IQOLanczosResizerImpl_SSE4_1.clang.s

# clang++-4.0 -I/home/yoffy/src/libiqo/include -I/home/yoffy/src/libiqo/src/../include -Wall -Wextra -Wconversion -Wno-sign-conversion -std=c++98 -fno-exceptions  -Wall -Wextra -Wconversion -Wno-sign-conversion  -Ofast   -march=core2 -msse4.1 -mtune=westmere -S -mllvm --x86-asm-syntax=intel /home/yoffy/src/libiqo/src/IQOLanczosResizerImpl_SSE4_1.cpp

.LBB7_11:                               # =>This Loop Header: Depth=1
                                        #     Child Loop BB7_12 Depth 2
	mov	rax, qword ptr [rsp - 16] # 8-byte Reload
	movdqu	xmm1, xmmword ptr [rax + 4*r15]
	movdqu	xmm2, xmmword ptr [rax + 4*r15 + 16]
	mov	rax, qword ptr [rsp - 8] # 8-byte Reload
	lea	rdi, [rax + 4*r12]
	xorps	xmm0, xmm0
	mov	ebp, dword ptr [rsp - 36] # 4-byte Reload
	mov	r11, rbx
	mov	r13d, ebx
	xorps	xmm3, xmm3
	.p2align	4, 0x90
.LBB7_12:                               #   Parent Loop BB7_11 Depth=1
                                        # =>  This Inner Loop Header: Depth=2
	movd	xmm4, ebp
	pshufd	xmm4, xmm4, 0           # xmm4 = xmm4[0,0,0,0]
	movdqa	xmm5, xmm4
	paddd	xmm5, xmm1
	paddd	xmm4, xmm2
	movq	r8, xmm5
	pextrq	r9, xmm5, 1
	movsxd	r10, r8d
	sar	r8, 32
	movsxd	r14, r9d
	sar	r9, 32
	movss	xmm5, dword ptr [rsi + 4*r10] # xmm5 = mem[0],zero,zero,zero
	insertps	xmm5, dword ptr [rsi + 4*r8], 16 # xmm5 = xmm5[0],mem[0],xmm5[2,3]
	insertps	xmm5, dword ptr [rsi + 4*r14], 32 # xmm5 = xmm5[0,1],mem[0],xmm5[3]
	insertps	xmm5, dword ptr [rsi + 4*r9], 48 # xmm5 = xmm5[0,1,2],mem[0]
	movq	rbx, xmm4
	pextrq	rcx, xmm4, 1
	movsxd	rdx, ebx
	sar	rbx, 32
	movsxd	rax, ecx
	sar	rcx, 32
	movss	xmm4, dword ptr [rsi + 4*rdx] # xmm4 = mem[0],zero,zero,zero
	insertps	xmm4, dword ptr [rsi + 4*rbx], 16 # xmm4 = xmm4[0],mem[0],xmm4[2,3]
	insertps	xmm4, dword ptr [rsi + 4*rax], 32 # xmm4 = xmm4[0,1],mem[0],xmm4[3]
	insertps	xmm4, dword ptr [rsi + 4*rcx], 48 # xmm4 = xmm4[0,1,2],mem[0]
	mulps	xmm5, xmmword ptr [rdi - 16]
	mulps	xmm4, xmmword ptr [rdi]
	addps	xmm0, xmm5
	addps	xmm3, xmm4
	add	rdi, 32
	inc	ebp
	dec	r13d
	jne	.LBB7_12
# BB#13:                                #   in Loop: Header=BB7_11 Depth=1
	add	r12, qword ptr [rsp - 24] # 8-byte Folded Reload
	roundps	xmm0, xmm0, 8
	roundps	xmm1, xmm3, 8
	cvtps2dq	xmm0, xmm0
	cvtps2dq	xmm1, xmm1
	packusdw	xmm0, xmm1
	packuswb	xmm0, xmm0
	mov	rax, qword ptr [rsp - 48] # 8-byte Reload
	movq	qword ptr [rax + r15], xmm0
	cmp	r12, qword ptr [rsp - 56] # 8-byte Folded Reload
	mov	eax, 0
	cmove	r12, rax
	add	r15, 8
	cmp	r15, qword ptr [rsp - 32] # 8-byte Folded Reload
	mov	rbx, r11
	jl	.LBB7_11

