double	0r8.3333333333220592498E-3
+#sin_p_2:
+#	.double	0r-1.9841269829591994520E-4
+#sin_p_3:
+#	.double	0r2.7557313627980839565E-6
+#sin_p_4:
+#	.double	0r-2.5050749232606622552E-8
+#sin_p_5:
+#	.double	0r1.5896334589820396677E-10
+
+sin_p_0:
+	.double	0r-1.6666666666666635210E-1
+sin_p_1:
+	.double	0r8.3333333333230012047E-3
+sin_p_2:
+	.double	0r-1.9841269830204560616E-4
+sin_p_3:
+	.double	0r2.7557313816586650493E-6
+sin_p_4:
+	.double	0r-2.5050776321203382309E-8
+sin_p_5:
+	.double	0r1.5897788850050641142E-10
+
+sin_p1_c:
+sin_p1_0h:
+	.double	0r5.6464247346312534948E-1
+sin_p1_0l:
+	.double	0r-6.06241788103829245082E-24
+sin_p1_1h:
+	.double	0r8.2533560693264007568E-1
+sin_p1_1l:
+	.double	0r7.9304552924597260244E-9
+sin_p1_2:
+	.double	0r-2.8232123673155168353E-1
+sin_p1_3:
+	.double	0r-1.3755593581121927649E-1
+sin_p1_4:
+	.double	0r2.3526769749636338575E-2
+sin_p1_5:
+	.double	0r6.8777964033410682668E-3
+sin_p1_6:
+	.double	0r-7.8422158064252829243E-4
+sin_p1_7:
+	.double	0r-1.6378316139497952112E-4
+sin_p1_8:
+	.double	0r1.4102652890564176127E-5
+sin_p1_9:
+	.double	0r2.0754701007995074451E-6
+
+sin_p2_c:
+sin_p2_0h:
+	.double	0r5.6464247346312534948E-1
+sin_p2_0l:
+	.double	0r-6.06241788103829245082E-24
+sin_p2_1h:
+	.double	0r8.2533560693264007568E-1
+sin_p2_1l:
+	.double	0r7.9304554034820284869E-9
+sin_p2_2:
+	.double	0r-2.8232123673155273824E-1
+sin_p2_3:
+	.double	0r-1.3755593580986152924E-1
+sin_p2_4:
+	.double	0r2.3526769748731638648E-2
+sin_p2_5:
+	.double	0r6.8777971733520841535E-3
+sin_p2_6:
+	.double	0r-7.8422151328288352071E-4
+sin_p2_7:
+	.double	0r-1.6373002515216467253E-4
+sin_p2_8:
+	.double	0r1.4106512050855138886E-5
+sin_p2_9:
+	.double	0r2.4764422854631738044E-6
+
+sin_p3_c:
+sin_p3_0h:
+	.double	0r-5.6464247346312534948E-1
+sin_p3_0l:
+	.double	0r6.06241788103829245082E-24
+sin_p3_1h:
+	.double	0r8.2533560693264007568E-1
+sin_p3_1l:
+	.double	0r7.9304552924597260244E-9
+sin_p3_2:
+	.double	0r2.8232123673155168353E-1
+sin_p3_3:
+	.double	0r-1.3755593581121927649E-1
+sin_p3_4:
+	.double	0r-2.3526769749636338575E-2
+sin_p3_5:
+	.double	0r6.8777964033410682668E-3
+sin_p3_6:
+	.double	0r7.8422158064252829243E-4
+sin_p3_7:
+	.double	0r-1.6378316139497952112E-4
+sin_p3_8:
+	.double	0r-1.4102652890564176127E-5
+sin_p3_9:
+	.double	0r2.0754701007995074451E-6
+
+sin_p4_c:
+sin_p4_0h:
+	.double	0r-5.6464247346312534948E-1
+sin_p4_0l:
+	.double	0r6.06241788103829245082E-24
+sin_p4_1h:
+	.double	0r8.2533560693264007568E-1
+sin_p4_1l:
+	.double	0r7.9304554034820284869E-9
+sin_p4_2:
+	.double	0r2.8232123673155273824E-1
+sin_p4_3:
+	.double	0r-1.3755593580986152924E-1
+sin_p4_4:
+	.double	0r-2.3526769748731638648E-2
+sin_p4_5:
+	.double	0r6.8777971733520841535E-3
+sin_p4_6:
+	.double	0r7.8422151328288352071E-4
+sin_p4_7:
+	.double	0r-1.6373002515216467253E-4
+sin_p4_8:
+	.double	0r-1.4106512050855138886E-5
+sin_p4_9:
+	.double	0r2.4764422854631738044E-6
+
+cos_p_0:
+	.double	0r4.1666666666666588026E-2
+cos_p_1:
+	.double	0r-1.3888888888872778175E-3
+cos_p_2:
+	.double	0r2.4801587288764459412E-5
+cos_p_3:
+	.double	0r-2.7557314167050702998E-7
+cos_p_4:
+	.double	0r2.0875701408714520036E-9
+cos_p_5:
+	.double	0r-1.1358718309736171193E-11
+
+# .. .5854
+#tan_p_0:
+#	.double	0r-18086151.600729089230
+#tan_p_1:
+#	.double	0r1161461.5827075217385
+#tan_p_2:
+#	.double	0r-13169.960960193349820
+#tan_q_0:
+#	.double	0r-54258454.802187263966
+#tan_q_1:
+#	.double	0r25187766.668997306377
+#tan_q_2:
+#	.double	0r-1329914.3444088697433
+#tan_q_3:
+#	.double	0r13760.338151945812569
+
+# .. .338
+tan_p_0:
+	.double	0r-18191117.789276178926
+tan_p_1:
+	.double	0r1167963.0645372841973
+tan_p_2:
+	.double	0r-13234.676172752202547
+tan_q_0:
+	.double	0r-54573353.367828540504
+tan_q_1:
+	.double	0r25333230.540743269026
+tan_q_2:
+	.double	0r-1337310.4614476819988
+tan_q_3:
+	.double	0r13827.503083158084337
+
+# ..7854
+#tan2_p_0:
+#	.double	0r3.3283674370620949801E-2
+#tan2_p_1:
+#	.double	0r2.5663601402049946582E-2
+#tan2_p_2:
+#	.double	0r-3.1180502523212879186E-4
+#tan2_p_3:
+#	.double	0r4.9814468378033982921E-7
+#tan2_q_0:
+#	.double	0r-1.3338299229604572149E-1
+#tan2_q_1:
+#	.double	0r3.4248261922569253959E-3
+#tan2_q_2:
+#	.double	0r-1.7860673544716858919E-5
+
+# ..3678
+#tan2_p_0:
+#	.double	0r3.3322645703933846673E-2
+#tan2_p_1:
+#	.double	0r2.5645883816593488536E-2
+#tan2_p_2:
+#	.double	0r-3.1101629629724883994E-4
+#tan2_p_3:
+#	.double	0r4.9436497805178155231E-7
+#tan2_q_0:
+#	.double	0r-1.3334402096273281768E-1
+#tan2_q_1:
+#	.double	0r3.4200990512379442383E-3
+#tan2_q_2:
+#	.double	0r-1.7781628650750523380E-5
+
+# ..3908
+tan2_p_0:
+	.double	0r3.3321283446731200795E-2
+tan2_p_1:
+	.double	0r2.5646503068216856824E-2
+tan2_p_2:
+	.double	0r-3.1104383115385493899E-4
+tan2_p_3:
+	.double	0r4.9449499280323978810E-7
+tan2_q_0:
+	.double	0r-1.3334538321993547050E-1
+tan2_q_1:
+	.double	0r3.4202642171272572627E-3
+tan2_q_2:
+	.double	0r-1.7784380596478880932E-5
+
+# .. .2
+tan3_p_0:
+	.double	0r-18224747.034562580287
+tan3_p_1:
+	.double	0r1170032.8518652219791
+tan3_p_2:
+	.double	0r-13254.922404240271135
+tan3_q_0:
+	.double	0r-54674241.103687740862
+tan3_q_1:
+	.double	0r25379794.997070763260
+tan3_q_2:
+	.double	0r-1339662.7778249811381
+tan3_q_3:
+	.double	0r13848.512058865804647
+
+tan_0_5_t:
+	.double	0r1.25
+	.double	0r0.5
+real_0_5_m_1_25_atan_0_5_49_l:
+	.double	0r-1.8102913770216112547E-16
+real_0_5_m_1_25_atan_0_5_49:
+	.double	0r-7.9559511251007464239E-2
+
+tan_n_0_5_t:
+	.double	0r1.25
+	.double	0r-0.5
+real_1_25_atan_0_5_m_0_5_49_l:
+	.double	0r1.8102913770216112547E-16
+real_1_25_atan_0_5_m_0_5_49:
+	.double	0r7.9559511251007464239E-2
+
+tan_s_0_5_t:
+	.double	0r1.5
+sqrt_0_5:
+	.double	0r7.07106781186547572737E-1
+sqrtn_0_5_m_1_5_atann_sqrt_0_5_l:
+	.double	0r-3.4791817252377832918E-18
+sqrtn_0_5_m_1_5_atann_sqrt_0_5:
+	.double	0r-2.1611278181903348372E-1
+
+tan_n_s_0_5_t:
+	.double	0r1.5
+n_sqrt_0_5:
+	.double	0r-7.07106781186547572737E-1
+n_sqrtn_0_5_m_1_5_atann_sqrt_0_5_l:
+	.double	0r3.4791817252377832918E-18
+n_sqrtn_0_5_m_1_5_atann_sqrt_0_5:
+	.double	0r2.1611278181903348372E-1
+
+tan_s_2_t:
+	.double	0r3.0
+sqrt_2_0:
+	.double	0r1.4142135623730951455
+sqrt_2_m_3_atan_sqrt_2_l:
+	.double	0r7.4004687159376949604E-17
+sqrt_2_m_3_atan_sqrt_2:
+	.double	0r-1.4517362920004328597
+
+tan_n_s_2_t:
+	.double	0r3.0
+n_sqrt_2_0:
+	.double	0r-1.4142135623730951455
+n_sqrt_2_m_3_atan_sqrt_2_l:
+	.double	0r-7.4004687159376949604E-17
+n_sqrt_2_m_3_atan_sqrt_2:
+	.double	0r1.4517362920004328597
+
+# .. 0.55
+#asin_p_0:
+#	.double	0r-25.316696852169688014
+#asin_p_1:
+#	.double	0r53.530200010422156254
+#asin_p_2:
+#	.double	0r-37.683203270692622766
+#asin_p_3:
+#	.double	0r9.8280386103977175338
+#asin_p_4:
+#	.double	0r-0.6933246588625101570
+#asin_q_0:
+#	.double	0r-151.90018111301861836
+#asin_q_1:
+#	.double	0r389.53628156345331490
+#asin_q_2:
+#	.double	0r-360.70299781814912876
+#asin_q_3:
+#	.double	0r144.63464250990605819
+#asin_q_4:
+#	.double	0r-23.247871426450686272
+
+asin_p_0:
+	.double	0r-31.133659782158286333
+asin_p_1:
+	.double	0r63.858157457633225818
+asin_p_2:
+	.double	0r-43.244454939408448979
+asin_p_3:
+	.double	0r10.711904434851501122
+asin_p_4:
+	.double	0r-0.7023102565817557208
+asin_q_0:
+	.double	0r-186.80195869294971089
+asin_q_1:
+	.double	0r467.20982615762636669
+asin_q_2:
+	.double	0r-419.67491247177241576
+asin_q_3:
+	.double	0r162.03208845920067915
+asin_q_4:
+	.double	0r-24.808857438965841169
+
+asin_c:
+
+asin2_p_0:
+	.double	0r31.367022265271895520
+asin2_p_1:
+	.double	0r-27.553457288163915706
+asin2_p_2:
+	.double	0r7.3142287109234631615
+asin2_p_3:
+	.double	0r-0.5695539268055010984
+asin2_p_4:
+	.double	0r2.7732700457810590539E-3
+asin2_q_0:
+	.double	0r376.40426718326244782
+asin2_q_1:
+	.double	0r-415.33244757417799065
+asin2_q_2:
+	.double	0r156.01490234281396852
+asin2_q_3:
+	.double	0r-22.702479601946876642
+
+// 0.55 ..
+#asin2_p_0:
+#	.double	0r29.355358314463529723
+#asin2_p_1:
+#	.double	0r-26.147018538045006864
+#asin2_p_2:
+#	.double	0r7.0703500716163327766
+#asin2_p_3:
+#	.double	0r-0.5652641176899680930
+#asin2_p_4:
+#	.double	0r2.9079909341710704143E-3
+#asin2_q_0:
+#	.double	0r352.26429977355559231
+#asin2_q_1:
+#	.double	0r-393.02368990519767067
+#asin2_q_2:
+#	.double	0r149.68540386246363028
+#asin2_q_3:
+#	.double	0r-22.170690309691252651
+
+asin_p1_c:
+asin_p1_0h:
+	.double	0r5.7043710961221538724E-1
+asin_p1_0l:
+	.double	0r4.9248428721475757358E-23
+asin_p1_1h:
+	.double	0r1.1881211400032043457
+asin_p1_1l:
+	.double	0r1.4630145983574038837E-9
+asin_p1_2:
+	.double	0r4.5284120326143617596E-1
+asin_p1_3:
+	.double	0r6.2472394124586794195E-1
+asin_p1_4:
+	.double	0r8.0835086833083236169E-1
+asin_p1_5:
+	.double	0r1.2595104015172415846
+asin_p1_6:
+	.double	0r2.0488537552293428767
+asin_p1_7:
+	.double	0r3.5089140995443472981
+asin_p1_8:
+	.double	0r6.2851334414813333495
+asin_p1_9:
+	.double	0r10.128040933414240499
+asin_p1_10:
+	.double	0r33.046781180882994988
+asin_p1_11:
+	.double	0r-52.817309641779630169
+asin_p1_12:
+	.double	0r540.99494117159986217
+asin_p1_13:
+	.double	0r-1326.1434641099594955
+asin_p1_14:
+	.double	0r2612.6799241572512074
+
+asin_p2_c:
+asin_p2_0h:
+	.double	0r5.7043710961221538724E-1
+asin_p2_0l:
+	.double	0r4.9248428721475757358E-23
+asin_p2_1h:
+	.double	0r1.1881211400032043457
+asin_p2_1l:
+	.double	0r1.4630165967588482090E-9
+asin_p2_2:
+	.double	0r4.5284120325997734291E-1
+asin_p2_3:
+	.double	0r6.2472394155362398482E-1
+asin_p2_4:
+	.double	0r8.0835082941011759949E-1
+asin_p2_5:
+	.double	0r1.2595128937378745615
+asin_p2_6:
+	.double	0r2.0487338667125807312
+asin_p2_7:
+	.double	0r3.5123184925937764866
+asin_p2_8:
+	.double	0r6.2076105483454178113
+asin_p2_9:
+	.double	0r11.227381892153145415
+asin_p2_10:
+	.double	0r20.429661082626292767
+asin_p2_11:
+	.double	0r35.926476241524127886
+asin_p2_12:
+	.double	0r55.782391042293575367
+asin_p2_13:
+	.double	0r64.653236450484357078
+asin_p2_14:
+	.double	0r39.681910714502151905
+
+asin_p3_c:
+asin_p3_0h:
+	.double	0r-5.7043710961221538724E-1
+asin_p3_0l:
+	.double	0r-4.9248428721475757358E-23
+asin_p3_1h:
+	.double	0r1.1881211400032043457
+asin_p3_1l:
+	.double	0r1.4630145983574038837E-9
+asin_p3_2:
+	.double	0r-4.5284120326143617596E-1
+asin_p3_3:
+	.double	0r6.2472394124586794195E-1
+asin_p3_4:
+	.double	0r-8.0835086833083236169E-1
+asin_p3_5:
+	.double	0r1.2595104015172415846
+asin_p3_6:
+	.double	0r-2.0488537552293428767
+asin_p3_7:
+	.double	0r3.5089140995443472981
+asin_p3_8:
+	.double	0r-6.2851334414813333495
+asin_p3_9:
+	.double	0r10.128040933414240499
+asin_p3_10:
+	.double	0r-33.046781180882994988
+asin_p3_11:
+	.double	0r-52.817309641779630169
+asin_p3_12:
+	.double	0r-540.99494117159986217
+asin_p3_13:
+	.double	0r-1326.1434641099594955
+asin_p3_14:
+	.double	0r-2612.6799241572512074
+
+asin_p4_c:
+asin_p4_0h:
+	.double	0r-5.7043710961221538724E-1
+asin_p4_0l:
+	.double	0r-4.9248428721475757358E-23
+asin_p4_1h:
+	.double	0r1.1881211400032043457
+asin_p4_1l:
+	.double	0r1.4630165967588482090E-9
+asin_p4_2:
+	.double	0r-4.5284120325997734291E-1
+asin_p4_3:
+	.double	0r6.2472394155362398482E-1
+asin_p4_4:
+	.double	0r-8.0835082941011759949E-1
+asin_p4_5:
+	.double	0r1.2595128937378745615
+asin_p4_6:
+	.double	0r-2.0487338667125807312
+asin_p4_7:
+	.double	0r3.5123184925937764866
+asin_p4_8:
+	.double	0r-6.2076105483454178113
+asin_p4_9:
+	.double	0r11.227381892153145415
+asin_p4_10:
+	.double	0r-20.429661082626292767
+asin_p4_11:
+	.double	0r35.926476241524127886
+asin_p4_12:
+	.double	0r-55.782391042293575367
+asin_p4_13:
+	.double	0r64.653236450484357078
+asin_p4_14:
+	.double	0r-39.681910714502151905
+
+acos_p_0:
+	.double	0r-23.490891740111308650
+acos_p_1:
+	.double	0r50.216364730997767651
+acos_p_2:
+	.double	0r-35.846130541173899076
+acos_p_3:
+	.double	0r9.5237685361261430472
+acos_p_4:
+	.double	0r-0.68995264971619463257
+acos_q_0:
+	.double	0r-140.94535044067245622
+acos_q_1:
+	.double	0r364.72359608470605963
+acos_q_2:
+	.double	0r-341.44918263146718118
+acos_q_3:
+	.double	0r138.79408611688072028
+acos_q_4:
+	.double	0r-22.704016182082593645
+
+acos2_p_0:
+	.double	0r30.187145498002834643
+acos2_p_1:
+	.double	0r-26.731562531992352660
+acos2_p_2:
+	.double	0r7.1724615825161839311
+acos2_p_3:
+	.double	0r-0.56708755287485912433
+acos2_p_4:
+	.double	0r2.8497262016782816223E-3
+acos2_q_0:
+	.double	0r362.24574597603304937
+acos2_q_1:
+	.double	0r-402.28404322843209684
+acos2_q_2:
+	.double	0r152.32592108178161538
+acos2_q_3:
+	.double	0r-22.394016829865698526
+
+acos_p1_c:
+acos_p1_0h:
+	.double	0r8.6321189001609244240E-1
+acos_p1_0l:
+	.double	0r-2.7467628254551864260E-23
+acos_p1_1h:
+	.double	0r-1.3159033656120300293
+acos_p1_1l:
+	.double	0r-2.4367664774871400368E-8
+acos_p1_2:
+	.double	0r-7.4055169144604560127E-1
+acos_p1_3:
+	.double	0r-1.2132914821382745352
+acos_p1_4:
+	.double	0r-2.1344587603015683364
+acos_p1_5:
+	.double	0r-4.3088049195660635959
+acos_p1_6:
+	.double	0r-9.2462207583022237856
+acos_p1_7:
+	.double	0r-20.778559358953092584
+acos_p1_8:
+	.double	0r-48.776431616826357640
+acos_p1_9:
+	.double	0r-105.97079800530472937
+acos_p1_10:
+	.double	0r-413.00216737064670269
+acos_p1_11:
+	.double	0r662.354234414432198719
+acos_p1_12:
+	.double	0r-11016.701631003812508
+acos_p1_13:
+	.double	0r35318.9338646810210776
+acos_p1_14:
+	.double	0r-97099.956120653485414
+
+acos_p2_c:
+acos_p2_0h:
+	.double	0r8.6321189001609244240E-1
+acos_p2_0l:
+	.double	0r-2.7467628254551864260E-23
+acos_p2_1h:
+	.double	0r-1.3159033656120300293
+acos_p2_1l:
+	.double	0r-2.4367664108737585593E-8
+acos_p2_2:
+	.double	0r-7.4055169144296173478E-1
+acos_p2_3:
+	.double	0r-1.2132914822347449224
+acos_p2_4:
+	.double	0r-2.1344586500282272112
+acos_p2_5:
+	.double	0r-4.3088086958273930094
+acos_p2_6:
+	.double	0r-9.2457168210006930309
+acos_p2_7:
+	.double	0r-20.790839559740305731
+acos_p2_8:
+	.double	0r-48.251128736941033992
+acos_p2_9:
+	.double	0r-114.02209884122218853
+acos_p2_10:
+	.double	0r-267.05630398433208939
+acos_p2_11:
+	.double	0r-584.65850073182002689
+acos_p2_12:
+	.double	0r-1076.6291620946751664
+acos_p2_13:
+	.double	0r-1407.9468675552657260
+acos_p2_14:
+	.double	0r-936.98633843723951031
+
+atan4_p_c:
+atan4_p_0:
+	.double	0r-51.285593469956815227
+atan4_p_1:
+	.double	0r-100.64117630107490697
+atan4_p_2:
+	.double	0r-64.297908354955467303
+atan4_p_3:
+	.double	0r-14.727705326536764119
+atan4_p_4:
+	.double	0r-0.8688437904180191795
+atan4_q_0:
+	.double	0r153.85678040987045279
+atan4_q_1:
+	.double	0r394.23759714914689312
+atan4_q_2:
+	.double	0r363.49766317870364674
+atan4_q_3:
+	.double	0r144.60833714965690433
+atan4_q_4:
+	.double	0r23.038413151078877661
+
+n_45_d_256:
+	.double	0r0.175781250000
+n_m_45_d_256:
+	.double	0r-0.175781250000
+
+atan_p_9_0:
+	.double	0r-3.3333333333333309279E-1
+atan_p_9_1:
+	.double	0r1.9999999999984593990E-1
+atan_p_9_2:
+	.double	0r-1.4285714282482467330E-1
+atan_p_9_3:
+	.double	0r1.1111110782523182794E-1
+atan_p_9_4:
+	.double	0r-9.0908902574698829957E-2
+atan_p_9_5:
+	.double	0r7.6916554867419328900E-2
+atan_p_9_6:
+	.double	0r-6.6526935532345926605E-2
+atan_p_9_7:
+	.double	0r5.6999867343679795628E-2
+atan_p_9_8:
+	.double	0r-3.9093313704495452621E-2
+
+atan_0_5_52:
+	.double	0r4.6364760900080603800E-1
+atan_0_5_52_l:
+	.double	0r7.8209928684219507951E-17
+atan_0_5_real_0_5:
+	.double	0r0.5
+
+n_atan_0_5_52:
+	.double	0r-4.6364760900080603800E-1
+n_atan_0_5_52_l:
+	.double	0r-7.8209928684219507951E-17
+n_atan_0_5_real_0_5:
+	.double	0r-0.5
+
+atan_n_0_5_8_c:
+atan_n_0_5_8_0:
+	.double	0r-3.3333333333333192705E-1
+atan_n_0_5_8_1:
+	.double	0r1.9999999999920992089E-1
+atan_n_0_5_8_2:
+	.double	0r-1.4285714269165267210E-1
+atan_n_0_5_8_3:
+	.double	0r1.1111109349119581424E-1
+atan_n_0_5_8_4:
+	.double	0r-9.0908019792184269203E-2
+atan_n_0_5_8_5:
+	.double	0r7.6884259191928067123E-2
+atan_n_0_5_8_6:
+	.double	0r-6.5833773159939656949E-2
+atan_n_0_5_8_7:
+	.double	0r4.8945313616423499758E-2
+
+atan_1_53:
+	.double	0r7.8539816339744827900E-1
+atan_1_53_l:
+	.double	0r3.0616169978683830179E-17
+
+n_atan_1_53:
+	.double	0r-7.8539816339744827900E-1
+n_atan_1_53_l:
+	.double	0r-3.0616169978683830179E-17
+
+atan_n_8_c:
+atan_n_8_0:
+	.double	0r-3.3333333333333325932E-1
+atan_n_8_1:
+	.double	0r1.9999999999993750555E-1
+atan_n_8_2:
+	.double	0r-1.4285714283792697032E-1
+atan_n_8_3:
+	.double	0r1.1111110814149251225E-1
+atan_n_8_4:
+	.double	0r-9.0908830368634924168E-2
+atan_n_8_5:
+	.double	0r7.6909493951295496150E-2
+atan_n_8_6:
+	.double	0r-6.6248652556378059830E-2
+atan_n_8_7:
+	.double	0r5.1752788427847688790E-2
+atan_n_8_real_40_bits:
+	.quad	0x0ffffffffffffe000
+atan_n_8_real_abs_40_bits:
+	.quad	0x07fffffffffffe000
+atan_n_8_real_2_p_m_12:
+	.double	0r2.44140625E-4
+