IQOLanczosResizerImpl_SSE4_1.vc.asm

; "C:\Program Files (x86)\Microsoft Visual Studio\2017\Community\VC\Tools\MSVC\14.11.25503\bin\HostX86\x64\CL.exe" /c /IC:\Users\yoffy\src\libiqo\include /IC:\Users\yoffy\src\libiqo\src\..\include /Zi /nologo /W3 /WX- /Ox /Ob2 /D WIN32 /D _WINDOWS /D NDEBUG /D "CMAKE_INTDIR=\"Release\"" /D _MBCS /Gm- /EHsc /MT /GS- /fp:fast /Zc:wchar_t /Zc:forScope /Zc:inline /GR /Gd /TP /FAsu C:\Users\yoffy\src\libiqo\src\IQOLanczosResizerImpl_SSE4_1.cpp

$LL4@resizeXmai:
	movdqu	xmm8, XMMWORD PTR [r11]
	xor	ebp, ebp
	movaps	xmm5, xmm7
	movaps	xmm6, xmm7
	movdqu	xmm9, XMMWORD PTR [r11+rax]
	test	ebx, ebx
	jle	$LN6@resizeXmai
	lea	rsi, QWORD PTR [r12+rcx*4]
	mov	eax, ebx
	cdq
	lea	rcx, QWORD PTR [rcx+rbx*8]
	sub	eax, edx
	sar	eax, 1
	neg	eax
	mov	r12d, eax
	npad	1
$LL7@resizeXmai:
	lea	edx, DWORD PTR [r12+rbp]
	inc	ebp
	movd	xmm1, edx
	pshufd	xmm1, xmm1, 0
	movdqa	xmm0, xmm1
	paddd	xmm1, xmm9
	paddd	xmm0, xmm8
	movq	r10, xmm1
	pextrq	rdx, xmm0, 1
	movq	r8, xmm0
	mov	rax, rdx
	shr	rax, 32					; 00000020H
	cdqe
	pextrq	r9, xmm1, 1
	movss	xmm4, DWORD PTR [r14+rax*4]
	movsxd	rax, edx
	movss	xmm3, DWORD PTR [r14+rax*4]
	mov	rax, r8
	shr	rax, 32					; 00000020H
	cdqe
	movss	xmm0, DWORD PTR [r14+rax*4]
	movsxd	rax, r8d
	movss	xmm2, DWORD PTR [r14+rax*4]
	mov	rax, r9
	insertps xmm2, xmm0, 16
	insertps xmm2, xmm3, 32				; 00000020H
	shr	rax, 32					; 00000020H
	insertps xmm2, xmm4, 48				; 00000030H
	mulps	xmm2, XMMWORD PTR [rsi]
	cdqe
	movss	xmm4, DWORD PTR [r14+rax*4]
	addps	xmm5, xmm2
	movsxd	rax, r9d
	movss	xmm3, DWORD PTR [r14+rax*4]
	mov	rax, r10
	shr	rax, 32					; 00000020H
	cdqe
	movss	xmm0, DWORD PTR [r14+rax*4]
	movsxd	rax, r10d
	movss	xmm2, DWORD PTR [r14+rax*4]
	insertps xmm2, xmm0, 16
	insertps xmm2, xmm3, 32				; 00000020H
	insertps xmm2, xmm4, 48				; 00000030H
	mulps	xmm2, XMMWORD PTR [rsi+16]
	add	rsi, 32					; 00000020H
	addps	xmm6, xmm2
	cmp	ebp, ebx
	jl	$LL7@resizeXmai
	mov	r12, QWORD PTR coefs$1$[rsp]
$LN6@resizeXmai:
	xor	eax, eax
	add	r11, 32					; 00000020H
	roundps	xmm0, xmm6, 8
	roundps	xmm1, xmm5, 8
	cvtps2dq xmm2, xmm0
	cvtps2dq xmm3, xmm1
	packusdw xmm3, xmm2
	packuswb xmm3, xmm3
	movq	QWORD PTR [rdi], xmm3
	add	rdi, 8
	cmp	rcx, r13
	cmovne	rax, rcx
	mov	rcx, rax
	mov	rax, QWORD PTR tv965[rsp]
	sub	r15, 1
	jne	$LL4@resizeXmai