+atan_p_8_c:
+atan_p_8_0:
+	.double	0r-3.3333333333333331483E-1
+atan_p_8_1:
+	.double	0r1.9999999999997875033E-1
+atan_p_8_2:
+	.double	0r-1.4285714284904510452E-1
+atan_p_8_3:
+	.double	0r1.1111110961868993219E-1
+atan_p_8_4:
+	.double	0r-9.0908937814301260527E-2
+atan_p_8_5:
+	.double	0r7.6913854250452912131E-2
+atan_p_8_6:
+	.double	0r-6.6341308897488418528E-2
+atan_p_8_7:
+	.double	0r5.2556144706942850564E-2
+atan_p_8_real_39_bits:
+	.quad	0x0ffffffffffffc000
+atan_p_8_abs_real_39_bits:
+	.quad	0x07fffffffffffc000
+atan_p_8_real_2_p_m_12:
+	.double	0r2.44140625E-4
+
+atan_2_0:
+	.double	0r1.1071487177940904090
+atan_2_0_l:
+	.double	0r9.4044713735663794124E-17
+atan_2_0_real_2_0:
+	.double	0r2.0
+
+n_atan_2_0:
+	.double	0r-1.1071487177940904090
+n_atan_2_0_l:
+	.double	0r-9.4044713735663794124E-17
+n_atan_2_0_real_2_0:
+	.double	0r-2.0
+
+atan_n_2_8_c:
+atan_n_2_8_0:
+	.double	0r-3.3333333333333042825E-1
+atan_n_2_8_1:
+	.double	0r1.9999999999849707444E-1
+atan_n_2_8_2:
+	.double	0r-1.4285714256949089140E-1
+atan_n_2_8_3:
+	.double	0r1.1111108323968375833E-1
+atan_n_2_8_4:
+	.double	0r-9.0907550811429177484E-2
+atan_n_2_8_5:
+	.double	0r7.6872320463257509671E-2
+atan_n_2_8_6:
+	.double	0r-6.5674898157514843833E-2
+atan_n_2_8_7:
+	.double	0r4.8083662211730243685E-2
+
+real_atan_large:
+	.double	0r5805358775541310.0840
+real_atan_m_large:
+	.double	0r-5805358775541310.0840
+
+exp_p0_c:
+exp_p0_2:
+	.double	0r4.9999999999999900080E-1
+exp_p0_3:
+	.double	0r1.6666666666678625619E-1
+exp_p0_4:
+	.double	0r4.1666666661027668139E-2
+exp_p0_5:
+	.double	0r8.3333334724719175524E-3
+exp_p0_6:
+	.double	0r1.3888868951434380893E-3
+exp_p0_7:
+	.double	0r1.9842988288729944998E-4
+exp_p0_8:
+	.double	0r2.4713876429938641863E-5
+exp_p0_9:
+	.double	0r2.9988651596576767477E-6
+
+exp_p1_c:
+exp_p1_0:
+	.double	0r1.1921772413532716328
+exp_p1_0_l:
+	.double	0r-5.3111925243256501606E-17
+exp_p1_1_h:
+	.double	0r1.1921772360801696777
+exp_p1_1_l:
+	.double	0r5.2731019550833480025E-9
+exp_p1_2:
+	.double	0r5.9608862067662515827E-1
+exp_p1_3:
+	.double	0r1.9869620689294639759E-1
+exp_p1_4:
+	.double	0r4.9674051698174313818E-2
+exp_p1_5:
+	.double	0r9.9348108199627475218E-3
+exp_p1_6:
+	.double	0r1.6557962681130399766E-3
+exp_p1_7:
+	.double	0r2.3658136938083061391E-4
+exp_p1_8:
+	.double	0r2.9408132193586159414E-5
+exp_p1_9:
+	.double	0r3.6463907908503166658E-6
+
+exp_m0_c:
+exp_m0_2:
+	.double	0r4.9999999999999916733E-1
+exp_m0_3:
+	.double	0r1.6666666666656548834E-1
+exp_m0_4:
+	.double	0r4.1666666661841822439E-2
+exp_m0_5:
+	.double	0r8.3333332132295139666E-3
+exp_m0_6:
+	.double	0r1.3888871537291139647E-3
+exp_m0_7:
+	.double	0r1.9839759941175428605E-4
+exp_m0_8:
+	.double	0r2.4723414228338120148E-5
+exp_m0_9:
+	.double	0r2.5328445339119633100E-6
+
+exp_m1_c:
+exp_m1_0:
+	.double	0r8.3880145108698256351E-1
+exp_m1_0_l:
+	.double	0r-3.1258096785010882408E-17
+exp_m1_1_h:
+	.double	0r8.3880144357681274414E-1
+exp_m1_1_l:
+	.double	0r7.5101698193691390770E-9
+exp_m1_2:
+	.double	0r4.1940072554349638878E-1
+exp_m1_3:
+	.double	0r1.3980024184811054577E-1
+exp_m1_4:
+	.double	0r3.4950060469065312441E-2
+exp_m1_5:
+	.double	0r6.9900121811488523035E-3
+exp_m1_6:
+	.double	0r1.1650024396098284090E-3
+exp_m1_7:
+	.double	0r1.6642701498218752582E-4
+exp_m1_8:
+	.double	0r2.0771521593253968396E-5
+exp_m1_9:
+	.double	0r2.1676257763527338661E-6
+
+# ln < 1
+
+#ln_s_c:
+#ln_s_p_0:
+#	.double	0r9.1503481690678860616
+#ln_s_p_1:
+#	.double	0r20.941485866336655874
+#ln_s_p_2:
+#	.double	0r16.487417954606193859
+#ln_s_p_3:
+#	.double	0r5.1186080224958212526
+#ln_s_p_4:
+#	.double	0r4.9845734945458103349E-1
+#ln_s_q_0:
+#	.double	0r27.451044507203782530
+#ln_s_q_1:
+#	.double	0r83.412740979424398802
+#ln_s_q_2:
+#	.double	0r95.551182894484938402
+#ln_s_q_3:
+#	.double	0r50.697088912299392405
+#ln_s_q_4:
+#	.double	0r12.129116214794825268
+
+# ln .83 .. 1
+
+#ln_s_c:
+#ln_s_p_0:
+#	.double	0r10.602228952834687803
+#ln_s_p_1:
+#	.double	0r23.654147261303222649
+#ln_s_p_2:
+#	.double	0r18.048399385285609497
+#ln_s_p_3:
+#	.double	0r5.3866317747411187611
+#ln_s_p_4:
+#	.double	0r4.9866367203621170257E-1
+#ln_s_q_0:
+#	.double	0r31.806686858504065185
+#ln_s_q_1:
+#	.double	0r94.817456927787816312
+#ln_s_q_2:
+#	.double	0r106.17427873659880788
+#ln_s_q_3:
+#	.double	0r54.803473649357279385
+#ln_s_q_4:
+#	.double	0r12.671320252177885379
+
+# ln .8243 .. 1
+
+ln_s_c:
+ln_s_p_0:
+	.double	0r10.533214407052481576
+ln_s_p_1:
+	.double	0r23.526643885226203423
+ln_s_p_2:
+	.double	0r17.976077060886193948
+ln_s_p_3:
+	.double	0r5.3744490269565492468
+ln_s_p_4:
+	.double	0r4.9865516811442134326E-1
+ln_s_q_0:
+	.double	0r31.599643221157446504
+ln_s_q_1:
+	.double	0r94.279664071546818604
+ln_s_q_2:
+	.double	0r105.67819330362900132
+ln_s_q_3:
+	.double	0r54.614015226380423940
+ln_s_q_4:
+	.double	0r12.646688742839442554
+
+# ln >= 1
+
+#ln_b_c:
+#ln_b_p_0:
+#	.double	0r17.467871503051810578
+#ln_b_p_1:
+#	.double	0r35.785852544901331385
+#ln_b_p_2:
+#	.double	0r24.552006690337957906
+#ln_b_p_3:
+#	.double	0r6.4047236758045062999
+#ln_b_p_4:
+#	.double	0r4.9917096570160635061E-1
+#ln_b_q_0:
+#	.double	0r52.403614509155588053
+#ln_b_q_1:
+#	.double	0r146.66026851655587393
+#ln_b_q_2:
+#	.double	0r152.20905275348258101
+#ln_b_q_3:
+#	.double	0r71.576606726420877180
+#ln_b_q_4:
+#	.double	0r14.725978745938864023
+
+# ln 1.0 .. 1.25
+
+ln_b_c:
+ln_b_p_0:
+	.double	0r15.580409683056638315
+ln_b_p_1:
+	.double	0r32.543914666072645048
+ln_b_p_2:
+	.double	0r22.877901706403754645
+ln_b_p_3:
+	.double	0r6.1556714513765760088
+ln_b_p_4:
+	.double	0r4.9907618814155391140E-1
+ln_b_q_0:
+	.double	0r46.741229049169916721
+ln_b_q_1:
+	.double	0r132.68766578509504939
+ln_b_q_2:
+	.double	0r140.10471702854880505
+ln_b_q_3:
+	.double	0r67.303567178530471438
+ln_b_q_4:
+	.double	0r14.223951326720150945
+
+# log10 < 1
+
+#log10_s_c:
+#log10_s_p_0:
+#	.double	0r3.9766162055986344903
+#log10_s_p_1:
+#	.double	0r9.0998450829915000782
+#log10_s_p_2:
+#	.double	0r7.1633761863796170388
+#log10_s_p_3:
+#	.double	0r2.2235094671281157019
+#log10_s_p_4:
+#	.double	0r2.1647774010462134120E-1
+#log10_s_q_0:
+#	.double	0r27.469491586710010012
+#log10_s_q_1:
+#	.double	0r83.461621599998238707
+#log10_s_q_2:
+#	.double	0r95.597370915186246521
+#log10_s_q_3:
+#	.double	0r50.715260287037992271
+#log10_s_q_4:
+#	.double	0r12.131569554734877414
+
+# log10 .83 .. 1
+
+log10_s_c:
+log10_s_p_0:
+	.double	0r4.6028186381430433727
+log10_s_p_1:
+	.double	0r10.269763424759014825
+log10_s_p_2:
+	.double	0r7.8365495824134949743
+log10_s_p_3:
+	.double	0r2.3390837249823710486
+log10_s_p_4:
+	.double	0r2.1656666260147883207E-1
+log10_s_q_0:
+	.double	0r31.795144745829979627
+log10_s_q_1:
+	.double	0r94.787371070649314220
+log10_s_q_2:
+	.double	0r106.14640820241447727
+log10_s_q_3:
+	.double	0r54.792773831672434426
+log10_s_q_4:
+	.double	0r12.669919891359308295
+
+# log10 >= 1
+
+#log10_b_c:
+#log10_b_p_0:
+#	.double	0r7.5811637929542987635
+#log10_b_p_1:
+#	.double	0r15.533033113504695066
+#log10_b_p_2:
+#	.double	0r10.658434759740712749
+#log10_b_p_3:
+#	.double	0r2.7808975967872808788
+#log10_b_p_4:
+#	.double	0r2.1678697192356721768E-1
+#log10_b_q_0:
+#	.double	0r52.368824211608433927
+#log10_b_q_1:
+#	.double	0r146.57500964710882840
+#log10_b_q_2:
+#	.double	0r152.13582168617784873
+#log10_b_q_3:
+#	.double	0r71.551032626191883423
+#log10_b_q_4:
+#	.double	0r14.723015954370795910
+
+# log10 1 .. 1.25
+
+log10_b_c:
+log10_b_p_0:
+	.double	0r6.7674014091633534207
+log10_b_p_1:
+	.double	0r14.135207984866871911
+log10_b_p_2:
+	.double	0r9.9365500005391176330
+log10_b_p_3:
+	.double	0r2.6734926843772655047
+log10_b_p_4:
+	.double	0r2.1674607755999492076E-1
+log10_b_q_0:
+	.double	0r46.747552809139307328
+log10_b_q_1:
+	.double	0r132.70322218382881374
+log10_b_q_2:
+	.double	0r140.11814067349982338
+log10_b_q_3:
+	.double	0r67.308281802680511419
+log10_b_q_4:
+	.double	0r14.224501421540916013
+
+log2_b_c:
+log2_b_p_0:
+	.double	0r5.7248655744376097942
+log2_b_p_1:
+	.double	0r9.7037982066316619267
+log2_b_p_2:
+	.double	0r4.9251104233553251177
+log2_b_p_3:
+	.double	0r0.71604182547154882066
+log2_b_p_4:
+	.double	0r0.00024384785289470863262
+log2_b_q_0:
+	.double	0r11.904523296020823153
+log2_b_q_1:
+	.double	0r29.106873574779058345
+log2_b_q_2:
+	.double	0r24.928920420570964467
+log2_b_q_3:
+	.double	0r8.6737948511229703286
+
+log2_s_c:
+log2_s_p_0:
+	.double	0r3.6543174401668601092
+log2_s_p_1:
+	.double	0r6.9236821167562956347
+log2_s_p_2:
+	.double	0r4.0552858971538414679
+log2_s_p_3:
+	.double	0r0.71265600724817212974
+log2_s_p_4:
+	.double	0r0.00061552202164347354606
+log2_s_q_0:
+	.double	0r7.5989394915663739383
+log2_s_q_1:
+	.double	0r20.096596833512705871
+log2_s_q_2:
+	.double	0r18.945813884202145516
+log2_s_q_3:
+	.double	0r7.4327985038948378715
+
+exp2_p0_25_c:
+exp2_p0_25_0h:
+	.double	0r1.1892071150027210269
+exp2_p0_25_0l:
+	.double	0r3.9820152314656461110E-17
+exp2_p0_25_1h:
+	.double	0r8.2429555058479309082E-1
+exp2_p0_25_1l:
+	.double	0r8.2811696389128996998E-9
+exp2_p0_25_2:
+	.double	0r2.8567907128801478533E-1
+exp2_p0_25_3:
+	.double	0r6.6005880936045882579E-2
+exp2_p0_25_4:
+	.double	0r1.1437947568180147420E-2
+exp2_p0_25_5:
+	.double	0r1.5856362256660922264E-3
+exp2_p0_25_6:
+	.double	0r1.8317976935970924383E-4
+exp2_p0_25_7:
+	.double	0r1.8139623479154579359E-5
+exp2_p0_25_8:
+	.double	0r1.5673360359583023396E-6
+exp2_p0_25_9:
+	.double	0r1.3055968895300254492E-7
+
+exp2_p0_c:
+exp2_p0_0h:
+	.double	0r1.0
+exp2_p0_0l:
+	.double	0r0.0
+exp2_p0_1h:
+	.double	0r6.9314716756343841553E-1
+exp2_p0_1l:
+	.double	0r1.2996506870699420233E-8
+exp2_p0_2:
+	.double	0r2.4022650695910330310E-1
+exp2_p0_3:
+	.double	0r5.5504108664725972100E-2
+exp2_p0_4:
+	.double	0r9.6181291092419882865E-3
+exp2_p0_5:
+	.double	0r1.3333558019987959767E-3
+exp2_p0_6:
+	.double	0r1.5403532681856270284E-4
+exp2_p0_7:
+	.double	0r1.5253065510864060166E-5
+exp2_p0_8:
+	.double	0r1.3190300302967161470E-6
+exp2_p0_9:
+	.double	0r1.0881074348847176580E-7
+
+exp2_m0_c:
+exp2_m0_0h:
+	.double	0r1.0
+exp2_m0_0l:
+	.double	0r0.0
+exp2_m0_1h:
+	.double	0r6.9314716756343841553E-1
+exp2_m0_1l:
+	.double	0r1.2996506870699420233E-8
+exp2_m0_2:
+	.double	0r2.4022650695909780749E-1
+exp2_m0_3:
+	.double	0r5.5504108664691832742E-2
+exp2_m0_4:
+	.double	0r9.6181291047027026725E-3
+exp2_m0_5:
+	.double	0r1.3333557764085299593E-3
+exp2_m0_6:
+	.double	0r1.5403499520940549562E-4
+exp2_m0_7:
+	.double	0r1.5251161994200753343E-5
+exp2_m0_8:
+	.double	0r1.3166061357593046239E-6
+exp2_m0_9:
+	.double	0r9.2942318847941285694E-8
+
+exp2_m0_25_c:
+exp2_m0_25_0h:
+	.double	0r8.4089641525371450204E-1
+exp2_m0_25_0l:
+	.double	0r4.0995050102907482601E-17
+exp2_m0_25_1h:
+	.double	0r5.8286496996879577637E-1
+exp2_m0_25_1l:
+	.double	0r9.4072813983103742430E-9
+exp2_m0_25_2:
+	.double	0r2.0200560855082275169E-1
+exp2_m0_25_3:
+	.double	0r4.6673206007799013240E-2
+exp2_m0_25_4:
+	.double	0r8.0878502812528896282E-3
+exp2_m0_25_5:
+	.double	0r1.1212140390502213245E-3
+exp2_m0_25_6:
+	.double	0r1.2952709156943405071E-4
+exp2_m0_25_7:
+	.double	0r1.2823060087120729985E-5
+exp2_m0_25_8:
+	.double	0r1.1036110895925728151E-6
+exp2_m0_25_9:
+	.double	0r7.4921899653437894396E-8
+
+real_m_1022:
+	.double	0r-1022.0
+real_1023:
+	.double	0r1023.0
+real_1024:
+	.double	0r1024.0
+real_1025:
+	.double	0r1025.0
+real_m_1076:
+	.double	0r-1076.0
+real_2_p_m_1022:
+	.quad	0x00010000000000000
+real_2_p_m_55:
+	.quad	0x03C80000000000000
+real_2_p_53:
+	.quad	0x04340000000000000
+real_m_2_p_53:
+	.quad	0x0C340000000000000
+real_2_p_55:
+	.quad	0x04360000000000000
+real_2_p_1023:
+	.quad	0x07fe0000000000000
+real_max:
+	.quad	0x07fefffffffffffff
+real_m_max:
+	.quad	0x0ffefffffffffffff
+
+real_4000_0:
+	.double	0r4000.0
+real_m_3810_0:
+	.double	0r-3810.0
+real_1401_0:
+	.double	0r1401.0
+real_m_1471_0:
+	.double	0r-1471.0
+real_4605_0:
+	.double	0r4605.0
+real_m_4834_0:
+	.double	0r-4834.0
+real_1_052:
+	.double	0r1.052
+real_m_1_052:
+	.double	0r-1.052
+real_power_exp_too_large:
+	.double	0r6711563375777760768.0
+real_power_exp_too_small:
+	.double	0r-6393154322601327104.0
+
diff --git a/macho64/astartup.s b/macho64/astartup.s
new file mode 100644
index 0000000..8f2bc01
--- /dev/null
+++ b/macho64/astartup.s
@@ -0,0 +1,5119 @@
+
+	.set LINUX,1
+	.set USE_LIBM,0
+	.set NEW_DESCRIPTORS,1
+	.set MEASURE_GC,0
+
+	.macro att_jmp
+	.att_syntax
+	jmp	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_call
+	.att_syntax
+	call	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_je
+	.att_syntax
+	je	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jne
+	.att_syntax
+	jne	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_ja
+	.att_syntax
+	ja	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jae
+	.att_syntax
+	jae	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jb
+	.att_syntax
+	jb	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jbe
+	.att_syntax
+	jbe	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jc
+	.att_syntax
+	jc	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jnc
+	.att_syntax
+	jnc	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jg
+	.att_syntax
+	jg	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jge
+	.att_syntax
+	jge	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jl
+	.att_syntax
+	jl	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jle
+	.att_syntax
+	jle	$0
+	.intel_syntax noprefix
+	.endmacro
+
+	.macro att_jz
+	.att_syntax
+	jz	$0
+	.intel_syntax noprefix
+	.endmacro
+
+/*	File:	astartup.s */
+/*	Author:	John van Groningen */
+/*	Machine:	amd64 */
+
+/* 	r10 = r10 */
+/* 	r11 = r11 */
+/* 	r12 = r12 */
+/* 	r13 = r13 */
+
+/* 	r11d = r11d */
+/* 	r12d = r12d */
+
+/* 	r10b = r10b */
+
+ .if LINUX
+	.intel_syntax noprefix
+ .endif
+
+ .if ! LINUX
+	.globl	convert_real_to_string
+ .endif
+ .if ! LINUX
+	.globl	write_heap
+ .endif
+	.globl	_return_code
+	.globl	_execution_aborted
+	.globl	e____system__kFinalizerGCTemp
+	.globl	e____system__kFinalizer
+
+ .if LINUX
+	.globl	_times
+	.globl	_exit
+ .else
+	.globl	GetTickCount
+	.globl	ExitProcess
+ .endif
+
+ .if USE_LIBM
+	.globl	cos
+	.globl	sin
+	.globl	tan
+	.globl	atan
+ .endif
+
+	.data
+	.align	3
+
+semi_space_size:	.quad	0
+
+heap_p1:	.quad	0
+heap_p2:	.quad	0
+heap_p3:	.quad	0
+neg_heap_p3:	.quad	0
+end_heap_p3:	.quad	0
+vector_p:	.quad	0
+vector_counter:	.quad	0
+neg_heap_vector_plus_4:	.quad	0
+
+heap_size_64_65:	.quad	0
+heap_vector:	.quad	0
+stack_top:	.quad	0
+end_vector:	.quad	0
+
+heap_size_257:	.quad	0
+heap_copied_vector:	.quad	0
+
+heap_end_after_gc:	.quad	0
+extra_heap:	.quad	0
+extra_heap_size:	.quad	0
+stack_p:	.quad	0
+halt_sp:	.quad	0
+	
+n_allocated_words:	.quad	0
+
+last_time:	.quad	0
+execute_time:	.quad	0
+garbage_collect_time:	.quad	0
+IO_time:	.quad	0
+
+compact_garbage_collect_time:	.quad	0
+mark_compact_garbage_collect_time:	.quad	0
+total_gc_bytes:	.quad	0
+total_compact_gc_bytes:	.quad	0
+
+	.globl	saved_heap_p
+saved_heap_p:
+	.quad	0
+	.quad	0
+	
+	.globl	saved_a_stack_p
+saved_a_stack_p:	.quad	0
+
+	.globl	end_a_stack
+end_a_stack:	.quad	0
+
+	.globl	int_to_real_scratch
+int_to_real_scratch:	.quad	0
+
+heap_end_write_heap:	.quad	0
+d3_flag_write_heap:	.quad	0
+heap2_begin_and_end:
+		.quad	0
+		.quad	0
+
+	.globl	a_stack_guard_page
+a_stack_guard_page:	.quad	0
+
+	.globl	profile_stack_pointer
+profile_stack_pointer:	.quad	0
+
+dll_initisialised:	.quad	0
+	.globl	end_b_stack
+end_b_stack:	.quad	0
+basic_only:	.quad	0
+heap_size_65:	.quad	0
+heap_copied_vector_size:	.quad	0
+heap_end_after_copy_gc:	.quad	0
+heap_mbp:	.quad	0
+heap_p:		.quad	0
+stack_mbp:	.quad	0
+
+bit_counter:
+	.quad	0
+bit_vector_p:
+	.quad	0
+zero_bits_before_mark:
+	.quad	1
+n_free_words_after_mark:
+	.quad	1000
+n_last_heap_free_bytes:
+	.quad	0
+lazy_array_list:
+	.quad	0
+n_marked_words:
+	.quad	0
+end_stack:
+	.quad	0
+
+bit_vector_size:
+	.quad	0
+
+caf_list:
+	.quad	0
+	.globl	caf_listp
+caf_listp:
+	.quad	0
+	
+zero_length_string:
+	.quad	__STRING__+2
+	.quad	0
+true_string:
+	.quad	__STRING__+2
+	.quad	4
+true_c_string:
+	.ascii	"True"
+	.byte	0,0,0,0
+false_string:
+	.quad	__STRING__+2
+	.quad	5
+false_c_string:
+	.ascii	"False"
+	.byte	0,0,0
+file_c_string:
+	.ascii	"File"
+	.byte	0,0,0,0
+garbage_collect_flag:
+	.byte	0
+	.byte	0,0,0
+
+	.comm	sprintf_buffer,32
+
+out_of_memory_string_1:
+	.ascii	"Not enough memory to allocate heap and stack"
+	.byte	10,0
+printf_int_string:
+	.ascii	"%d"
+	.byte	0
+printf_real_string:
+	.ascii	"%.15g"
+	.byte	0
+printf_string_string:
+	.ascii	"%s"
+	.byte	0
+printf_char_string:
+	.ascii	"%c"
+	.byte	0
+garbage_collect_string_1:
+	.ascii	"A stack: "
+	.byte	0
+garbage_collect_string_2:
+	.ascii	" bytes. BC stack: "
+	.byte	0
+garbage_collect_string_3:
+	.ascii	" bytes."
+	.byte	10,0
+heap_use_after_gc_string_1:
+	.ascii	"Heap use after garbage collection: "
+	.byte	0
+heap_use_after_compact_gc_string_1:
+	.ascii	"Heap use after compacting garbage collection: "
+	.byte	0
+heap_use_after_gc_string_2:
+	.ascii	" Bytes."
+	.byte	10,0
+stack_overflow_string:
+	.ascii	"Stack overflow."
+	.byte	10,0
+out_of_memory_string_4:
+	.ascii	"Heap full."
+	.byte	10,0
+time_string_1:
+	.ascii	"Execution: "
+	.byte	0
+time_string_2:
+	.ascii	"  Garbage collection: "
+	.byte	0
+
+time_string_3:
+	.ascii	" "
+	.byte	0
+
+time_string_4:
+	.ascii	"  Total: "
+	.byte	0
+high_index_string:
+	.ascii	"Index too high in UPDATE string."
+	.byte	10,0
+low_index_string:
+	.ascii	"Index negative in UPDATE string."
+	.byte	10,0
+IO_error_string:
+	.ascii	"IO error: "
+	.byte	0
+new_line_string:
+	.byte	10,0
+	
+sprintf_time_string:
+	.ascii	"%d.%02d"
+	.byte	0
+
+marked_gc_string_1:
+	.ascii	"Marked: "
+	.byte	0
+
+ .if PROFILE
+	.align	3
+m_system:
+	.long	6
+	.ascii	"System"
+	.byte	0
+	.byte	0
+
+	.long	m_system-.
+garbage_collector_name:
+	.quad	0
+	.ascii	"garbage_collector"
+	.byte	0
+	.align	3
+ .endif
+
+	.align	4
+	.globl	sign_real_mask
+sign_real_mask:
+	.quad	0x8000000000000000,0x8000000000000000
+	.globl	abs_real_mask
+abs_real_mask:
+	.quad	0x7fffffffffffffff,0x7fffffffffffffff
+
+	.align 3
+NAN_real:
+	.long	0x0ffffffff,0x7fffffff
+one_real:
+	.long	0x00000000,0x3ff00000
+zero_real:
+	.long	0x00000000,0x00000000
+
+	.align 2 
+bit_set_table:
+	.long	0x00000001,0x00000002,0x00000004,0x00000008
+	.long	0x00000010,0x00000020,0x00000040,0x00000080
+	.long	0x00000100,0x00000200,0x00000400,0x00000800
+	.long	0x00001000,0x00002000,0x00004000,0x00008000
+	.long	0x00010000,0x00020000,0x00040000,0x00080000
+	.long	0x00100000,0x00200000,0x00400000,0x00800000
+	.long	0x01000000,0x02000000,0x04000000,0x08000000
+	.long	0x10000000,0x20000000,0x40000000,0x80000000
+	.long	0
+bit_set_table2:
+	.long	0x00000001,0,0x00000002,0,0x00000004,0,0x00000008,0
+	.long	0x00000010,0,0x00000020,0,0x00000040,0,0x00000080,0
+	.long	0x00000100,0,0x00000200,0,0x00000400,0,0x00000800,0
+	.long	0x00001000,0,0x00002000,0,0x00004000,0,0x00008000,0
+	.long	0x00010000,0,0x00020000,0,0x00040000,0,0x00080000,0
+	.long	0x00100000,0,0x00200000,0,0x00400000,0,0x00800000,0
+	.long	0x01000000,0,0x02000000,0,0x04000000,0,0x08000000,0
+	.long	0x10000000,0,0x20000000,0,0x40000000,0,0x80000000,0
+	.long	0,0
+bit_clear_table:
+	.long	0x0fffffffe,0x0fffffffd,0x0fffffffb,0x0fffffff7
+	.long	0x0ffffffef,0x0ffffffdf,0x0ffffffbf,0x0ffffff7f
+	.long	0x0fffffeff,0x0fffffdff,0x0fffffbff,0x0fffff7ff
+	.long	0x0ffffefff,0x0ffffdfff,0x0ffffbfff,0x0ffff7fff
+	.long	0x0fffeffff,0x0fffdffff,0x0fffbffff,0x0fff7ffff
+	.long	0x0ffefffff,0x0ffdfffff,0x0ffbfffff,0x0ff7fffff
+	.long	0x0feffffff,0x0fdffffff,0x0fbffffff,0x0f7ffffff
+	.long	0x0efffffff,0x0dfffffff,0x0bfffffff,0x7fffffff
+	.long	0x0ffffffff
+bit_clear_table2:
+	.long	0x0fffffffe,-1,0x0fffffffd,-1,0x0fffffffb,-1,0x0fffffff7,-1
+	.long	0x0ffffffef,-1,0x0ffffffdf,-1,0x0ffffffbf,-1,0x0ffffff7f,-1
+	.long	0x0fffffeff,-1,0x0fffffdff,-1,0x0fffffbff,-1,0x0fffff7ff,-1
+	.long	0x0ffffefff,-1,0x0ffffdfff,-1,0x0ffffbfff,-1,0x0ffff7fff,-1
+	.long	0x0fffeffff,-1,0x0fffdffff,-1,0x0fffbffff,-1,0x0fff7ffff,-1
+	.long	0x0ffefffff,-1,0x0ffdfffff,-1,0x0ffbfffff,-1,0x0ff7fffff,-1
+	.long	0x0feffffff,-1,0x0fdffffff,-1,0x0fbffffff,-1,0x0f7ffffff,-1
+	.long	0x0efffffff,-1,0x0dfffffff,-1,0x0bfffffff,-1,0x7fffffff,-1
+	.long	0x0ffffffff,-1
+first_one_bit_table:
+	.byte	-1,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	5,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	6,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	5,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	7,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	5,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	6,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	5,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+	.byte	4,0,1,0,2,0,1,0,3,0,1,0,2,0,1,0
+
+	.align	2 
+	.comm	sprintf_time_buffer,20
+
+	.align	3
+
+/*	.globl	small_integers */
+	.comm	small_integers,33*16
+/*	.globl	static_characters */
+	.comm	static_characters,256*16
+
+/*	.globl	clean_exception_handler */
+/*	.globl	clean_unwind_info */
+/*clean_unwind_info: */
+/*	DD	000000009H */
+/*	DD	imagerel(clean_exception_handler) */
+
+
+	.text
+
+	.globl	_abc_main
+	.globl	print
+	.globl	print_char
+	.globl	print_int
+	.globl	print_real
+	.globl	print__string__
+	.globl	print__chars__sc
+	.globl	print_sc
+	.globl	print_symbol
+	.globl	print_symbol_sc
+	.globl	printD
+	.globl	DtoAC
+	.globl	push_t_r_args
+	.globl	push_a_r_args
+	.globl	halt
+	.globl	dump
+
+	.globl	catAC
+	.globl	sliceAC
+	.globl	updateAC
+	.globl	eqAC
+	.globl	cmpAC
+
+	.globl	string_to_string_node
+	.globl	int_array_to_node
+	.globl	real_array_to_node
+
+	.globl	_create_arrayB
+	.globl	_create_arrayC
+	.globl	_create_arrayI
+	.globl	_create_arrayI32
+	.globl	_create_arrayR
+	.globl	_create_arrayR32
+	.globl	_create_r_array
+	.globl	create_array
+	.globl	create_arrayB
+	.globl	create_arrayC
+	.globl	create_arrayI
+	.globl	create_arrayI32
+	.globl	create_arrayR
+	.globl	create_arrayR32
+	.globl	create_R_array
+
+	.globl	BtoAC
+	.globl	ItoAC
+	.globl	RtoAC
+	.globl	eqD
+
+	.globl	collect_0
+	.globl	collect_1
+	.globl	collect_2
+	.globl	collect_3
+
+	.globl	yet_args_needed
+	.globl	yet_args_needed_0
+	.globl	yet_args_needed_1
+	.globl	yet_args_needed_2
+	.globl	yet_args_needed_3
+	.globl	yet_args_needed_4
+
+	.globl	_c3,_c4,_c5,_c6,_c7,_c8,_c9,_c10,_c11,_c12
+	.globl	_c13,_c14,_c15,_c16,_c17,_c18,_c19,_c20,_c21,_c22
+	.globl	_c23,_c24,_c25,_c26,_c27,_c28,_c29,_c30,_c31,_c32
+
+	.globl	e__system__nind
+	.globl	e__system__eaind
+/* old names of the previous two labels for compatibility, remove later */
+	.globl	__indirection,__eaind
+	.globl	e__system__dind
+	.globl	eval_fill
+
+	.globl	eval_upd_0,eval_upd_1,eval_upd_2,eval_upd_3,eval_upd_4
+	.globl	eval_upd_5,eval_upd_6,eval_upd_7,eval_upd_8,eval_upd_9
+	.globl	eval_upd_10,eval_upd_11,eval_upd_12,eval_upd_13,eval_upd_14
+	.globl	eval_upd_15,eval_upd_16,eval_upd_17,eval_upd_18,eval_upd_19
+	.globl	eval_upd_20,eval_upd_21,eval_upd_22,eval_upd_23,eval_upd_24
+	.globl	eval_upd_25,eval_upd_26,eval_upd_27,eval_upd_28,eval_upd_29
+	.globl	eval_upd_30,eval_upd_31,eval_upd_32
+
+	.globl	repl_args_b
+	.globl	push_arg_b
+	.globl	del_args
+
+	.globl	add_IO_time
+	.globl	add_execute_time
+	.globl	_IO_error
+	.globl	stack_overflow
+
+	.globl	out_of_memory_4
+	.globl	print_error
+
+ .if LINUX
+	.globl	__start
+ .else
+	.globl	_start
+ .endif
+
+ .if PROFILE
+/*	.globl	init_profiler */
+/*	.globl	profile_n */
+/*	.globl	profile_s */
+/*	.globl	profile_r */
+/*	.globl	write_profile_information */
+/*	.globl	write_profile_stack */
+ .endif
+
+  .if USE_LIBM
+	.globl	cos_real
+	.globl	sin_real
+	.globl	tan_real
+	.globl	asin_real
+	.globl	acos_real
+	.globl	atan_real
+	.globl	ln_real
+	.globl	log10_real
+	.globl	exp_real
+	.globl	pow_real
+ .endif
+	.globl	entier_real
+	.globl	r_to_i_real
+  .if USE_LIBM
+	.globl	_c_pow
+	.globl	_c_log10
+	.globl	_c_entier
+ .endif
+
+	.globl	__driver
+
+/* from system.abc: */
+	.globl	dINT
+	.globl	INT32
+	.globl	CHAR
+	.globl	BOOL
+	.globl	REAL
+	.globl	REAL32
+	.globl	FILE
+	.globl	__STRING__
+	.globl	__ARRAY__
+	.globl	__cycle__in__spine
+	.globl	__print__graph
+	.globl	__eval__to__nf
+
+/* from wcon.c: */
+	.globl	_w_print_char
+	.globl	_w_print_string
+	.globl	_w_print_text
+	.globl	_w_print_int
+	.globl	_w_print_real
+
+	.globl	_ew_print_char
+	.globl	_ew_print_text
+	.globl	_ew_print_string
+	.globl	_ew_print_int
+
+	.globl	_ew_print_real
+
+	.globl	_ab_stack_size
+	.globl	_heap_size
+	.globl	_flags
+
+/* from standard c library: */
+
+ .if ! LINUX
+	.globl	allocate_memory
+	.globl	allocate_memory_with_guard_page_at_end
+	.globl	free_memory
+ .endif
+ 
+	.globl	_heap_size_multiple
+	.globl	_initial_heap_size
+
+	.globl	_min_write_heap_size
+
+	.globl	__Nil
+/*	.globl	finalizer_list */
+	.comm	finalizer_list,8
+/*	.globl	free_finalizer_list */
+	.comm	free_finalizer_list,8
+
+_abc_main:
+	push	rbx
+	push	rcx
+	push	rdx 
+	push	rbp 
+	push	rsi 
+	push	rdi 
+
+	call	init_clean
+	test	rax,rax
+	jne	init_error
+
+	call	init_timer
+
+	mov	halt_sp[rip],rsp
+
+ .if PROFILE
+	call	init_profiler
+ .endif
+
+ .if LINUX
+	att_call	__start
+
+exit_:
+ .else
+	call	_start
+
+exit:
+ .endif
+
+	call	exit_clean
+
+init_error:
+	pop	rdi
+	pop	rsi
+	pop	rbp
+	pop	rdx
+	pop	rcx
+	pop	rbx
+
+ .if LINUX
+	mov	eax,dword ptr _return_code[rip]
+	jne	return_code_set_1
+	mov	eax,-1
+return_code_set_1:
+ .endif
+	ret
+
+
+	.globl	DllMain
+DllMain:
+	cmp	edx,1
+	je	DLL_PROCESS_ATTACH
+	jb	DLL_PROCESS_DETACH
+	ret
+
+DLL_PROCESS_ATTACH:
+	push	rbx 
+	push	rbp 
+	push	rsi 
+	push	rdi
+ .if ! LINUX
+	.byte	0x49
+	push	rsp
+	.byte	0x49
+	push	rbp
+	.byte	0x49
+	push	rsi
+	.byte	0x49
+	push	rdi
+ .else
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ .endif
+	mov	qword ptr dll_initisialised[rip],1
+
+	att_call	init_clean
+	test	rax,rax
+	jne	init_dll_error
+
+	att_call	init_timer
+
+	mov	halt_sp[rip],rsp 
+
+ .if PROFILE
+	att_call	init_profiler
+ .endif
+
+	mov	qword ptr saved_heap_p[rip],rdi
+	mov	qword ptr saved_heap_p+8[rip],r15
+	mov	saved_a_stack_p[rip],rsi
+
+	mov	rax,1
+	jmp	exit_dll_init
+
+init_dll_error:
+	xor	rax,rax
+	att_jmp	exit_dll_init
+	
+DLL_PROCESS_DETACH:
+	push	rbx 
+	push	rbp 
+	push	rsi 
+	push	rdi
+ .if ! LINUX
+	.byte	0x49
+	push	rsp
+	.byte	0x49
+	push	rbp
+	.byte	0x49
+	push	rsi
+	.byte	0x49
+	push	rdi
+ .else
+	push	r12
+	push	r13
+	push	r14
+	push	r15
+ .endif
+	
+	mov	rdi,qword ptr saved_heap_p[rip]
+	mov	r15,qword ptr saved_heap_p+8[rip]
+	mov	rsi,saved_a_stack_p[rip]
+
+	att_call	exit_clean
+
+exit_dll_init:
+ .if ! LINUX
+	.byte	0x49
+	pop	rdi
+	.byte	0x49
+	pop	rsi
+	.byte	0x49
+	pop	rbp
+	.byte	0x49
+	pop	rsp
+ .else
+	pop	r15
+	pop	r14
+	pop	r13
+	pop	r12
+ .endif
+	pop	rdi 
+	pop	rsi 
+	pop	rbp 
+	pop	rbx 
+	ret
+
+init_clean:
+	lea	rax,128[rsp]
+	sub	rsp,32+8
+
+	sub	rax,qword ptr _ab_stack_size[rip]
+	mov	end_b_stack[rip],rax 
+
+	mov	rax,qword ptr _flags[rip]
+	and	rax,1
+	mov	basic_only[rip],rax 
+
+/*	call	allow_prefetch_for_athlon */
+
+	mov	rax,qword ptr _heap_size[rip]
+	sub	rax,7
+	xor	rdx,rdx 
+	mov	rbx,65
+	div	rbx
+	mov	qword ptr heap_size_65[rip],rax 
+
+	mov	rax,qword ptr _heap_size[rip]
+	sub	rax,7
+	xor	rdx,rdx 
+	mov	rbx,257
+	div	rbx
+	mov	heap_size_257[rip],rax
+	add	rax,7
+	and	rax,-8
+	mov	qword ptr heap_copied_vector_size[rip],rax
+	mov	qword ptr heap_end_after_copy_gc[rip],0
+
+	mov	rax,qword ptr _heap_size[rip]
+	add	rax,7
+	and	rax,-8
+	mov	qword ptr _heap_size[rip],rax 
+	add	rax,7
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	rdi,rax
+	call	_malloc
+ .else
+	mov	rcx,rax
+	call	allocate_memory
+ .endif
+	mov	rsp,rbp
+
+	test	rax,rax 
+	je	no_memory_2
+
+	mov	heap_mbp[rip],rax
+	lea	rdi,7[rax]
+	and	rdi,-8
+	mov	heap_p[rip],rdi 
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r14,rdi
+	mov	rdi,qword ptr _ab_stack_size[rip]
+	add	rdi,7
+	att_call	_malloc
+	mov	rdi,r14
+ .else
+	mov	rcx,qword ptr _ab_stack_size
+	add	rcx,7
+	call	allocate_memory_with_guard_page_at_end
+ .endif
+	mov	rsp,rbp
+	
+	test	rax,rax 
+	je	no_memory_3
+	
+	mov	stack_mbp[rip],rax 
+
+	add	rax,qword ptr _ab_stack_size[rip]
+	add	rax,7+4095
+	and	rax,-4096
+	mov	qword ptr a_stack_guard_page[rip],rax 
+	sub	rax,qword ptr _ab_stack_size[rip]
+
+	add	rax,7
+	and	rax,-8
+
+	mov	rsi,rax 
+	mov	stack_p[rip],rax 
+
+	add	rax,qword ptr _ab_stack_size[rip]
+	sub	rax,64
+	mov	qword ptr end_a_stack[rip],rax 
+
+	lea	rcx,small_integers[rip]
+	xor	rax,rax 
+	lea	rbx,(dINT+2)[rip]
+
+make_small_integers_lp:
+	mov	[rcx],rbx 
+	mov	8[rcx],rax 
+	inc	rax 
+	add	rcx,16
+	cmp	rax,33
+	att_jne	make_small_integers_lp
+
+	lea	rcx,static_characters[rip]
+	xor	rax,rax 
+	lea	rbx,(CHAR+2)[rip]
+
+make_static_characters_lp:
+	mov	[rcx],rbx 
+	mov	8[rcx],rax 
+	inc	rax 
+	add	rcx,16
+	cmp	rax,256
+	att_jne	make_static_characters_lp
+
+	lea	rcx,(caf_list+8[rip])
+	mov	qword ptr caf_listp[rip],rcx 
+
+	lea	rcx,__Nil-8[rip]
+	mov	qword ptr finalizer_list[rip],rcx
+	mov	qword ptr free_finalizer_list[rip],rcx
+
+	mov	heap_p1[rip],rdi
+
+	mov	rbp,qword ptr heap_size_257[rip]
+	shl	rbp,4
+	lea	rax,[rdi+rbp*8]
+	mov	heap_copied_vector[rip],rax 
+	add	rax,heap_copied_vector_size[rip]
+	mov	heap_p2[rip],rax
+
+	mov	byte ptr garbage_collect_flag[rip],0
+
+	test	byte ptr _flags[rip],64
+	je	no_mark1
+
+	mov	rax,qword ptr heap_size_65[rip]
+	mov	qword ptr heap_vector[rip],rdi 
+	add	rdi,rax
+
+	add	rdi,7
+	and	rdi,-8
+
+	mov	qword ptr heap_p3[rip],rdi
+	lea	rbp,[rax*8]
+	mov	byte ptr garbage_collect_flag[rip],-1
+
+no_mark1:
+	mov	rax,qword ptr _initial_heap_size[rip]
+
+	mov	rbx,4000
+	test	byte ptr _flags[rip],64
+	jne	no_mark9
+	add	rbx,rbx 
+no_mark9:
+
+	cmp	rax,rbx 
+	jle	too_large_or_too_small
+	shr	rax,3
+	cmp	rax,rbp 
+	att_jge	too_large_or_too_small
+	mov	rbp,rax 
+too_large_or_too_small:
+
+	lea	rax,[rdi+rbp*8]
+	mov	heap_end_after_gc[rip],rax 
+
+	test	byte ptr _flags[rip],64
+	att_je	no_mark2
+	mov	qword ptr bit_vector_size[rip],rbp
+no_mark2:
+
+	mov	r15,rbp
+
+	add	rsp,32+8
+	xor	rax,rax
+	ret
+
+no_memory_2:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	lea	rdi,out_of_memory_string_1[rip]
+ .else
+	lea	rcx,out_of_memory_string_1
+ .endif
+	att_call	_ew_print_string
+	mov	rsp,rbp
+
+	mov	qword ptr _execution_aborted[rip],1
+
+	add	rsp,32
+	mov	rax,1
+	ret
+
+no_memory_3:
+	mov	rbp,rsp
+	and	rsp,-16
+
+ .if LINUX
+	lea	rdi,out_of_memory_string_1[rip]
+ .else
+	lea	ecx,out_of_memory_string_1
+ .endif
+	att_call	_ew_print_string
+
+	mov	qword ptr _execution_aborted[rip],1
+ 
+ .if LINUX
+	mov	rdi,heap_mbp[rip]
+	att_call	_free
+ .else
+	mov	rcx,heap_mbp
+	call	free_memory
+ .endif
+
+	mov	rsp,rbp
+
+	add	rsp,32
+	mov	rax,1
+	ret
+
+exit_clean:
+	att_call	add_execute_time
+
+	mov	rax,qword ptr _flags[rip]
+	test	al,8
+	je	no_print_execution_time
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if ! LINUX
+	sub	rsp,32
+ .endif
+
+ .if LINUX
+	lea	rdi,time_string_1[rip]
+ .else
+	lea	rcx,time_string_1
+ .endif
+	att_call	_ew_print_string
+	
+	mov	rax,execute_time[rip]
+	call	print_time
+	
+ .if LINUX
+	lea	rdi,time_string_2[rip]
+ .else
+	lea	rcx,time_string_2
+ .endif
+	att_call	_ew_print_string
+
+	mov	rax,garbage_collect_time[rip]
+ .if MEASURE_GC
+ .else
+	add	rax,mark_compact_garbage_collect_time[rip]
+	add	rax,compact_garbage_collect_time[rip]
+ .endif
+	att_call	print_time
+
+ .if MEASURE_GC
+
+  .if LINUX
+	lea	rdi,time_string_3
+  .else
+	lea	rcx,time_string_3
+  .endif
+	call	_ew_print_string
+
+	mov	rax,mark_compact_garbage_collect_time
+	call	print_time
+
+  .if LINUX
+	lea	rdi,time_string_3
+  .else
+	lea	rcx,time_string_3
+  .endif
+	call	_ew_print_string
+
+	mov	rax,compact_garbage_collect_time
+	call	print_time
+ 
+ .endif
+
+ .if LINUX
+	lea	rdi,time_string_4[rip]
+ .else
+	lea	rcx,time_string_4
+ .endif
+	att_call	_ew_print_string
+
+	mov	rax,execute_time[rip]
+	add	rax,garbage_collect_time[rip]
+	add	rax,IO_time[rip]
+
+	add	rax,mark_compact_garbage_collect_time[rip]
+	add	rax,compact_garbage_collect_time[rip]
+
+	att_call	print_time
+
+ .if LINUX
+	mov	rdi,10
+ .else
+	mov	rcx,10
+ .endif
+	att_call	_ew_print_char
+
+ .if MEASURE_GC
+
+  .if LINUX
+	mov	rdi,total_gc_bytes
+  .else
+	mov	rcx,total_gc_bytes
+  .endif
+	call	_ew_print_int
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+  .if LINUX
+	mov rdi,total_compact_gc_bytes
+  .else
+	mov rcx,total_compact_gc_bytes
+  .endif
+	call	_ew_print_int
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm0,qword ptr garbage_collect_time
+	divsd	xmm0,xmm1
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm0,qword ptr mark_compact_garbage_collect_time
+	divsd	xmm0,xmm1
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm0,qword ptr compact_garbage_collect_time
+	divsd	xmm0,xmm1
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,10
+  .else
+	mov	rcx,10
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm2,qword ptr garbage_collect_time
+	divsd	xmm2,xmm1
+	mov	rax,qword ptr total_gc_bytes
+	cvtsi2sd	xmm0,rax
+	divsd	xmm0,xmm2
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm2,qword ptr mark_compact_garbage_collect_time
+	divsd	xmm2,xmm1
+	mov	rax,qword ptr total_compact_gc_bytes
+	cvtsi2sd	xmm0,rax
+	divsd	xmm0,xmm2
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm2,qword ptr compact_garbage_collect_time
+	divsd	xmm2,xmm1
+	mov	rax,qword ptr total_compact_gc_bytes
+	cvtsi2sd	xmm0,rax
+	divsd	xmm0,xmm2
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	call	_ew_print_char
+
+	mov	rax,1000
+	cvtsi2sd	xmm1,rax
+	cvtsi2sd	xmm2,qword ptr mark_compact_garbage_collect_time
+	cvtsi2sd	xmm3,qword ptr compact_garbage_collect_time
+	addsd	xmm2,xmm3
+	divsd	xmm2,xmm1
+	mov	rax,qword ptr total_compact_gc_bytes
+	cvtsi2sd	xmm0,rax
+	divsd	xmm0,xmm2
+	call	_ew_print_real
+
+  .if LINUX
+	mov	rdi,10
+  .else
+	mov	rcx,10
+  .endif
+	call	_ew_print_char
+ 
+ .endif
+
+	mov	rsp,rbp
+
+no_print_execution_time:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	rdi,stack_mbp[rip]
+	att_call	_free
+
+	mov	rdi,heap_mbp[rip]
+	att_call	_free
+ .else
+	mov	rcx,stack_mbp
+	sub	rsp,32
+	call	free_memory
+
+	mov	rcx,heap_mbp
+	call	free_memory
+	add	rsp,32
+ .endif
+	mov	rsp,rbp
+
+ .if PROFILE
+  .if ! TRACE
+	call	write_profile_information
+  .endif
+ .endif
+
+	ret
+
+__driver:
+	mov	rbp,qword ptr _flags[rip]
+	test	rbp,16
+	att_je	__print__graph
+	att_jmp	__eval__to__nf
+
+print_time:
+	push	rbp
+
+	xor	rdx,rdx 
+	mov	rbx,1000
+	div	rbx 
+	mov	rcx,rax 
+	mov	rax,rdx 
+	xor	rdx,rdx 
+	mov	rbx,10
+	div	rbx
+
+	push	rax
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	rdi,rcx
+ .else
+	sub	rsp,32
+ .endif
+	att_call	_ew_print_int
+	mov	rsp,rbp
+
+	lea	rcx,sprintf_time_buffer[rip]
+
+	xor	rdx,rdx 
+	mov	rbx,10
+
+/*	movb	$'.',(%rcx ) */
+	mov	byte ptr [rcx],46
+
+	pop	rax
+
+	div	rbx 
+	add	rax,48
+	add	rdx,48
+	mov	byte ptr 1[rcx],al 
+	mov	byte ptr 2[rcx],dl 
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	rsi,3
+	mov	rdi,rcx
+ .else
+	mov	rdx,3
+	sub	rsp,32
+ .endif
+	att_call	_ew_print_text
+	mov	rsp,rbp
+
+	pop	rbp
+	ret
+
+print_sc:
+	mov	rbp,basic_only[rip]
+	test	rbp,rbp 
+	jne	end_print
+
+print:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rax
+ .else
+	mov	rcx,rax
+	sub	rsp,32
+ .endif
+	att_call	_w_print_string
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+
+end_print:
+	ret
+
+dump:
+	att_call	print
+	att_jmp	halt
+
+printD:	test	al,2
+	jne	printD_
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+
+	lea	rdi,4[rax]
+	mov	esi,0[rax]
+ .else
+	lea	rcx,4[rax]
+	mov	edx,dword ptr [rax]
+	sub	rsp,32
+ .endif
+	att_call	_w_print_text
+ .if LINUX
+	mov	rsi,r13
+	mov	rsi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+DtoAC_record:
+ .if NEW_DESCRIPTORS
+	movsxd	rbp,dword ptr (-6)[rax]
+ .else
+	movsx	rbp,dword ptr (-4)[rbp]
+ .endif
+	jmp	DtoAC_string_a2
+
+DtoAC:	test	al,2
+	jne	DtoAC_
+
+	mov	rbp,rax 
+	att_jmp	DtoAC_string_a2
+
+DtoAC_:
+ .if NEW_DESCRIPTORS
+	cmp	word ptr (-2)[rax],256
+	att_jae	DtoAC_record
+
+	movzx	rbx,word ptr [rax]
+	lea	rbp,10[rax+rbx]
+ .else
+	lea	rbp,(-2)[rax]
+	movsx	rbx,word ptr [rbp]
+	cmp	rbx,256
+	jae	DtoAC_record
+
+	shl	rbx,3
+	sub	rbp,rbx 
+
+ 	movzx	rbx,word ptr (-2)[rbp]
+	lea	rbp,4[rbp+rbx*8]
+ .endif
+
+DtoAC_string_a2:
+	mov	eax,dword ptr [rbp]
+	lea	rcx,4[rbp]
+	jmp	build_string
+
+print_symbol:
+	xor	rbx,rbx 
+	jmp	print_symbol_2
+
+print_symbol_sc:
+	mov	rbx,basic_only[rip]
+print_symbol_2:
+	mov	rax,[rcx]
+
+	lea	rbp,dINT+2[rip]
+	cmp	rax,rbp
+	je	print_int_node
+
+	lea	rbp,CHAR+2[rip]
+	cmp	rax,rbp
+	je	print_char_denotation
+
+	lea	rbp,BOOL+2[rip]
+	cmp	rax,rbp
+	je	print_bool
+
+	lea	rbp,REAL+2[rip]
+	cmp	rax,rbp
+	je	print_real_node
+	
+	test	rbx,rbx 
+	jne	end_print_symbol
+
+printD_:
+	cmp	word ptr (-2)[rax],256
+	jae	print_record
+
+	movzx	rbx,word ptr [rax]
+	lea	rbp,10[rax+rbx]
+	jmp	print_string_a2
+
+print_record:
+	movsxd	rbp,dword ptr (-6)[rax]
+	lea	rbp,-6[rax+rbp]
+	att_jmp	print_string_a2
+
+end_print_symbol:
+	ret
+
+print_int_node:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,8[rcx]
+ .else
+	sub	rsp,32
+	mov	rcx,8[rcx]
+ .endif
+	att_call	_w_print_int
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_int:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rax
+ .else
+	mov	rcx,rax
+	sub	rsp,32
+ .endif
+	att_call	_w_print_int
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_char_denotation:
+	test	rbx,rbx 
+	jne	print_char_node
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+ .else
+	sub	rsp,32
+ .endif
+	mov	rbx,8[rcx]
+
+ .if LINUX
+	mov	rdi,0x27
+ .else
+	mov	rcx,0x27
+ .endif
+	att_call	_w_print_char
+
+ .if LINUX
+	mov	rdi,rbx
+ .else
+	mov	rcx,rbx
+ .endif
+	att_call	_w_print_char
+
+ .if LINUX
+	mov	rdi,0x27
+ .else
+	mov	rcx,0x27
+ .endif
+	att_call	_w_print_char
+
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_char_node:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+ 
+	mov	rdi,8[rcx]
+.else
+	mov	rcx,8[rcx]
+	sub	rsp,32
+ .endif
+	att_call	_w_print_char
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+	
+print_char:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	
+	mov	rdi,rax
+ .else
+	mov	rcx,rax 
+	sub	rsp,32
+ .endif
+	att_call	_w_print_char
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_bool:
+	movsx	rcx,byte ptr 8[rcx]
+	test	rcx,rcx 
+	je	print_false
+
+print_true:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	lea	rdi,true_c_string[rip]
+ .else
+	lea	rcx,true_c_string
+	sub	rsp,32
+ .endif
+	att_call	_w_print_string
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_false:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	lea	rdi,false_c_string[rip]
+ .else
+	lea	rcx,false_c_string
+	sub	rsp,32
+ .endif
+	att_call	_w_print_string
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_real_node:
+	movlpd	xmm0,qword ptr 8[rcx]
+print_real:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+ .else
+	sub	rsp,32
+ .endif
+	att_call	_w_print_real
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print_string_a2:
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	lea	rdi,4[rbp]
+	mov	esi,0[rbp]
+	mov	rbp,rsp
+	and	rsp,-16
+ .else
+	lea	rcx,4[rbp]
+	mov	edx,0[rbp]
+	mov	rbp,rsp
+	and	rsp,-16
+	sub	rsp,32
+ .endif
+	att_call	_w_print_text
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+	ret
+
+print__chars__sc:
+	mov	rbp,basic_only[rip]
+	test	rbp,rbp 
+	jne	no_print_chars
+
+print__string__:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rsi,8[rcx]
+	lea	rdi,16[rcx]
+ .else
+	mov	rdx,8[rcx]
+	lea	rcx,16[rcx]
+	sub	rsp,32
+ .endif
+	att_call	_w_print_text
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+no_print_chars:
+	ret
+
+push_a_r_args:
+	push	rdi
+
+	mov	rdx,qword ptr 16[rcx]
+	sub	rdx,2
+	movzx	rdi,word ptr [rdx]
+	sub	rdi,256
+	movzx	rbx,word ptr 2[rdx]
+	add	rdx,4
+	push	rdx 
+		
+	mov	rdx,rdi 
+	sub	rdx,rbx 
+
+	shl	rax,3
+	lea	rcx,24[rcx+rbx*8]
+	dec	rdi 
+mul_array_size_lp:
+	add	rcx,rax 
+	sub	rdi,1
+	att_jnc	mul_array_size_lp
+
+	lea	rdi,[rcx+rdx*8]
+	jmp	push_a_elements
+push_a_elements_lp:
+	mov	rax,qword ptr (-8)[rcx]
+	sub	rcx,8
+	mov	qword ptr [rsi],rax 
+	add	rsi,8
+push_a_elements:
+	sub	rbx,1
+	att_jnc	push_a_elements_lp
+
+	mov	rcx,rdi
+	pop	rax
+	pop	rdi
+
+	pop	rbp 
+	jmp	push_b_elements
+push_b_elements_lp:
+	push	(-8)[rcx]
+	sub	rcx,8
+push_b_elements:
+	sub	rdx,1
+	att_jnc	push_b_elements_lp
+
+	jmp	rbp
+
+push_t_r_args:
+	pop	rbp
+
+	mov	rdx,qword ptr [rcx]
+	add	rcx,8
+	sub	rdx,2
+	movzx	rax,word ptr [rdx]
+	sub	rax,256
+	movzx	rbx,word ptr 2[rdx]
+	add	rdx,4
+
+	mov	qword ptr [rsi],rdx
+	mov	qword ptr 8[rsi],rbx
+
+	sub	rbx,rax
+	neg	rbx 
+
+	lea	rdx,[rcx+rax*8]
+	cmp	rax,2
+	jbe	small_record
+	mov	rdx,qword ptr 8[rcx]
+	lea	rdx,(-8)[rdx+rax*8]
+small_record:
+	jmp	push_r_b_elements
+
+push_r_b_elements_lp:
+	dec	rax 
+	jne	not_first_arg_b
+	
+	push	[rcx]
+	att_jmp	push_r_b_elements
+not_first_arg_b:
+	push	(-8)[rdx]
+	sub	rdx,8
+push_r_b_elements:
+	sub	rbx,1
+	att_jnc	push_r_b_elements_lp
+
+	mov	rbx,qword ptr 8[rsi]
+	push	rbp
+	push	[rsi]
+	jmp	push_r_a_elements
+
+push_r_a_elements_lp:
+	dec	rax 
+	jne	not_first_arg_a
+	
+	mov	rbp,qword ptr [rcx]
+	mov	qword ptr [rsi],rbp
+	add	rsi,8
+	att_jmp	push_r_a_elements
+not_first_arg_a:
+	mov	rbp,qword ptr (-8)[rdx]
+	sub	rdx,8
+	mov	qword ptr [rsi],rbp
+	add	rsi,8
+push_r_a_elements:
+	sub	rbx,1
+	att_jnc	push_r_a_elements_lp
+
+	pop	rax 
+	ret
+
+BtoAC:
+	test	al,al 
+	je	BtoAC_false
+BtoAC_true:
+	lea	rcx,true_string[rip]
+	ret
+BtoAC_false:
+	lea	rcx,false_string[rip]
+	ret
+
+RtoAC:
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	lea	rsi,printf_real_string[rip]
+	lea	rdi,sprintf_buffer[rip]
+	mov	rax,1
+	call	_sprintf
+	mov	rsi,r13
+	mov	rdi,r14
+ .else
+	lea	rdx,sprintf_buffer
+	sub	rsp,32
+	call	convert_real_to_string
+ .endif
+	mov	rsp,rbp
+	jmp	return_sprintf_buffer
+
+ItoAC:
+	lea	rcx,sprintf_buffer[rip]
+	call	int_to_string
+	
+	mov	rax,rcx
+	lea	rcx,sprintf_buffer[rip]
+	sub	rax,rcx
+
+	jmp	sprintf_buffer_to_string
+
+	.globl	convert_int_to_string
+convert_int_to_string:
+	push	rbp
+	push	rbx
+	mov	rax,rdx
+	att_call	int_to_string
+	mov	rax,rcx 
+	pop	rbx
+	pop	rbp
+	ret
+
+int_to_string:
+	test	rax,rax
+	jns	no_minus
+	mov	byte ptr [rcx],45
+	inc	rcx
+	neg	rax 
+no_minus:
+	mov	rbp,rcx
+
+	je	zero_digit
+	
+calculate_digits:
+	cmp	rax,10
+	jb	last_digit
+
+	mov	rdx,0x0cccccccccccccccd
+	mov	rbx,rax 
+
+	mul	rdx  
+
+	mov	rax,rdx 
+	and	rdx,-8
+	add	rbx,48
+
+	shr	rax,3
+	sub	rbx,rdx
+	shr	rdx,2
+
+	sub	rbx,rdx
+	mov	byte ptr [rcx],bl 
+
+	inc	rcx
+	att_jmp	calculate_digits
+
+last_digit:
+	test	rax,rax 
+	je	no_zero
+zero_digit:
+	add	rax,48
+	mov	byte ptr [rcx],al 
+	inc	rcx
+no_zero:
+	mov	rdx,rcx
+
+reverse_digits:
+	dec	rdx
+	cmp	rbp,rdx
+	jae	end_reverse_digits
+	mov	bl,byte ptr [rbp]
+	mov	al,byte ptr [rdx] 
+	mov	byte ptr [rdx],bl
+	mov	byte ptr [rbp],al
+	inc	rbp
+	att_jmp	reverse_digits
+
+end_reverse_digits:
+	mov	byte ptr [rcx],0
+	ret
+
+return_sprintf_buffer:
+	lea	rax,(sprintf_buffer-1)[rip]
+skip_characters:
+	inc	rax 
+	cmp	byte ptr [rax],0
+	att_jne	skip_characters
+
+	lea	rcx,sprintf_buffer[rip]
+	sub	rax,rcx
+
+sprintf_buffer_to_string:
+	lea	rcx,sprintf_buffer[rip]
+build_string:
+
+	lea	rbx,16+7[rax]
+	shr	rbx,3
+
+	sub	r15,rbx
+	jge	D_to_S_no_gc
+
+	push	rcx 
+	att_call	collect_0
+	pop	rcx 
+
+D_to_S_no_gc:
+	sub	rbx,2
+	mov	rbp,rdi 
+	lea	r9,__STRING__+2[rip]
+	mov	qword ptr [rdi],r9
+	mov	8[rdi],rax 
+	add	rdi,16
+	jmp	D_to_S_cp_str_2
+
+D_to_S_cp_str_1:
+	mov	rax,[rcx]
+	add	rcx,8
+	mov	[rdi],rax 
+	add	rdi,8
+D_to_S_cp_str_2:
+	sub	rbx,1
+	att_jnc	D_to_S_cp_str_1
+	
+	mov	rcx,rbp 
+	ret
+
+eqD:	mov	rax,[rcx]
+	cmp	rax,[rdx]
+	jne	eqD_false
+
+	lea	rbp,dINT+2[rip]
+	cmp	rax,rbp
+	je	eqD_INT
+	lea	rbp,CHAR+2[rip]
+	cmp	rax,rbp
+	je	eqD_CHAR
+	lea	rbp,BOOL+2[rip]
+	cmp	rax,rbp
+	je	eqD_BOOL
+	lea	rbp,REAL+2[rip]
+	cmp	rax,rbp
+	je	eqD_REAL
+
+	mov	rax ,1
+	ret
+
+eqD_CHAR:
+eqD_INT:
+	mov	rbx,8[rcx]
+	xor	rax,rax 
+	cmp	rbx,8[rdx]
+	sete	al
+	ret
+
+eqD_BOOL:
+	mov	bl,byte ptr 8[rcx]
+	xor	rax,rax 
+	cmp	bl,byte ptr 8[rdx]
+	sete	al 
+	ret
+
+eqD_REAL:
+	movlpd	xmm0,qword ptr 8[rcx]
+	comisd	xmm0,qword ptr 8[rdx]
+	fnstsw	ax
+	and	ah,68
+	xor	ah,64
+	sete	al
+	and	rax,1
+	ret
+
+eqD_false:
+	xor	rax ,rax 
+	ret
+/* */
+/*	the timer */
+/* */
+
+
+init_timer:
+	mov	rbp,rsp
+	and	rsp,-16
+	sub	rsp,32
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rsp
+	att_call    _times
+	mov	rsi,r13
+	mov	rdi,r14
+	mov	eax,[rsp]
+	imul	eax,10
+ .else
+	call	GetTickCount
+ .endif
+	mov	rsp,rbp
+
+	mov	last_time[rip],rax 
+	xor	rax ,rax 
+	mov	execute_time[rip],rax 
+	mov	garbage_collect_time[rip],rax 
+	mov	IO_time[rip],rax 
+
+	mov	mark_compact_garbage_collect_time[rip],rax 
+	mov	compact_garbage_collect_time[rip],rax 
+
+	ret
+
+get_time_diff:
+	mov	rbp,rsp
+	and	rsp,-16
+	sub	rsp,32
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,rsp
+	att_call    _times
+	mov	rsi,r13
+	mov	rdi,r14
+	mov	eax,[rsp]
+	imul	eax,10
+ .else
+	call	GetTickCount
+ .endif
+	mov	rsp,rbp
+
+	lea	rcx,last_time[rip]
+	mov	rdx,[rcx]
+	mov	[rcx],rax 
+	sub	rax,rdx
+	ret
+	
+add_execute_time:
+	att_call	get_time_diff
+	lea	rcx,execute_time[rip]
+
+add_time:
+	add	rax,[rcx]
+	mov	[rcx],rax 
+	ret
+
+add_garbage_collect_time:
+	att_call	get_time_diff
+	lea	rcx,garbage_collect_time[rip]
+	att_jmp	add_time
+
+add_IO_time:
+	att_call	get_time_diff
+	lea	rcx,IO_time[rip]
+	att_jmp	add_time
+
+add_mark_compact_garbage_collect_time:
+	att_call	get_time_diff
+	lea	rcx,mark_compact_garbage_collect_time[rip]
+	att_jmp	add_time
+
+add_compact_garbage_collect_time:
+	att_call	get_time_diff
+	lea	rcx,compact_garbage_collect_time[rip]
+	att_jmp	add_time
+/* */
+/*	the garbage collector */
+/* */
+
+collect_3:
+ .if PROFILE
+	lea	rbp,garbage_collector_name[rip]
+	att_call	profile_s
+ .endif
+	mov	[rsi],rcx 
+	mov	8[rsi],rdx 
+	mov	16[rsi],r8
+	add	rsi,24
+	call	collect_0_
+	mov	r8,(-8)[rsi]
+	mov	rdx,(-16)[rsi]
+	mov	rcx,(-24)[rsi]
+	sub	rsi,24
+ .if PROFILE
+	jmp	profile_r
+ .else
+	ret
+ .endif
+
+collect_2:
+ .if PROFILE
+	lea	rbp,garbage_collector_name[rip]
+	att_call	profile_s
+ .endif
+	mov	[rsi],rcx 
+	mov	8[rsi],rdx 
+	add	rsi,16
+	att_call	collect_0_
+	mov	rdx,(-8)[rsi]
+	mov	rcx,(-16)[rsi]
+	sub	rsi,16
+ .if PROFILE
+	att_jmp	profile_r
+ .else
+	ret
+ .endif
+
+collect_1:
+ .if PROFILE
+	lea	rbp,garbage_collector_name[rip]
+	att_call	profile_s
+ .endif
+	mov	[rsi],rcx 
+	add	rsi,8
+	att_call	collect_0_
+	mov	rcx,(-8)[rsi]
+	sub	rsi,8
+ .if PROFILE
+	att_jmp	profile_r
+ .else
+	ret
+ .endif
+
+collect_0:
+ .if PROFILE
+	lea	rbp,garbage_collector_name[rip]
+	att_call	profile_s
+ .endif
+	att_call	collect_0_
+ .if PROFILE
+	att_jmp	profile_r
+ .else
+	ret
+ .endif
+
+collect_0_:
+	mov	rbp,rdi 
+
+	push	rax
+	push	rbx
+
+	mov	rbx,qword ptr heap_end_after_gc[rip]
+	sub	rbx,rdi
+
+	shr	rbx,3
+	sub	rbx,r15
+	mov	qword ptr n_allocated_words[rip],rbx
+
+	test	byte ptr _flags[rip],64
+	je	no_mark3
+
+	mov	rbp,qword ptr bit_counter[rip]
+	test	rbp,rbp 
+	je	no_scan
+
+	push	rsi
+	mov	rsi,rbx
+
+	xor	rbx,rbx 
+	mov	rcx,qword ptr bit_vector_p[rip]
+
+scan_bits:
+	cmp	ebx,dword ptr[rcx]
+	je	zero_bits
+	mov	dword ptr [rcx],ebx 
+	add	rcx,4
+	sub	rbp,1
+	att_jne	scan_bits
+
+	jmp	end_scan
+
+zero_bits:
+	lea	rdx,4[rcx]
+	add	rcx,4
+	sub	rbp,1
+	jne	skip_zero_bits_lp1
+	jmp	end_bits
+
+skip_zero_bits_lp:
+	test	rax,rax 
+	jne	end_zero_bits
+skip_zero_bits_lp1:
+	mov	eax,dword ptr [rcx]
+	add	rcx,4
+	sub	rbp,1
+	att_jne	skip_zero_bits_lp
+
+	test	rax,rax 
+	att_je	end_bits
+	mov	dword ptr (-4)[rcx],ebx 
+	mov	rax,rcx 
+	sub	rax,rdx 
+	jmp	end_bits2
+
+end_zero_bits:
+	mov	rax,rcx 
+	sub	rax,rdx 
+	shl	rax,3
+	add	qword ptr n_free_words_after_mark[rip],rax
+	mov	dword ptr (-4)[rcx],ebx
+
+	cmp	rax,rsi
+	att_jb	scan_bits
+
+found_free_memory:
+	mov	qword ptr bit_counter[rip],rbp 
+	mov	qword ptr bit_vector_p[rip],rcx 
+
+	lea	rbp,(-4)[rdx]
+	sub	rbp,qword ptr heap_vector[rip]
+	shl	rbp,6
+	mov	rdi,qword ptr heap_p3[rip]
+	add	rdi,rbp 
+
+	lea	rbp,[rdi+rax*8]
+	mov	qword ptr heap_end_after_gc[rip],rbp 
+
+	mov	r15,rax
+	sub	r15,rsi
+
+	pop	rsi 
+	pop	rbx 
+	pop	rax 
+	ret
+
+end_bits:
+	mov	rax,rcx 
+	sub	rax,rdx 
+	add	rax,4
+end_bits2:
+	shl	rax,3
+	add	qword ptr n_free_words_after_mark[rip],rax 
+	cmp	rax,rsi 
+	att_jae	found_free_memory
+
+end_scan:
+	pop	rsi 
+	mov	qword ptr bit_counter[rip],rbp
+
+no_scan:
+
+no_mark3:
+	movsx	rax,byte ptr garbage_collect_flag[rip]
+	test	rax,rax 
+	jle	collect
+
+	sub	rax,2
+	mov	byte ptr garbage_collect_flag[rip],al 
+
+	mov	rbp,qword ptr extra_heap_size[rip]
+	cmp	rbx,rbp 
+	att_ja	collect
+
+	mov	rdi,qword ptr extra_heap[rip]
+
+	mov	r15,rbp
+
+	lea	rbp,[rdi+rbp*8]
+	mov	qword ptr heap_end_after_gc[rip],rbp 
+
+	sub	r15,rbx
+
+	pop	rbx 
+	pop	rax 
+	ret
+
+collect:
+ .if LINUX
+	sub	rsp,104
+ .else
+	sub	rsp,88
+ .endif
+	mov	32[rsp],r10
+	mov	24[rsp],r11
+	mov	16[rsp],r12
+	mov	8[rsp],r13
+	mov	[rsp],r14
+	movsd	40[rsp],xmm0
+	movsd	48[rsp],xmm1
+	movsd	56[rsp],xmm2
+	movsd	64[rsp],xmm3
+	movsd	72[rsp],xmm4
+	movsd	80[rsp],xmm5
+ .if LINUX
+	movsd	88[rsp],xmm6
+	movsd	96[rsp],xmm7
+ .endif
+
+	att_call	add_execute_time
+
+	test	qword ptr _flags[rip],4
+	je	no_print_stack_sizes
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+ 	mov	r13,rsi
+	mov	r14,rdi
+ .else
+	sub	rsp,32
+ .endif
+
+ .if 0
+  .if LINUX
+	mov	rdi,qword ptr 64[rsp]
+  .else
+	mov	rcx,qword ptr 96[rsp]
+  .endif
+	call	_ew_print_int
+
+  .if LINUX
+	mov	rdi,32
+  .else
+	mov	rcx,32
+  .endif
+	att_call	_ew_print_char
+ .endif
+
+ .if LINUX
+	lea	rdi,garbage_collect_string_1[rip]
+ .else
+	lea	rcx,garbage_collect_string_1
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rdi,r13
+	sub	rdi,stack_p[rip]
+ .else
+	mov	rcx,rsi 
+	sub	rcx,stack_p
+ .endif
+	att_call	_ew_print_int
+
+ .if LINUX
+	lea	rdi,garbage_collect_string_2[rip]
+ .else
+	lea	rcx,garbage_collect_string_2
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rdi,halt_sp[rip]
+	sub	rdi,rsp 
+ .else
+	mov	rcx,halt_sp
+	sub	rcx,rsp 
+ .endif
+	att_call	_ew_print_int
+
+ .if LINUX
+	lea	rdi,garbage_collect_string_3[rip]
+ .else
+	lea	rcx,garbage_collect_string_3
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	mov	rsp,rbp
+
+no_print_stack_sizes:
+	mov	rax,stack_p[rip]
+	add	rax,qword ptr _ab_stack_size[rip]
+	cmp	rsi,rax 
+	att_ja	stack_overflow
+
+	test	byte ptr _flags[rip],64
+	jne	compacting_collector
+
+	cmp	byte ptr garbage_collect_flag[rip],0
+	att_jne	compacting_collector
+
+	mov	rbp,heap_copied_vector[rip]
+
+	cmp	qword ptr heap_end_after_copy_gc[rip],0
+	je	zero_all
+
+	mov	rax,rdi
+	sub	rax,qword ptr heap_p1[rip]
+	add	rax,127*8
+	shr	rax,9
+	call	zero_bit_vector
+
+	mov	rdx,qword ptr heap_end_after_copy_gc[rip]
+	sub	rdx,qword ptr heap_p1[rip]
+	shr	rdx,7
+	and	rdx,-4
+
+	mov	rbp,qword ptr heap_copied_vector[rip]
+	mov	rax,qword ptr heap_copied_vector_size[rip]
+	add	rbp,rdx 
+	sub	rax,rdx 
+	shr	rax,2
+
+	mov	qword ptr heap_end_after_copy_gc[rip],0
+
+	att_call	zero_bit_vector
+	jmp	end_zero_bit_vector
+
+zero_all:
+	mov	rax,heap_copied_vector_size[rip]
+	shr	rax,2
+	att_call	zero_bit_vector
+
+end_zero_bit_vector:
+
+	.include	"acopy.s"
+
+	mov	qword ptr heap2_begin_and_end[rip],rsi 
+
+	mov	r15,rsi
+	sub	r15,rdi
+
+	mov	rax,heap_size_257[rip]
+	shl	rax,7
+	sub	rax,r15
+	add	qword ptr total_gc_bytes[rip],rax
+
+	shr	r15,3
+
+	pop	rsi 
+
+	att_call	add_garbage_collect_time
+
+	sub	r15,qword ptr n_allocated_words[rip]
+	jc	switch_to_mark_scan
+
+	lea	rax,[r15+r15*4]
+	shl	rax,6
+	mov	rbx,qword ptr _heap_size[rip]
+	mov	rcx,rbx 
+	shl	rbx,2
+	add	rbx,rcx 
+	add	rbx,rbx 
+	add	rbx,rcx
+	cmp	rax,rbx
+	jnc	no_mark_scan
+
+switch_to_mark_scan:
+	mov	rax,qword ptr heap_size_65[rip]
+	shl	rax,6
+	mov	rbx,qword ptr heap_p[rip]
+
+	mov	rcx,qword ptr heap_p1[rip]
+	cmp	rcx,qword ptr heap_p2[rip]
+	jc	vector_at_begin
+	
+vector_at_end:
+	mov	qword ptr heap_p3[rip],rbx 
+	add	rbx,rax 
+	mov	qword ptr heap_vector[rip],rbx 
+	
+	mov	rax,qword ptr heap_p1[rip]
+	mov	qword ptr extra_heap[rip],rax 
+	sub	rbx,rax 
+	shr	rbx,3
+	mov	qword ptr extra_heap_size[rip],rbx
+	jmp	switch_to_mark_scan_2
+
+vector_at_begin:
+	mov	qword ptr heap_vector[rip],rbx
+	add	rbx,qword ptr _heap_size[rip]
+	sub	rbx,rax 
+	mov	qword ptr heap_p3[rip],rbx 
+	
+	mov	qword ptr extra_heap[rip],rbx
+	mov	rcx,qword ptr heap_p2[rip]
+	sub	rcx,rbx 
+	shr	rcx,3
+	mov	qword ptr extra_heap_size[rip],rcx 
+
+switch_to_mark_scan_2:
+	mov	rax,heap_size_257[rip]
+	shl	rax,7-3
+	sub	rax,r15
+	shl	rax,3
+
+	mov	byte ptr garbage_collect_flag[rip],1
+
+	lea	rcx,heap_use_after_gc_string_1[rip]
+
+	test	r15,r15
+	jns	end_garbage_collect
+	
+	mov	byte ptr garbage_collect_flag[rip],-1
+	
+	mov	rbx,qword ptr extra_heap_size[rip]
+	mov	r15,rbx
+	sub	r15,qword ptr n_allocated_words[rip]
+	js	out_of_memory_4_3
+
+	mov	rdi,qword ptr extra_heap[rip]
+	shl	rbx,3
+	add	rbx,rdi
+	mov	qword ptr heap_end_after_gc[rip],rbx
+
+	mov	qword ptr heap_end_write_heap[rip],rdi
+
+	mov	qword ptr d3_flag_write_heap[rip],1
+	jmp	end_garbage_collect_
+
+no_mark_scan:
+/* exchange the semi_spaces */
+	mov	rax,heap_p1[rip]
+	mov	rbx,heap_p2[rip]
+	mov	heap_p2[rip],rax 
+	mov	heap_p1[rip],rbx 
+
+	mov	rax,heap_size_257[rip]
+	shl	rax,7-3
+	mov	rbx,rax
+	sub	rax,r15
+
+	mov	rcx,rax
+	imul	qword ptr _heap_size_multiple[rip]
+	shrd	rax,rdx,9
+	shr	rdx,9
+	jne	no_small_heap1
+
+	cmp	rax,4000
+	jge	not_too_small1
+	mov	rax,4000
+not_too_small1:
+	sub	rbx,rax 
+	att_jb	no_small_heap1
+
+	sub	r15,rbx
+	shl	rbx,3
+	mov	rbp,qword ptr heap_end_after_gc[rip]
+	mov	qword ptr heap_end_after_copy_gc[rip],rbp 
+	sub	rbp,rbx 
+	mov	qword ptr heap_end_after_gc[rip],rbp 
+
+no_small_heap1:
+	mov	rax,rcx
+	shl	rax,3
+
+	lea	rcx,heap_use_after_gc_string_1[rip]
+
+end_garbage_collect:
+
+	mov	qword ptr heap_end_write_heap[rip],rdi 
+	mov	qword ptr d3_flag_write_heap[rip],0
+
+end_garbage_collect_:
+	test	qword ptr _flags[rip],2
+	je	no_heap_use_message
+
+	push	rax
+
+	mov	rbp,rsp
+	and	rsp,-16
+
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi	
+
+	mov	rdi,rcx
+ .else
+	sub	rsp,32
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rdi,[rbp]
+ .else
+	mov	rcx,[rbp]
+ .endif
+	att_call	_ew_print_int
+
+ .if LINUX
+	lea	rdi,heap_use_after_gc_string_2[rip]
+ .else
+	lea	rcx,heap_use_after_gc_string_2
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .else
+ 	add	rsp,32
+ .endif
+	mov	rsp,rbp
+
+	pop	rax
+
+no_heap_use_message:
+	call	call_finalizers
+	 
+	test	byte ptr _flags[rip],32
+	je	no_write_heap
+
+	cmp	rax,qword ptr _min_write_heap_size[rip]
+	att_jb	no_write_heap
+
+	push	rcx 
+	push 	rdx 
+	push	rbp 
+	push	rsi 
+	push	rdi 
+
+	sub	rsp,128
+
+	mov	rax,qword ptr d3_flag_write_heap[rip]
+	test	rax,rax 
+	jne	copy_to_compact_with_alloc_in_extra_heap
+
+	movsx	rax,byte ptr garbage_collect_flag[rip]
+	
+	mov	rcx,qword ptr heap2_begin_and_end[rip]
+	mov	rdx,qword ptr (heap2_begin_and_end+8)[rip]
+
+	lea	rbx,heap_p1[rip]
+	
+	test	rax,rax 
+	je	gc0
+	
+	lea	rbx,heap_p2[rip]
+	jg	gc1
+
+	lea	rbx,heap_p3[rip]
+	xor	rcx,rcx 
+	xor	rdx,rdx 
+
+gc0:
+gc1:
+	mov	rbx,qword ptr [rbx]
+	
+	mov	rax,rsp 
+	
+	mov	qword ptr [rax],rbx 
+	mov	qword ptr 8[rax],rdi 
+	
+	mov	qword ptr 16[rax],rcx 
+	mov	qword ptr 24[rax],rdx 
+	
+	mov	rbx ,qword ptr stack_p[rip]
+
+	mov	qword ptr 32[rax],rbx 
+
+	mov	qword ptr 40[rax],rsi 
+	mov	qword ptr 48[rax],0
+	mov	qword ptr 56[rax],0
+	
+	lea	rbp,small_integers[rip]
+	mov	qword ptr 64[rax],rbp
+	lea	rbp,static_characters[rip]
+	mov	qword ptr 72[rax],rbp
+
+	lea	rbp,dINT+2[rip]
+	mov	qword ptr 80[rax],rbp
+	lea	rbp,CHAR+2[rip]
+	mov	qword ptr 88[rax],rbp
+	lea	rbp,REAL+2[rip]
+	mov	qword ptr 96[rax],rbp
+	lea	rbp,BOOL+2[rip]
+	mov	qword ptr 104[rax],rbp
+	lea	rbp,__STRING__+2[rip]
+	mov	qword ptr 112[rax],rbp
+	lea	rbp,__ARRAY__+2[rip]
+	mov	qword ptr 120[rax],rbp
+
+	mov	rbp,rsp
+	and	rsp,-16
+ .if LINUX
+	mov	rdi,rax
+ .else
+	mov	rcx,rax
+	sub	rsp,32
+ .endif
+ .if ! LINUX
+	call	write_heap
+ .endif
+	mov	rsp,rbp
+
+	add	rsp,128
+	
+	pop	rdi 
+	pop	rsi 
+	pop	rbp 
+	pop	rdx 
+	pop	rcx 
+no_write_heap:
+
+restore_registers_after_gc_and_return:
+	mov	r10,32[rsp]
+	mov	r11,24[rsp]
+	mov	r12,16[rsp]
+	mov	r13,8[rsp]
+	mov	r14,[rsp]
+	movlpd	xmm0,40[rsp]
+	movlpd	xmm1,48[rsp]
+	movlpd	xmm2,56[rsp]
+	movlpd	xmm3,64[rsp]
+	movlpd	xmm4,72[rsp]
+	movlpd	xmm5,80[rsp]
+ .if LINUX
+	movlpd	xmm6,88[rsp]
+	movlpd	xmm7,96[rsp]
+	add	rsp,104
+ .else
+	add	rsp,88
+ .endif
+	pop	rbx 
+	pop	rax 
+	ret
+
+call_finalizers:
+	mov	rax,qword ptr free_finalizer_list[rip]
+
+call_finalizers_lp:
+	lea	r9,__Nil-8[rip]
+	cmp	rax,r9
+	je	end_call_finalizers
+	push	8[rax]
+	mov	rbx,qword ptr 16[rax]
+	push	8[rbx]
+	call	qword ptr [rbx]
+	add	rsp,8
+	pop	rax 
+	att_jmp	call_finalizers_lp
+end_call_finalizers:
+
+	lea	r9,__Nil-8[rip]
+	mov	qword ptr free_finalizer_list[rip],r9
+	ret
+
+copy_to_compact_with_alloc_in_extra_heap:
+	mov	rcx,qword ptr heap2_begin_and_end[rip]
+	mov	rdx,qword ptr (heap2_begin_and_end+8)[rip]
+	lea	rbx,heap_p2[rip]
+	att_jmp	gc1
+
+allow_prefetch_for_athlon:
+	test	qword ptr _flags[rip],4096
+	jne	no_prefetch_flag
+
+	xor	rax,rax
+	cpuid
+	test	rax,rax
+	jz	disable_prefetch_flag
+
+ .if LINUX
+	cmp	rbx,'A+('u*0x100)+('t*0x10000)+('h*0x1000000)
+	att_jne	disable_prefetch_flag
+	cmp	rdx,'e+('n*0x100)+('t*0x10000)+('i*0x1000000)
+	att_jne	disable_prefetch_flag
+	cmp	rcx,'c+('A*0x100)+('M*0x10000)+('D*0x1000000)
+	att_jne	disable_prefetch_flag
+ .else
+	cmp	rbx,'A'+('u' shl 8)+('t' shl 16)+('h' shl 24)
+	jne	disable_prefetch_flag
+	cmp	rdx,'e'+('n' shl 8)+('t' shl 16)+('i' shl 24)
+	jne	disable_prefetch_flag
+	cmp	rcx,'c'+('A' shl 8)+('M' shl 16)+('D' shl 24)
+	jne	disable_prefetch_flag
+ .endif
+
+/*	mov	rax,1 */
+/*	cpuid */
+/*	and	rax,0x0f00 */
+/*	cmp	rax,0x600 */
+/*	je	keep_prefetch_flag */
+
+	ret
+
+disable_prefetch_flag:
+	and	qword ptr _flags[rip],-4097
+keep_prefetch_flag:
+no_prefetch_flag:
+	ret
+
+out_of_memory_4_3:
+out_of_memory_4_2:
+out_of_memory_4_1:
+out_of_memory_4:
+	att_call	add_garbage_collect_time
+	
+	lea	rbp,out_of_memory_string_4[rip]
+	att_jmp	print_error
+
+zero_bit_vector:
+	xor	rdx,rdx 
+	test	al,1
+	je	zero_bits1_1
+	mov	dword ptr [rbp],edx 
+	add	rbp,4
+zero_bits1_1:
+	shr	rax,1
+
+	mov	rbx,rax 
+	shr	rax,1
+	test	bl,1
+	je	zero_bits1_5
+
+	sub	rbp,8
+	jmp	zero_bits1_2
+
+zero_bits1_4:
+	mov	dword ptr [rbp],edx 
+	mov	dword ptr 4[rbp],edx 
+zero_bits1_2:
+	mov	dword ptr 8[rbp],edx 
+	mov	dword ptr 12[rbp],edx 
+	add	rbp,16
+zero_bits1_5:
+	sub	rax,1
+	att_jae	zero_bits1_4
+	ret
+
+reorder:
+	push	rsi 
+	push	rbp 
+
+	mov	rbp,rax 
+	shl	rbp,3
+	mov	rsi,rbx 
+	shl	rsi,3
+	add	rcx,rsi 
+	sub	rdx,rbp 
+
+	push	rsi 
+	push	rbp 
+	push	rbx 
+	push	rax 
+	jmp	st_reorder_lp
+
+reorder_lp:
+	mov	rbp,qword ptr [rcx]
+	mov	rsi,qword ptr (-8)[rdx]
+	mov	qword ptr (-8)[rdx],rbp 
+	sub	rdx,8
+	mov	qword ptr [rcx],rsi 
+	add	rcx,8
+	
+	dec	rax
+	jne	next_b_in_element
+	mov	rax,qword ptr [rsp]
+	add	rcx,qword ptr 24[rsp]
+next_b_in_element:
+	dec	rbx 
+	jne	next_a_in_element
+	mov	rbx,qword ptr 8[rsp]
+	sub	rdx,qword ptr 16[rsp]
+next_a_in_element:
+st_reorder_lp:
+	cmp	rdx,rcx 
+	att_ja	reorder_lp
+
+	pop	rax 
+	pop	rbx 
+	add	rsp,16
+	pop	rbp 
+	pop	rsi 
+	ret
+
+/* */
+/*	the sliding compacting garbage collector */
+/* */
+
+compacting_collector:
+/* zero all mark bits */
+
+	mov	rax,qword ptr heap_p3[rip]
+	neg	rax
+	mov	qword ptr neg_heap_p3[rip],rax 
+
+	mov	qword ptr stack_top[rip],rsi
+
+	mov	rdi,qword ptr heap_vector[rip]
+
+	test	byte ptr _flags[rip],64
+	je	no_mark4
+
+	cmp	qword ptr zero_bits_before_mark[rip],0
+	je	no_zero_bits
+
+	mov	qword ptr zero_bits_before_mark[rip],0
+
+no_mark4:
+	mov	rbp,rdi 
+	mov	rax,qword ptr heap_size_65[rip]
+	add	rax,3
+	shr	rax,2
+
+	xor	rbx,rbx 
+
+	test	al,1
+	je	zero_bits_1
+	mov	dword ptr [rbp],ebx 
+	add	rbp,4
+zero_bits_1:
+	mov	rcx,rax 
+	shr	rax,2
+
+	test	cl,2
+	je	zero_bits_5
+
+	sub	rbp,8
+	jmp	zero_bits_2
+
+zero_bits_4:
+	mov	dword ptr [rbp],ebx 
+	mov	dword ptr 4[rbp],ebx 
+zero_bits_2:
+	mov	dword ptr 8[rbp],ebx 
+	mov	dword ptr 12[rbp],ebx 
+	add	rbp,16
+zero_bits_5:
+	sub	rax,1
+	att_jnc	zero_bits_4
+
+	test	byte ptr _flags[rip],64
+	je	no_mark5
+
+no_zero_bits:
+	mov	rax,qword ptr n_last_heap_free_bytes[rip]
+	mov	rbx,qword ptr n_free_words_after_mark[rip]
+	shl	rbx,3
+
+	mov	rbp,rbx 
+	shl	rbp,3
+	add	rbp,rbx 
+	shr	rbp,2
+
+	cmp	rax,rbp 
+	jg	compact_gc
+
+	mov	rbx,qword ptr bit_vector_size[rip]
+	shl	rbx,3
+
+	sub	rax,rbx 
+	neg	rax
+
+	imul	qword ptr _heap_size_multiple[rip]
+	shrd	rax,rdx,7
+	shr	rdx,7
+	jne	no_smaller_heap
+	
+	cmp	rax,rbx 
+	att_jae	no_smaller_heap
+	
+	cmp	rbx,8000
+	att_jbe	no_smaller_heap
+	
+	att_jmp	compact_gc
+no_smaller_heap:
+	test	qword ptr _flags[rip],4096
+	jne	pmark
+
+	.include	"amark.s"
+
+	.include	"amark_prefetch.s"
+
+compact_gc:
+	mov	qword ptr zero_bits_before_mark[rip],1
+	mov	qword ptr n_last_heap_free_bytes[rip],0
+	mov	qword ptr n_free_words_after_mark[rip],1000
+
+no_mark5:
+
+	.include	"acompact.s"
+
+	mov	rsi,qword ptr stack_top[rip]
+
+	mov	rbx,qword ptr heap_size_65[rip]
+	shl	rbx,6
+	add	rbx,qword ptr heap_p3[rip]
+
+	mov	qword ptr heap_end_after_gc[rip],rbx 
+
+	sub	rbx,rdi
+	shr	rbx,3
+
+	sub	rbx,qword ptr n_allocated_words[rip]
+	mov	r15,rbx
+	att_jc	out_of_memory_4_1
+
+	mov	rax,rbx
+	shl	rax,2
+	add	rax,rbx
+	shl	rax,4
+	cmp	rax,qword ptr _heap_size[rip]
+	att_jc	out_of_memory_4_2
+
+ 	test	byte ptr _flags[rip],64
+	je	no_mark_6
+
+	mov	rax,qword ptr neg_heap_p3[rip]
+	add	rax,rdi 
+	mov	rbx,qword ptr n_allocated_words[rip]
+	lea	rax,[rax+rbx*8]
+
+	mov	rbx,qword ptr heap_size_65[rip]
+	shl	rbx,6
+	
+	imul	qword ptr _heap_size_multiple[rip]
+	shrd	rax,rdx,8
+	shr	rdx,8
+	jne	no_small_heap2
+
+	and	rax,-4
+
+	cmp	rax,8000
+	jae	not_too_small2
+	mov	rax,8000
+not_too_small2:
+	mov	rcx,rbx 
+	sub	rcx,rax 
+	att_jb	no_small_heap2
+	
+	sub	qword ptr heap_end_after_gc[rip],rcx 
+	shr	rcx,3
+	sub	r15,rcx
+
+	mov	rbx,rax
+
+no_small_heap2:
+	shr	rbx,3
+	mov	qword ptr bit_vector_size[rip],rbx
+
+no_mark_6:
+	jmp	no_copy_garbage_collection
+
+no_copy_garbage_collection:
+	att_call	add_compact_garbage_collect_time
+
+	mov	rax,rdi
+	sub	rax,qword ptr heap_p3[rip]
+
+	add	qword ptr total_compact_gc_bytes[rip],rax 
+
+	mov	rax,rdi 
+	sub	rax,qword ptr heap_p3[rip]
+	mov	rbx,qword ptr n_allocated_words[rip]
+	lea	rax,[rax+rbx*8]
+
+	lea	rcx,heap_use_after_compact_gc_string_1[rip]
+	att_jmp	end_garbage_collect
+
+	.globl	clean_exception_handler_
+
+clean_exception_handler_:
+
+	att_jmp	clean_exception_handler_
+
+	mov	rax,qword ptr [rcx]
+	cmp	dword ptr [rax],0x0c00000fd
+	je  	stack_overflow_exception
+
+	cmp	dword ptr [rax],0x80000001
+	je  	guard_page_or_access_violation_exception
+
+	cmp	dword ptr [rax] ,0x0c0000005
+	att_je  	guard_page_or_access_violation_exception
+
+no_stack_overflow_exception:
+	mov	rax,0
+	ret
+
+guard_page_or_access_violation_exception:
+	mov	rax,qword ptr 16[rax]
+	and	rax,-4096
+	cmp	qword ptr a_stack_guard_page[rip],rax
+	att_jne 	no_stack_overflow_exception
+	
+	cmp	qword ptr a_stack_guard_page[rip],0
+	att_je  	no_stack_overflow_exception
+
+stack_overflow_exception:
+	mov	rax,qword ptr 8[rcx]
+	lea	rax,stack_overflow[rip]
+	mov	qword ptr (0x0F8)[rax],rax
+
+	mov	rax,-1
+	ret
+
+stack_overflow:
+	att_call	add_execute_time
+
+	lea	rbp,stack_overflow_string[rip]
+	att_jmp	print_error
+
+_IO_error:
+	mov	rbp,rsp
+	and	rsp,-16
+
+	mov	rbx,rcx
+ .if LINUX
+	lea	rdi,IO_error_string[rip]
+ .else
+	sub	rsp,32
+	lea	rcx,IO_error_string
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	rdi,rbx
+ .else
+	mov	rcx,rbx
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	lea	rdi,new_line_string[rip]
+ .else
+	lea	rcx,new_line_string
+ .endif
+	att_call	_ew_print_string
+
+	mov	rsp,rbp
+
+	att_jmp	halt
+
+print_error:
+ .if LINUX
+	mov	rdi,rbp
+ .else
+	mov	rcx,rbp 
+ .endif
+	mov	rbp,rsp
+	and	rsp,-16
+	att_call	_ew_print_string
+	mov	rsp,rbp
+
+halt:
+	mov	rsp,halt_sp[rip]
+
+ .if PROFILE
+	call	write_profile_stack
+ .endif
+
+	mov	qword ptr _execution_aborted[rip],1
+
+	cmp	qword ptr dll_initisialised[rip],0
+ .if LINUX
+	att_je	exit_
+ .else
+	je	exit
+ .endif
+ .if LINUX
+	cmp	dword ptr _return_code[rip],0
+ .else
+	cmp	qword ptr return_code,0
+ .endif
+	jne	return_code_set
+ .if LINUX
+	mov	dword ptr _return_code[rip],-1
+ .else
+	mov	qword ptr return_code,-1
+ .endif
+return_code_set:
+ .if LINUX
+	mov	edi,dword ptr _return_code[rip]
+	and	rsp,-16
+	att_call	_exit
+ .else
+	push	qword ptr return_code
+	call	(ExitProcess)
+ .endif
+	att_jmp	return_code_set
+
+e__system__eaind:
+__eaind:
+eval_fill:
+	mov	[rsi],rcx 
+	add	rsi,8
+	mov	rcx,rdx 
+	call	qword ptr [rdx]
+	mov	rdx,rcx 
+	mov	rcx,(-8)[rsi]
+	sub	rsi,8
+	
+	mov	rbp,[rdx]
+	mov	[rcx],rbp 
+	mov	rbp,8[rdx]
+	mov	8[rcx],rbp 
+	mov	rbp,16[rdx]
+	mov	16[rcx],rbp 
+	ret
+
+	.align	2 
+	lea	rax,e__system__eaind[rip]
+	jmp	rax
+	.byte	0,0,0
+	.long	0x80000000 /* e__system__dind */
+	.long	-2
+e__system__nind:
+__indirection:
+	mov	rdx,8[rcx]
+	mov	rax,[rdx]
+	test	al,2
+
+	je	eval_fill2
+
+	mov	[rcx],rax 
+	mov	rbp,8[rdx]
+	mov	8[rcx],rbp 
+	mov	rbp,16[rdx]
+	mov	16[rcx],rbp 
+	ret
+
+eval_fill2:
+	lea	r9,__cycle__in__spine[rip]
+	mov	qword ptr [rcx],r9
+	mov	qword ptr [rsi],rcx 
+
+	test	byte ptr _flags[rip],64
+	att_je	__cycle__in__spine	
+
+	add	rsi,8
+	mov	rcx,rdx 
+	call	rax 
+	mov	rdx,rcx
+	mov	rcx,qword ptr (-8)[rsi]
+	sub	rsi,8
+	
+	mov	rbp,[rdx]
+	mov	[rcx],rbp
+	mov	rbp,8[rdx]
+	mov	8[rcx],rbp
+	mov	rbp,16[rdx]
+	mov	16[rcx],rbp 
+	ret
+
+ .if PROFILE
+	call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_0:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	8[rdx],rcx 
+	jmp	rbp
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_1:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	rax,8[rdx]
+	mov	8[rdx],rcx 
+	mov	rdx,rax 
+	jmp	rbp
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_2:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	8[rdx],rcx
+	mov	rdx,16[rdx]
+	jmp	rbp 
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_3:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	8[rdx],rcx
+	mov	[rsi],rcx
+	mov	rcx,24[rdx]
+	add	rsi,8
+	mov	rdx,16[rdx]
+	jmp	rbp
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_4:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	8[rdx],rcx 
+	mov	[rsi],rcx 
+	mov	rbx,32[rdx]
+	mov	8[rsi],rbx 
+	mov	rcx,24[rdx]
+	add	rsi,16
+	mov	rdx,16[rdx]
+	jmp	rbp
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_5:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	[rsi],rcx 
+	mov	8[rdx],rcx 
+	mov	rbx,40[rdx]
+	mov	8[rsi],rbx 
+	mov	rbx,32[rdx]
+	mov	16[rsi],rbx 
+	mov	rcx,24[rdx]
+	add	rsi,24
+	mov	rdx,16[rdx]
+	jmp	rbp
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_6:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	[rsi],rcx 
+	mov	8[rdx],rcx 
+	mov	rbx,48[rdx]
+	mov	8[rsi],rbx
+	mov	rbx,40[rdx]
+	mov	16[rsi],rbx 
+	mov	rbx,32[rdx]
+	mov	24[rsi],rbx 
+	mov	rcx,24[rdx]
+	add	rsi,32
+	mov	rdx,16[rdx]
+	jmp	rbp 
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_7:
+	mov	rax,0
+	mov	rbx,40
+eval_upd_n:
+	lea	r8,__indirection[rip]
+	mov	qword ptr [rdx],r8
+	mov	r8,8[rdx]
+	mov	[rsi],rcx 
+	mov	8[rdx],rcx 
+	add	rdx,rbx 
+	mov	rbx,16[rdx ]
+	mov	8[rsi],rbx 
+	mov	rbx,8[rdx]
+	mov	16[rsi],rbx 
+	mov	rbx,[rdx]
+	mov	24[rsi],rbx 
+	add	rsi,32
+
+eval_upd_n_lp:
+	mov	rbx,(-8)[rdx]
+	sub	rdx,8
+	mov	[rsi],rbx 
+	add	rsi,8
+	sub	rax,1
+	att_jnc	eval_upd_n_lp
+
+	mov	rcx,(-8)[rdx]
+	mov	rdx,(-16)[rdx ]
+	jmp	rbp 
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_8:
+	mov	rax,1
+	mov	rbx,48
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_9:
+	mov	rax,2
+	mov	rbx,56
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_10:
+	mov	rax,3
+	mov	rbx,64
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_11:
+	mov	rax,4
+	mov	rbx,72
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_12:
+	mov	rax,5
+	mov	rbx,80
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_13:
+	mov	rax,6
+	mov	rbx,88
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_14:
+	mov	rax,7
+	mov	rbx,96
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_15:
+	mov	rax,8
+	mov	rbx,104
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_16:
+	mov	rax,9
+	mov	rbx,112
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_17:
+	mov	rax,10
+	mov	rbx,120
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_18:
+	mov	rax,11
+	mov	rbx,128
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_19:
+	mov	rax,12
+	mov	rbx,136
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_20:
+	mov	rax,13
+	mov	rbx,144
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_21:
+	mov	rax,14
+	mov	rbx,152
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_22:
+	mov	rax,15
+	mov	rbx,160
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_23:
+	mov	rax,16
+	mov	rbx,168
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_24:
+	mov	rax,17
+	mov	rbx,176
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_25:
+	mov	rax,18
+	mov	rbx,184
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_26:
+	mov	rax,19
+	mov	rbx,192
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_27:
+	mov	rax,20
+	mov	rbx,200
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_28:
+	mov	rax,21
+	mov	rbx,208
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_29:
+	mov	rax,22
+	mov	rbx,216
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_30:
+	mov	rax,23
+	mov	rbx,224
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_31:
+	mov	rax,24
+	mov	rbx,232
+	att_jmp	eval_upd_n
+
+ .if PROFILE
+	att_call	profile_n
+	mov	rbp,rax
+ .endif
+eval_upd_32:
+	mov	rax,25
+	mov	rbx,240
+	att_jmp	eval_upd_n
+
+/* */
+/*	STRINGS */
+/* */
+
+catAC:
+	mov	rax,8[rcx]
+	mov	rbx,8[rdx]
+	lea	rbp,16+7[rax+rbx]
+	shr	rbp,3
+	sub	r15,rbp
+	jl	gc_3
+gc_r_3:
+	add	rcx,16
+	add	rdx,16
+
+/* fill_node */
+
+	mov	r8,rdi
+	lea	rbp,__STRING__+2[rip]
+	mov	qword ptr [rdi],rbp
+
+/* store length */
+
+	lea	rbp,[rax+rbx]
+	mov	8[rdi],rbp 
+	add	rdi,16
+
+/* copy string 1 */
+
+	lea	rbp,7[rbx]
+	shr	rbp,3
+	add	rbx,rdi 
+
+	xchg	rcx,rbp 
+	xchg	rsi,rdx 
+	cld
+	rep movsq
+	mov	rsi,rdx 
+	mov	rcx,rbp 
+
+	mov	rdi,rbx 
+
+/* copy_string 2 */
+
+cat_string_6:
+	mov	rbp,rax 
+	shr	rbp,3
+	je	cat_string_9
+
+cat_string_7:
+	mov	rbx,[rcx]
+	add	rcx,8
+	mov	[rdi],rbx 
+	add	rdi,8
+	dec	rbp 
+	att_jne	cat_string_7
+	
+cat_string_9:
+	test	al,4
+	je	cat_string_10
+	mov	ebx,dword ptr [rcx]
+	add	rcx,4
+	mov	dword ptr [rdi],ebx 
+	add	rdi,4
+cat_string_10:
+	test	al,2
+	je	cat_string_11
+	mov	bx,word ptr [rcx]
+	add	rcx,2
+	mov	word ptr [rdi],bx 
+	add	rdi,2
+cat_string_11:
+	test	al,1
+	je	cat_string_12
+	mov	bl,byte ptr [rcx]
+	mov	byte ptr [rdi],bl 
+	inc	rdi
+cat_string_12:
+
+	mov	rcx,r8
+/* .align heap pointer */
+	add	rdi,7
+	and	rdi,-8
+	ret
+
+gc_3:	att_call	collect_2
+	att_jmp	gc_r_3
+
+empty_string:
+	lea	rcx,zero_length_string[rip]
+	ret
+
+sliceAC:
+	mov	rbp,8[rcx]
+	test	rbx,rbx 
+	jns	slice_string_1
+	xor	rbx,rbx 
+slice_string_1:
+	cmp	rbx,rbp 
+	att_jge	empty_string
+	cmp	rax,rbx 
+	att_jl	empty_string
+	inc	rax 
+	cmp	rax,rbp 
+	jle	slice_string_2
+	mov	rax,rbp 
+slice_string_2:
+	sub	rax,rbx 
+
+	lea	rbp,(16+7)[rax]
+	shr	rbp,3
+
+	sub	r15,rbp
+	jl	gc_4
+r_gc_4:
+	sub	rbp,2
+	lea	rdx,16[rcx+rbx]
+
+	lea	rcx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rcx
+	mov	8[rdi],rax 
+
+/* copy part of string */
+	mov	rcx,rbp 
+	mov	rbp,rdi 
+	add	rdi,16
+
+	xchg	rsi,rdx 
+	cld
+	rep movsq
+	mov	rsi,rdx 
+	mov	rcx,rbp 
+	ret
+
+gc_4:
+	mov	rbp,rdx 
+	att_call	collect_1
+	lea	rbp,(16+7)[rax]
+	shr	rbp,3
+	att_jmp	r_gc_4
+
+updateAC:
+	mov	rbp,8[rcx]
+	cmp	rbx,rbp 
+	jae	update_string_error
+
+	add	rbp,16+7
+	shr	rbp,3
+
+	sub	r15,rbp
+	jl	gc_5
+r_gc_5:
+	mov	rbp,8[rcx]
+	add	rbp,7
+	shr	rbp,3
+
+	mov	rdx,rcx 
+	mov	r8,rdi 
+	lea	rcx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rcx
+	mov	rcx,8[rdx]
+	add	rdx,16
+	mov	8[rdi],rcx 
+	add	rdi,16
+
+	add	rbx,rdi
+
+	mov	rcx,rbp 
+	xchg	rsi,rdx 
+	cld
+	rep movsq
+	mov	rsi,rdx 
+
+	mov	byte ptr [rbx],al 
+	mov	rcx,r8 
+	ret
+
+gc_5:	att_call	collect_1
+	att_jmp	r_gc_5
+
+update_string_error:
+	lea	rbp,high_index_string[rip]
+	test	rax,rax 
+	jns	update_string_error_2
+	lea	rbp,low_index_string[rip]
+update_string_error_2:
+	att_jmp	print_error
+
+eqAC:
+	mov	rax,8[rcx]
+	cmp	rax,8[rdx]
+	jne	equal_string_ne
+	add	rcx,16
+	add	rdx,16
+	mov	rbx,rax 
+	and	rbx,7
+	shr	rax,3
+	je	equal_string_d
+equal_string_1:
+	mov	rbp,[rcx]
+	cmp	rbp,[rdx]
+	att_jne	equal_string_ne
+	add	rcx,8
+	add	rdx,8
+	dec	rax
+	att_jne	equal_string_1
+equal_string_d:
+	test	bl,4
+	je	equal_string_w
+	mov	eax,dword ptr [rcx]
+	cmp	eax,dword ptr [rdx]
+	att_jne	equal_string_ne
+	add	rcx,4
+	add	rdx,4
+equal_string_w:
+	test	bl,2
+	je	equal_string_b
+	mov	ax,word ptr [rcx]
+	cmp	ax,word ptr [rdx]
+	att_jne	equal_string_ne
+	add	rcx,2
+	add	rdx,2
+equal_string_b:
+	test	bl,1
+	je	equal_string_eq
+	mov	bl,byte ptr [rcx]
+	cmp	bl,byte ptr [rdx]
+	att_jne	equal_string_ne
+equal_string_eq:
+	mov	rax,1
+	ret
+equal_string_ne:
+	xor	rax,rax 
+	ret
+
+cmpAC:
+	mov	rbx,8[rcx]
+	mov	rbp,8[rdx]
+	add	rcx,16
+	add	rdx,16
+	cmp	rbp,rbx 
+	jb	cmp_string_less
+	ja	cmp_string_more
+	xor	rax,rax 
+	jmp	cmp_string_chars
+cmp_string_more:
+	mov	rax,1
+	att_jmp	cmp_string_chars
+cmp_string_less:
+	mov	rax,-1
+	mov	rbx,rbp 
+	att_jmp	cmp_string_chars
+
+cmp_string_1:
+	mov	rbp,[rdx]
+	cmp	rbp,[rcx]
+	jne	cmp_string_ne_q
+	add	rdx,8
+	add	rcx,8
+cmp_string_chars:
+	sub	rbx,8
+	att_jnc	cmp_string_1
+cmp_string_d:
+	test	bl,4
+	je	cmp_string_w
+	mov	ebp,dword ptr [rdx]
+	cmp	ebp,dword ptr [rcx]
+	jne	cmp_string_ne_d
+	add	rdx,4
+	add	rcx,4
+cmp_string_w:
+	test	bl,2
+	je	cmp_string_b
+	mov	bpl,byte ptr [rdx]
+	cmp	bpl,byte ptr [rcx]
+	jne	cmp_string_ne
+	mov	bpl,byte ptr 1[rdx]
+	cmp	bpl,byte ptr 1[rcx]
+	att_jne	cmp_string_ne
+	add	rdx,2
+	add	rcx,2
+cmp_string_b:
+	test	bl,1
+	je	cmp_string_eq
+	mov	bl,byte ptr [rdx]
+	cmp	bl,byte ptr [rcx]
+	att_jne	cmp_string_ne
+cmp_string_eq:
+	ret
+cmp_string_ne_d:
+	mov	r10d,[rcx]
+	bswap	ebp
+	bswap	r10d
+	cmp	ebp,r10d
+	att_jmp	cmp_string_ne
+cmp_string_ne_q:
+	mov	r10,[rcx]
+	bswap	rbp
+	bswap	r10
+	cmp	rbp,r10
+cmp_string_ne:
+	ja	cmp_string_r1
+	mov	rax,-1
+	ret
+cmp_string_r1:
+	mov	rax,1
+	ret
+
+string_to_string_node:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+
+	lea	rbx,16+7[rax]
+	shr	rbx,3
+
+	sub	r15,rbx
+	jl	string_to_string_node_gc
+
+string_to_string_node_r:
+	sub	rbx,2
+	lea	rbp,__STRING__+2[rip]
+	mov	qword ptr [rdi],rbp
+	mov	qword ptr 8[rdi],rax 
+	mov	rbp,rdi
+	add	rdi,16
+	jmp	string_to_string_node_4
+	
+string_to_string_node_2:
+	mov	rax,qword ptr [rcx]
+	add	rcx,8
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+string_to_string_node_4:
+	sub	rbx,1
+	att_jge	string_to_string_node_2
+
+	mov	rcx,rbp 
+	ret
+
+string_to_string_node_gc:
+	push	rcx
+	att_call	collect_0
+	pop	rcx
+	att_jmp	string_to_string_node_r
+
+
+int_array_to_node:
+	mov	rax,qword ptr -16[rcx]
+	lea	rbx,3[rax]
+	sub	r15,rbx
+	jl	int_array_to_node_gc
+
+int_array_to_node_r:
+	lea	rbx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rbx
+	mov	rdx,rcx
+	mov	qword ptr 8[rdi],rax
+	mov	rcx,rdi
+	lea	rbx,dINT+2[rip]
+	mov	qword ptr 16[rdi],rbx
+	add	rdi,24
+	jmp	int_or_real_array_to_node_4
+
+int_or_real_array_to_node_2:
+	mov	rbx,qword ptr [rdx]
+	add	rdx,8
+	mov	qword ptr [rdi],rbx
+	add	rdi,8
+int_or_real_array_to_node_4:
+	sub	rax,1
+	att_jge	int_or_real_array_to_node_2
+
+	ret
+
+int_array_to_node_gc:
+	push	rcx
+	att_call	collect_0
+	pop	rcx
+	att_jmp	int_array_to_node_r
+
+
+real_array_to_node:
+	mov	rax,qword ptr -16[rcx]
+	lea	rbx,3[rax]
+	sub	r15,rbx
+	jl	real_array_to_node_gc
+
+real_array_to_node_r:
+	lea	rbx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rbx
+	mov	rdx,rcx
+	mov	qword ptr 8[rdi],rax
+	mov	rcx,rdi
+	lea	rbx,REAL+2[rip]
+	mov	qword ptr 16[rdi],rbx
+	add	rdi,24
+	att_jmp	int_or_real_array_to_node_4
+
+real_array_to_node_gc:
+	push	rcx
+	att_call	collect_0
+	pop	rcx
+	att_jmp	real_array_to_node_r
+
+
+	.align	2
+	.long	3
+_c3:	att_jmp	__cycle__in__spine
+	.align	2
+
+	.long	4
+_c4:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	5
+_c5:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	6
+_c6:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	7
+_c7:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	8
+_c8:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	9
+_c9:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	10
+_c10:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	11
+_c11:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	12
+_c12:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	13
+_c13:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	14
+_c14:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	15
+_c15:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	16
+_c16:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	17
+_c17:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	18
+_c18:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	19
+_c19:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	20
+_c20:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	21
+_c21:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	22
+_c22:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	23
+_c23:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	24
+_c24:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	25
+_c25:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	26
+_c26:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	27
+_c27:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	28
+_c28:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	29
+_c29:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	30
+_c30:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	31
+_c31:	att_jmp	__cycle__in__spine
+	.align	2
+	.long	32
+_c32:	att_jmp	__cycle__in__spine
+
+/* */
+/*	ARRAYS */
+/* */
+
+_create_arrayB:
+	mov	rbx,rax 
+	add	rax,24+7
+	shr	rax,3
+	sub	r15,rax
+	jge	no_collect_4574
+	att_call	collect_0
+no_collect_4574:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rbx 
+	lea	rdx,BOOL+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	lea	rdi,[rdi+rax*8]
+	ret
+
+_create_arrayC:
+	mov	rbx,rax
+	add	rax,16+7
+	shr	rax,3
+	sub	r15,rax
+	jge	no_collect_4573
+	att_call	collect_0
+no_collect_4573:
+	mov	rcx,rdi 
+	lea	rdx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rbx 
+	lea	rdi,[rdi+rax*8]
+	ret
+
+_create_arrayI:
+	lea	rbp,3[rax]
+	sub	r15,rbp
+	jge	no_collect_4572
+	att_call	collect_0
+no_collect_4572:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax
+	lea	rbp,dINT+2[rip]
+	mov	qword ptr 16[rdi],rbp
+	lea	rdi,24[rdi+rax*8]
+	ret
+
+_create_arrayI32:
+	mov	rbx,rax
+	add	rax,6+1
+	shr	rax,1
+	sub	r15,rax
+	jge	no_collect_3572
+	att_call	collect_0
+no_collect_3572:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rbx
+	lea	rdx,INT32+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	lea	rdi,[rdi+rax*8]
+	ret
+
+_create_arrayR:
+	lea	rbp,3[rax]
+	sub	r15,rbp
+	jge	no_collect_4580
+	att_call	collect_0
+no_collect_4580:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax
+	lea	rdx,REAL+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	lea	rdi,24[rdi+rax*8]
+	ret
+
+_create_arrayR32:
+	mov	rbx,rax
+	add	rax,6+1
+	shr	rax,1
+	sub	r15,rax
+	jge	no_collect_3580
+	att_call	collect_0
+no_collect_3580:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax
+	lea	rdx,REAL32+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	lea	rdi,[rdi+rax*8]
+	ret
+
+/* rax : number of elements, rbx: element descriptor */
+/* r10 : element size, r11 : element a size, rcx :a_element-> rcx : array */
+
+_create_r_array:
+	mov	rbp,rax
+	imul	rbp,r10
+	add	rbp,3
+	sub	r15,rbp
+	jge	no_collect_4586
+	att_call	collect_1
+no_collect_4586:
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	rdx,rcx
+	mov	qword ptr 8[rdi],rax
+	mov	rcx,rdi 
+	mov	qword ptr 16[rdi],rbx
+	add	rdi,24
+
+	test	r11,r11
+	je	_create_r_array_0
+	sub	r11,2
+	jc	_create_r_array_1
+	je	_create_r_array_2
+	sub	r11,2
+	jc	_create_r_array_3
+	je	_create_r_array_4
+	jmp	_create_r_array_5
+
+_create_r_array_0:
+	imul	r10,rax
+	lea	rdi,[rdi+r10*8]
+	ret
+
+_create_r_array_1:
+	shl	r10,3
+	jmp	_st_fillr1_array
+_fillr1_array:
+	mov	qword ptr [rdi],rdx
+	add	rdi,r10 
+_st_fillr1_array:
+	sub	rax,1
+	att_jnc	_fillr1_array
+	ret
+
+_create_r_array_2:
+	shl	r10,3
+	jmp	_st_fillr2_array
+_fillr2_array:
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rdx
+	add	rdi,r10 
+_st_fillr2_array:
+	sub	rax,1
+	att_jnc	_fillr2_array
+	ret
+
+_create_r_array_3:
+	shl	r10,3
+	jmp	_st_fillr3_array
+_fillr3_array:
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rdx
+	mov	qword ptr 16[rdi],rdx
+	add	rdi,r10 
+_st_fillr3_array:
+	sub	rax,1
+	att_jnc	_fillr3_array
+	ret
+
+_create_r_array_4:
+	shl	r10,3
+	jmp	_st_fillr4_array
+_fillr4_array:
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rdx
+	mov	qword ptr 16[rdi],rdx
+	mov	qword ptr 24[rdi],rdx
+	add	rdi,r10 
+_st_fillr4_array:
+	sub	rax,1
+	att_jnc	_fillr4_array
+	ret
+
+_create_r_array_5:
+	sub	r10,4
+	sub	r10,r11
+	sub	r11,1
+	shl	r10,3
+	jmp	_st_fillr5_array
+
+_fillr5_array:
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rdx
+	mov	qword ptr 16[rdi],rdx
+	mov	qword ptr 24[rdi],rdx
+	add	rdi,32
+
+	mov	rbx,r11
+_copy_elem_5_lp:
+	mov	qword ptr [rdi],rdx
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	_copy_elem_5_lp
+
+	add	rdi,r10 
+_st_fillr5_array:
+	sub	rax,1
+	att_jnc	_fillr5_array
+	
+	ret
+
+create_arrayB:
+	mov	r10,rbx
+	add	rbx,24+7
+	shr	rbx,3
+	sub	r15,rbx
+	jge	no_collect_4575
+	att_call	collect_0
+no_collect_4575:
+	mov	rbp,rax
+	sub	rbx,3
+	shl	rbp,8
+	or	rax,rbp
+	mov	rbp,rax
+	shl	rbp,16
+	or	rax,rbp
+	mov	rbp,rax 
+	shl	rbp,32
+	or	rax,rbp
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],r10
+	lea	rdx,BOOL+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	add	rdi,24
+	jmp	create_arrayBCI
+
+create_arrayC:
+	mov	r10,rbx
+	add	rbx,16+7
+	shr	rbx,3
+	sub	r15,rbx
+	jge	no_collect_4578
+	att_call	collect_0
+no_collect_4578:
+	mov	rbp,rax
+	sub	rbx,2
+	shl	rbp,8
+	or	rax,rbp
+	mov	rbp,rax
+	shl	rbp,16
+	or	rax,rbp
+	mov	rbp,rax
+	shl	rbp,32
+	or	rax,rbp
+	mov	rcx,rdi 
+	lea	rdx,__STRING__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],r10
+	add	rdi,16
+	att_jmp	create_arrayBCI
+
+create_arrayI32:
+	mov	r10,rbx
+	add	rbx,6+1
+	shr	rbx,1
+	sub	r15,rbx
+	jge	no_collect_3577
+	att_call	collect_0
+no_collect_3577:
+	mov	rcx,rdi 
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],r10
+	lea	rdx,INT32+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	add	rdi,24
+
+	sub	rbx,3
+
+	mov	ebp,eax
+	shl	rax,32
+	or	rax,rbp
+	att_jmp	create_arrayBCI
+
+create_arrayI:
+	lea	rbp,3[rbx]
+	sub	r15,rbp
+	jge	no_collect_4577
+	att_call	collect_0
+no_collect_4577:
+	mov	rcx,rdi 
+	lea	rbp,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rbp
+	mov	qword ptr 8[rdi],rbx 
+	lea	rbp,dINT+2[rip]
+	mov	qword ptr 16[rdi],rbp
+	add	rdi,24
+create_arrayBCI:
+	mov	rdx,rbx
+	shr	rbx,1
+	test	dl,1
+	je	st_filli_array
+
+	mov	qword ptr [rdi],rax 
+	add	rdi,8
+	att_jmp	st_filli_array
+
+filli_array:
+	mov	qword ptr [rdi],rax 
+	mov	qword ptr 8[rdi],rax 
+	add	rdi,16
+st_filli_array:
+	sub	rbx,1
+	att_jnc	filli_array
+
+	ret
+
+create_arrayR32:
+	cvtsd2ss	xmm0,xmm0
+	movss	dword ptr (-8)[rsp],xmm0
+	mov	r10,rax
+	add	rax,6+1
+	shr	rax,1
+	mov	ebx,dword ptr (-8)[rsp]
+	sub	r15,rax
+	jge	no_collect_3579
+	att_call	collect_0
+no_collect_3579:
+	mov	rcx,rdi 
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],r10
+	lea	rdx,REAL32+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	add	rdi,24
+
+	sub	rax,3
+
+	mov	edx,ebx
+	shl	rbx,32
+	or	rbx,rdx
+	jmp	st_fillr_array
+
+create_arrayR:
+	movsd	qword ptr (-8)[rsp],xmm0
+	lea	rbp,3[rax]
+
+	mov	rbx,qword ptr (-8)[rsp]
+
+	sub	r15,rbp
+	jge	no_collect_4579
+	att_call	collect_0
+no_collect_4579:
+	mov	rcx,rdi 
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax
+	lea	rdx,REAL+2[rip]
+	mov	qword ptr 16[rdi],rdx
+	add	rdi,24
+	att_jmp	st_fillr_array
+fillr_array:
+	mov	qword ptr [rdi],rbx 
+	add	rdi,8
+st_fillr_array:
+	sub	rax,1
+	att_jnc	fillr_array
+
+	ret
+
+create_array:
+	lea	rbp,3[rax]
+	sub	r15,rbp
+	jge	no_collect_4576
+	att_call	collect_1
+no_collect_4576:
+	mov	rbx,rcx 
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax 
+	mov	qword ptr 16[rdi],0
+	add	rdi,24
+
+	jmp	fillr1_array
+
+
+
+
+/* in rax: number of elements, rbx: element descriptor */
+/* r10 : element size, r11 : element a size -> rcx : array */
+
+create_R_array:
+	sub	r10,2
+	jc	create_R_array_1
+	je	create_R_array_2
+	sub	r10,2
+	jc	create_R_array_3
+	je	create_R_array_4
+	jmp	create_R_array_5
+
+create_R_array_1:
+	lea	rbp,3[rax]
+	sub	r15,rbp
+	jge	no_collect_4581
+	att_call	collect_0
+no_collect_4581:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax 
+	mov	qword ptr 16[rdi],rbx 
+	add	rdi,24
+
+	test	r11,r11
+	je	r_array_1_b
+
+	mov	rbx,qword ptr (-8)[rsi]
+	att_jmp	fillr1_array
+
+r_array_1_b:
+	mov	rbx,qword ptr 8[rsp]
+
+fillr1_array:
+	mov	rdx,rax 
+	shr	rax,1
+	test	dl,1
+	je	st_fillr1_array_1
+
+	mov	qword ptr [rdi],rbx
+	add	rdi,8
+	att_jmp	st_fillr1_array_1
+
+fillr1_array_lp:
+	mov	qword ptr [rdi],rbx 
+	mov	qword ptr 8[rdi],rbx 
+	add	rdi,16
+st_fillr1_array_1:
+	sub	rax,1
+	att_jnc	fillr1_array_lp
+
+	ret
+
+create_R_array_2:
+	lea	rbp,3[rax*2]
+	sub	r15,rbp
+	jge	no_collect_4582
+	att_call	collect_0
+no_collect_4582:
+	mov	rcx,rdi 
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax 
+	mov	qword ptr 16[rdi],rbx 
+	add	rdi,24
+
+	sub	r11,1
+	jc	r_array_2_bb
+	je	r_array_2_ab
+r_array_2_aa:
+	mov	rbx,qword ptr (-8)[rsi]
+	mov	rbp,qword ptr (-16)[rsi]
+	jmp	st_fillr2_array
+r_array_2_ab:
+	mov	rbx,qword ptr (-8)[rsi]
+	mov	rbp,qword ptr 8[rsp]
+	att_jmp	st_fillr2_array
+r_array_2_bb:
+	mov	rbx,qword ptr 8[rsp]
+	mov	rbp,qword ptr 16[rsp]
+	att_jmp	st_fillr2_array
+
+fillr2_array_1:
+	mov	qword ptr [rdi],rbx 
+	mov	qword ptr 8[rdi],rbp 
+	add	rdi,16
+st_fillr2_array:
+	sub	rax,1
+	att_jnc	fillr2_array_1
+
+	ret
+
+create_R_array_3:
+	lea	rbp,3[rax+rax*2]
+	sub	r15,rbp
+	jge	no_collect_4583
+	att_call	collect_0
+no_collect_4583:
+	mov	rcx,rdi
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax 
+	mov	qword ptr 16[rdi],rbx 
+	add	rdi,24
+
+	pop	rdx 
+	mov	r12,rsp 
+
+	test	r11,r11 
+	je	r_array_3
+	
+	lea	r13,0[r11*8]
+	mov	rbp,rsi
+	sub	rbp,r13
+
+	sub	r11,1
+
+copy_a_to_b_lp3:
+	push	[rbp]
+	add	rbp,8
+	sub	r11,1
+	att_jnc	copy_a_to_b_lp3
+
+r_array_3:
+	mov	rbx,qword ptr [rsp]
+	mov	r13,qword ptr 8[rsp]
+	mov	rbp,qword ptr 16[rsp]
+	
+	mov	rsp,r12
+	push	rdx
+
+	jmp	st_fillr3_array
+
+fillr3_array_1:
+	mov	qword ptr [rdi],rbx 
+	mov	qword ptr 8[rdi],r13
+	mov	qword ptr 16[rdi],rbp 
+	add	rdi,24
+st_fillr3_array:
+	sub	rax,1
+	att_jnc	fillr3_array_1
+
+	ret
+
+create_R_array_4:
+	lea	rbp,3[rax+4]
+	sub	r15,rbp
+	jge	no_collect_4584
+	att_call	collect_0
+no_collect_4584:
+	mov	rcx,rdi 
+	lea	rdx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rdx
+	mov	qword ptr 8[rdi],rax 
+	mov	qword ptr 16[rdi],rbx 
+	add	rdi,24
+
+	pop	rdx 
+	mov	r12,rsp 
+
+	test	r11,r11 
+	je	r_array_4
+
+	lea	r13,0[r11*8]
+	mov	rbp,rsi
+	sub	rbp,r13
+	sub	r11,1
+
+copy_a_to_b_lp4:
+	push	[rbp]
+	add	rbp,8
+	sub	r11,1
+	att_jnc	copy_a_to_b_lp4
+
+r_array_4:
+	mov	rbx,qword ptr [rsp]
+	mov	r13,qword ptr 8[rsp]
+	mov	r14,qword ptr 16[rsp]
+	mov	rbp,qword ptr 24[rsp]
+	
+	mov	rsp,r12
+	push	rdx
+	
+	jmp	st_fillr4_array
+
+fillr4_array:
+	mov	qword ptr [rdi],rbx 
+	mov	qword ptr 8[rdi],r13 
+	mov	qword ptr 16[rdi],r14 
+	mov	qword ptr 24[rdi],rbp 
+	add	rdi,32
+st_fillr4_array:
+	sub	rax,1
+	att_jnc	fillr4_array
+
+	ret
+
+create_R_array_5:
+	lea	r12,4[r10]
+	mov	rbp,rax
+	imul	rbp,r12
+	add	rbp,3
+	sub	r15,rbp
+	jge	no_collect_4585
+	att_call	collect_0
+no_collect_4585:
+	lea	rcx,__ARRAY__+2[rip]
+	mov	qword ptr [rdi],rcx
+	mov	qword ptr 8[rdi],rax
+	mov	qword ptr 16[rdi],rbx
+	mov	rcx,rdi
+	add	rdi,24
+
+	pop	rdx
+	mov	r12,rsp 
+	
+	test	r11,r11 
+	je	r_array_5
+
+	lea	r13,0[r11*8]
+	mov	rbp,rsi
+	sub	rbp,r13
+	sub	r11,1
+
+copy_a_to_b_lp5:
+	push	[rbp]
+	add	rbp,8
+	sub	r11,1
+	att_jnc	copy_a_to_b_lp5
+
+r_array_5:
+	mov	r13,qword ptr [rsp]
+	mov	r14,qword ptr 8[rsp]
+	mov	r8,qword ptr 16[rsp]
+	mov	r9,qword ptr 24[rsp]
+	add	rsp,32
+
+	sub	r10,1
+	jmp	st_fillr5_array
+
+fillr5_array_1:
+	mov	qword ptr [rdi],r13 
+	mov	qword ptr 8[rdi],r14
+
+	mov	r11,rsp
+	mov	rbx,r10 
+
+	mov	qword ptr 16[rdi],r8
+	mov	qword ptr 24[rdi],r9
+	add	rdi,32
+
+copy_elem_lp5:
+	mov	rbp,qword ptr [r11]
+	add	r11,8
+	mov	qword ptr [rdi],rbp 
+	add	rdi,8
+	sub	rbx,1
+	att_jnc	copy_elem_lp5
+
+st_fillr5_array:
+	sub	rax,1
+	att_jnc	fillr5_array_1
+
+	mov	rsp,r12
+	jmp	rdx 
+
+ .if ! NEW_DESCRIPTORS
+yet_args_needed:
+/* for more than 4 arguments */
+	mov	r10,[rdx]
+	movzx	rax,word ptr (-2)[r10]
+	add	rax,3
+	sub	r15,rax
+	jl	gc_1
+gc_r_1:	sub	rax,3+1+4
+	mov	rbx,8[rdx]
+	add	r10,8
+	mov	rdx,16[rdx]
+	mov	rbp,rdi
+	mov	r8,[rdx]
+	mov	[rdi],r8 
+	mov	r8,8[rdx]
+	mov	8[rdi],r8 
+	mov	r8,16[rdx]
+	mov	16[rdi],r8 
+	add	rdx,24
+	add	rdi,24
+
+cp_a:	mov	r8,[rdx]
+	add	rdx,8
+	mov	[rdi],r8 
+	add	rdi,8
+	sub	rax,1
+	jge	cp_a
+
+	mov	[rdi],rcx 
+	mov	8[rdi],r10 
+	lea	rcx,8[rdi]
+	mov	16[rdi],rbx 
+	mov	24[rdi],rbp 
+	add	rdi,32
+	ret
+
+gc_1:
+	call	collect_2
+	jmp	gc_r_1
+
+yet_args_needed_0:
+	sub	r15,2
+	jl	gc_20
+gc_r_20:
+	mov	8[rdi],rcx 
+	mov	rax,[rdx]
+	mov	rcx,rdi
+	add	rax,8
+	mov	[rdi],rax 
+	add	rdi,16
+	ret
+
+gc_20:
+	call	collect_2
+	jmp	gc_r_20
+
+yet_args_needed_1:
+	sub	r15,3
+	jl	gc_21
+gc_r_21:
+	mov	16[rdi],rcx 
+	mov	rax,[rdx]
+	mov	rcx,rdi
+	add	rax,8
+	mov	[rdi],rax 
+	mov	rbx,8[rdx]
+	mov	8[rdi],rbx 
+	add	rdi,24
+	ret
+
+gc_21:
+	call	collect_2
+	jmp	gc_r_21
+
+yet_args_needed_2:
+	sub	r15,5
+	jl	gc_22
+gc_r_22:
+	mov	rax,[rdx]
+	mov	8[rdi],rcx 
+	add	rax,8
+	mov	rbp,8[rdx]
+	mov	16[rdi],rax 
+	lea	rcx,16[rdi]
+	mov	24[rdi],rbp 
+	mov	rbp,16[rdx]
+	mov	[rdi],rbp 
+	mov	32[rdi],rdi 
+	add	rdi,40
+	ret
+
+gc_22:
+	call	collect_2
+	jmp	gc_r_22
+
+yet_args_needed_3:
+	sub	r15,6
+	jl	gc_23
+gc_r_23:
+	mov	rax,[rdx]
+	mov	16[rdi],rcx 
+	add	rax,8
+	mov	rbp,8[rdx]
+	mov	24[rdi],rax 
+	mov	rdx,16[rdx]
+	mov	32[rdi],rbp 
+	mov	rbp,[rdx]
+	mov	40[rdi],rdi 
+	mov	[rdi],rbp 
+	mov	rbp,8[rdx]
+	lea	rcx,24[rdi]
+	mov	8[rdi],rbp 
+	add	rdi,48
+	ret
+
+gc_23:
+	call	collect_2
+	jmp	gc_r_23
+
+yet_args_needed_4:
+	sub	r15,7
+	jl	gc_24
+gc_r_24:
+	mov	rax,[rdx]
+	mov	24[rdi],rcx 
+	add	rax,8
+	mov	rbp,8[rdx]
+	mov	32[rdi],rax 
+	mov	rdx,16[rdx]
+	mov	40[rdi],rbp 
+	mov	rbp,[rdx]
+	mov	48[rdi],rdi 
+	mov	[rdi],rbp 
+	mov	rbp,8[rdx]
+	lea	rcx,32[rdi]
+	mov	8[rdi],rbp 
+	mov	rbp,16[rdx ]
+	mov	16[rdi],rbp 
+	add	rdi,56
+	ret
+
+gc_24:
+	call	collect_2
+	jmp	gc_r_24
+ .endif
+
+repl_args_b:
+	test	rax,rax 
+	jle	repl_args_b_1
+
+	dec	rax 
+	je	repl_args_b_4
+
+	mov	rdx,16[rcx]
+	sub	rbx,2
+	jne	repl_args_b_2
+
+	mov	[rsi],rdx 
+	add	rsi,8
+	att_jmp	repl_args_b_4
+
+repl_args_b_2:
+	lea	rdx,[rdx+rax*8]
+
+repl_args_b_3:
+	mov	rbp,(-8)[rdx]
+	sub	rdx,8
+	mov	[rsi],rbp 
+	add	rsi,8
+	dec	rax 
+	att_jne	repl_args_b_3
+
+repl_args_b_4:
+	mov	rbp,8[rcx]
+	mov	[rsi],rbp 
+	add	rsi,8
+repl_args_b_1:
+	ret
+
+push_arg_b:
+	cmp	rbx,2
+	jb	push_arg_b_1
+	jne	push_arg_b_2
+	cmp	rbx,rax 
+	att_je	push_arg_b_1
+push_arg_b_2:
+	mov	rcx,16[rcx]
+	sub	rbx,2
+push_arg_b_1:
+	mov	rcx,[rcx+rbx*8]
+	ret
+
+del_args:
+	mov	rbx,[rcx]
+	sub	rbx,rax 
+	movsx	rax,word ptr (-2)[rbx]
+	sub	rax,2
+	jge	del_args_2
+
+	mov	[rdx],rbx 
+	mov	rbp,8[rcx]
+	mov	8[rdx],rbp 
+	mov	rbp,16[rcx]
+	mov	16[rdx],rbp 
+	ret
+
+del_args_2:
+	jne	del_args_3
+
+	mov	[rdx],rbx 
+	mov	rbp,8[rcx]
+	mov	8[rdx],rbp 
+	mov	rbp,16[rcx]
+	mov	rbp,[rbp]
+	mov	16[rdx],rbp 
+	ret
+
+del_args_3:
+	sub	r15,rax
+	jl	del_args_gc
+del_args_r_gc:
+	mov	[rdx],rbx 
+	mov	16[rdx],rdi 
+	mov	rbp,8[rcx]
+	mov	rcx,16[rcx]
+	mov	8[rdx],rbp 
+
+del_args_copy_args:
+	mov	rbp,[rcx]
+	add	rcx,8
+	mov	[rdi],rbp 
+	add	rdi,8
+	sub	rax,1
+	att_jg	del_args_copy_args
+
+	ret
+
+del_args_gc:
+	att_call	collect_2
+	att_jmp	del_args_r_gc
+
+ .if USE_LIBM
+cos_real:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	call	cos
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	ret
+
+sin_real:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	call	sin
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	ret
+
+tan_real:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	call	tan
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	ret
+
+atan_real:
+	mov	rbp,rsp
+	and	rsp,-16
+	mov	r13,rsi
+	mov	r14,rdi
+	call	atan
+	mov	rsp,rbp
+	mov	rsi,r13
+	mov	rdi,r14
+	ret
+
+asin_real:
+acos_real:
+ln_real:
+log10_real:
+exp_real:
+pow_real:
+exp2_real_:
+_c_log10:
+_c_pow:
+_c_entier:
+	int 3
+	ret
+ .endif
+	
+entier_real:
+	cvttsd2si rax,xmm0
+	ucomisd	xmm0,qword ptr real_0_0[rip]
+	jb	entier_real_m
+	ret
+
+entier_real_m:
+	movsd	qword ptr (-8)[rsp],xmm0
+	mov	rcx,qword ptr (-8)[rsp]
+	mov	rbx,rcx
+	shr rcx,52
+	cmp rcx,0x0bff
+	jb	entier_real_m_small
+	cmp	rcx,0x0bff+52
+	jae	entier_real_m_large
+	sub	rcx,0x0bff-12
+	shl	rbx,cl
+	je	entier_m_exact
+entier_real_m_small:
+	sub	rax,1
+entier_real_m_large:
+entier_m_exact:
+	ret
+
+r_to_i_real:
+	cvtsd2si rax,xmm0
+	ret
+
+	.globl	getheapend
+
+getheapend:
+	lea	rbx,[rdi+r15*8]
+	mov	rax,heap_end_after_gc[rip]
+	ret
+
+	.include	"areals.s"
+
+ .if PROFILE
+  .if TRACE
+	.include	"atrace.s"
+  .else
+	.include	"aprofile.s"
+  .endif
+ .endif
+
+ .if NEW_DESCRIPTORS
+	.include	"aap.s"
+ .endif
+
diff --git a/macho64/atrace.s b/macho64/atrace.s
new file mode 100644
index 0000000..4f9bf6d
--- /dev/null
+++ b/macho64/atrace.s
@@ -0,0 +1,474 @@
+
+	.text
+
+	.globl	init_profiler
+	.globl	profile_r
+	.globl profile_l
+	.globl profile_l2
+	.globl profile_n
+	.globl profile_n2
+	.globl profile_s
+	.globl profile_s2
+	.globl profile_t
+	.globl	write_profile_stack
+	.globl	stack_trace_depth
+
+ .if ! LINUX
+	.globl	allocate_memory
+ .endif
+	.globl	__STRING__
+	.globl	_ab_stack_size
+	.globl	_ew_print_string
+	.globl	_ew_print_char
+	.globl	_ew_print_text
+/*	.globl	print_error */
+/*	.globl	profile_stack_pointer */
+
+next		= 0
+name_		= 8
+FunctionProfile	= 16
+
+profile_t:
+	sub	qword ptr profile_stack_pointer[rip],8
+	ret
+
+profile_r:
+	sub	qword ptr profile_stack_pointer[rip],8
+	ret
+
+profile_l:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+
+	test	rbx,rbx
+	je	allocate_function_profile_record_l
+allocate_function_profile_record_lr:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	add	rbp,8
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_l:
+	call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_lr
+
+profile_l2:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+
+	test	rbx,rbx
+	je	allocate_function_profile_record_l2
+allocate_function_profile_record_l2r:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	mov	qword ptr 8[rbp],rbx
+	add	rbp,16
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_l2:
+	att_call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_l2r
+
+profile_n:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+	
+	test	rbx,rbx
+	je	allocate_function_profile_record_n
+allocate_function_profile_record_nr:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	add	rbp,8
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_n:
+	att_call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_nr
+
+profile_n2:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+
+	test	rbx,rbx
+	je	allocate_function_profile_record_n2
+allocate_function_profile_record_n2r:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	mov	qword ptr 8[rbp],rbx
+	add	rbp,16
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_n2:
+	att_call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_n2r
+
+profile_s2:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+	
+	test	rbx,rbx
+	je	allocate_function_profile_record_s2
+allocate_function_profile_record_s2r:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	mov	qword ptr 8[rbp],rbx
+	add	rbp,16
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_s2:
+	att_call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_s2r
+
+profile_s:
+	push	rbx
+	mov	rbx,qword ptr [rbp]
+	
+	test	rbx,rbx
+	je	allocate_function_profile_record_s
+allocate_function_profile_record_sr:
+	mov	rbp,qword ptr profile_stack_pointer[rip]
+
+	mov	qword ptr [rbp],rbx
+	add	rbp,8
+	mov	qword ptr profile_stack_pointer[rip],rbp
+
+	pop	rbx
+	ret
+
+allocate_function_profile_record_s:
+	att_call	allocate_function_profile_record
+	att_jmp	allocate_function_profile_record_sr
+
+/* argument: rbp: function name adress-4 */
+/* result:   rbx: function profile record adress */
+
+allocate_function_profile_record:
+	push	rax
+	mov	rax,qword ptr global_n_free_records_in_block[rip]
+	mov	rbx,qword ptr global_last_allocated_block[rip]
+
+	test	rax,rax
+	jne	no_alloc
+
+	push	rcx
+	push	rdx
+	push	rbp
+
+ .if LINUX
+	sub	rsp,104
+	mov	qword ptr [rsp],rsi
+	mov	qword ptr 8[rsp],rdi
+	mov	qword ptr 16[rsp],r8
+	mov	qword ptr 24[rsp],r10
+	mov	qword ptr 32[rsp],r11
+	movsd	qword ptr 40[rsp],xmm0
+	movsd	qword ptr 48[rsp],xmm1
+	movsd	qword ptr 56[rsp],xmm2
+	movsd	qword ptr 64[rsp],xmm3
+	movsd	qword ptr 72[rsp],xmm4
+	movsd	qword ptr 80[rsp],xmm5
+	movsd	qword ptr 88[rsp],xmm6
+	movsd	qword ptr 96[rsp],xmm7
+ .else
+	sub	rsp,72
+	mov	qword ptr [rsp],r8
+	mov	qword ptr 8[rsp],r10
+	mov	qword ptr 16[rsp],r11
+	movsd	qword ptr 24[rsp],xmm0
+	movsd	qword ptr 32[rsp],xmm1
+	movsd	qword ptr 40[rsp],xmm2
+	movsd	qword ptr 48[rsp],xmm3
+	movsd	qword ptr 56[rsp],xmm4
+	movsd	qword ptr 64[rsp],xmm5
+ .endif
+
+	mov	rbp,rsp
+	sub	rsp,40
+	and	rsp,-16
+ .if LINUX
+	mov	rdi,8192
+			/* 512*FunctionProfile */
+	att_call	_malloc
+ .else
+	mov	rcx,512*FunctionProfile
+	call	allocate_memory
+ .endif
+	mov	rsp,rbp
+
+ .if LINUX
+	mov	rsi,qword ptr [rsp]
+	mov	rdi,qword ptr 8[rsp]
+	mov	r8,qword ptr 16[rsp]
+	mov	r10,qword ptr 24[rsp]
+	mov	r11,qword ptr 32[rsp]
+	movlpd	xmm0,qword ptr 40[rsp]
+	movlpd	xmm1,qword ptr 48[rsp]
+	movlpd	xmm2,qword ptr 56[rsp]
+	movlpd	xmm3,qword ptr 64[rsp]
+	movlpd	xmm4,qword ptr 72[rsp]
+	movlpd	xmm5,qword ptr 80[rsp]
+	movlpd	xmm6,qword ptr 88[rsp]
+	movlpd	xmm7,qword ptr 96[rsp]
+	add	rsp,104
+ .else
+	mov	r8,qword ptr [rsp]
+	mov	r10,qword ptr 8[rsp]
+	mov	r11,qword ptr 16[rsp]
+	movlpd	xmm0,qword ptr 24[rsp]
+	movlpd	xmm1,qword ptr 32[rsp]
+	movlpd	xmm2,qword ptr 40[rsp]
+	movlpd	xmm3,qword ptr 48[rsp]
+	movlpd	xmm4,qword ptr 56[rsp]
+	movlpd	xmm5,qword ptr 64[rsp]
+	add	rsp,72
+ .endif
+
+	test	rax,rax
+
+	pop	rbp
+	pop	rdx
+	pop	rcx
+
+	je	no_memory
+
+	mov	rbx,rax
+	mov	rax,512
+	mov	qword ptr global_last_allocated_block[rip],rbx
+
+no_alloc:	
+	dec	rax
+	mov	qword ptr global_n_free_records_in_block[rip],rax
+	lea	rax,FunctionProfile[rbx]
+	mov	qword ptr global_last_allocated_block[rip],rax
+
+	mov	rax,qword ptr global_profile_records[rip]
+	mov	qword ptr name_[rbx],rbp
+
+	mov	qword ptr next[rbx],rax
+	mov	qword ptr global_profile_records[rip],rbx
+
+	mov	qword ptr [rbp],rbx
+	pop	rax
+	ret
+
+no_memory:
+	lea	rbp,not_enough_memory_for_profiler[rip]
+	pop	rax
+	att_jmp	print_error
+
+write_profile_stack:
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+ .endif
+	mov	rax,qword ptr profile_stack_pointer[rip]
+
+	test	rax,rax
+	je	stack_not_initialised
+
+	push	rax
+
+	mov	rbp,rsp
+	sub	rsp,40
+	and	rsp,-16
+ .if LINUX
+	lea	rdi,stack_trace_string[rip]
+ .else
+	lea	rcx,stack_trace_string
+ .endif
+	att_call	_ew_print_string
+	mov	rsp,rbp
+
+	pop	rax
+
+/*	mov	rbp,12 */
+	mov	rbp,qword ptr stack_trace_depth[rip]
+write_functions_on_stack:
+	mov	rbx,qword ptr (-8)[rax]
+	sub	rax,8
+
+	test	rbx,rbx
+	je	end_profile_stack
+
+	push	rax
+	mov	rcx,qword ptr name_[rbx]
+
+	push	rbp
+
+ .if LINUX
+	movsx	rdx,dword ptr (-4)[rcx]
+	lea	rdx,-4[rcx+rdx]
+	lea	rdi,8[rcx]
+	mov	r12,rdx
+ .else
+	mov	edx,dword ptr (-4)[rcx]
+	add	rcx,8
+
+	mov	r12d,dword ptr [rdx]
+	lea	r13,4[rdx]
+ .endif
+
+	mov	rbp,rsp
+	sub	rsp,40
+	and	rsp,-16
+
+	att_call	_ew_print_string
+
+ .if LINUX
+	lea	rdi,module_string[rip]
+ .else
+	lea	rcx,module_string
+ .endif
+	att_call	_ew_print_string
+
+ .if LINUX
+	mov	esi,dword ptr [r12]
+	lea	rdi,4[r12]
+ .else
+	mov	rdx,r12
+	mov	rcx,r13
+ .endif
+	att_call	_ew_print_text
+
+ .if LINUX
+	mov	rdi,']
+ .else
+	mov	rcx,']
+ .endif
+	att_call	_ew_print_char
+
+ .if LINUX
+	mov	rdi,10
+ .else
+	mov	rcx,10
+ .endif
+	att_call	_ew_print_char
+
+	mov	rsp,rbp
+
+	pop	rbp
+	pop	rax
+
+	sub	rbp,1
+	att_jne	write_functions_on_stack
+
+end_profile_stack:
+stack_not_initialised:
+ .if LINUX
+	mov	rsi,r13
+	mov	rdi,r14
+ .endif
+	ret
+
+init_profiler:
+	mov	rbp,rsp
+	sub	rsp,40
+	and	rsp,-16
+ .if LINUX
+	mov	r13,rsi
+	mov	r14,rdi
+	mov	rdi,qword ptr _ab_stack_size[rip]
+	att_call	_malloc
+	mov	rsi,r13
+	mov	rdi,r14
+ .else
+ 	mov	rcx,qword ptr ab_stack_size
+	call	allocate_memory
+ .endif
+	mov	rsp,rbp
+	
+	test	rax,rax
+	je	init_profiler_error
+
+	push	rax
+
+	lea	rbp,start_string[rip]
+	att_call	allocate_function_profile_record
+
+	pop	rdx
+
+	mov	qword ptr 8[rdx],rbx
+	mov	qword ptr [rdx],0
+	add	rdx,16
+	mov	qword ptr profile_stack_pointer[rip],rdx
+	ret
+
+init_profiler_error:
+	mov	qword ptr profile_stack_pointer[rip],0
+	lea	rbp,not_enough_memory_for_profile_stack[rip]
+	att_jmp	print_error
+
+
+
+	.data
+
+	.align 8
+
+global_n_free_records_in_block:
+	.quad 0
+/* 0 n free records in block */
+global_last_allocated_block:
+	.quad 0
+/* 8 latest allocated block */
+global_profile_records:
+	.quad 0
+/* 16 profile record list */
+
+stack_trace_depth:
+	.quad	12
+	.align	8
+
+/* m_system also defined in istartup.s */
+/*
+m_system:
+	.quad	6
+	.ascii	"System"
+	.byte	0
+	.byte	0
+*/
+	.long	m_system-.
+start_string:
+	.quad	0
+	.ascii	"start"
+	.byte	0
+	.align	8
+not_enough_memory_for_profile_stack:
+	.ascii	"not enough memory for profile stack"
+	.byte	10
+	.byte	0
+not_enough_memory_for_profiler:
+	.ascii	"not enough memory for profiler"
+	.byte	10
+	.byte	0
+stack_trace_string:
+	.ascii	"Stack trace:"
+	.byte	10
+	.byte	0
+module_string:
+	.ascii	" [module: "
+	.byte	0
+	.align	8
+
+
+
+/*	end */
diff --git a/macho64/startup.s b/macho64/startup.s
new file mode 100644
index 0000000..b5713f2
--- /dev/null
+++ b/macho64/startup.s
@@ -0,0 +1,6 @@
+
+        .set PROFILE,0
+        .set TRACE,0
+
+	.include "astartup.s"
+
diff --git a/macho64/startupTrace.s b/macho64/startupTrace.s
new file mode 100644
index 0000000..1738171
--- /dev/null
+++ b/macho64/startupTrace.s
@@ -0,0 +1,6 @@
+
+        .set PROFILE,1
+        .set TRACE,1
+
+	.include "astartup.s"
+