idx
int64 0
2.11M
| name
stringlengths 1
118k
| code
stringlengths 6
516k
| asm
stringlengths 21
4.64M
| file
stringlengths 39
143
| opt
stringclasses 1
value | path
stringlengths 20
133
|
|---|---|---|---|---|---|---|
2,113,100 |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_avx_functor::binary_op_max>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_2_3_4_20(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = b.w;
int h = b.h;
int d = b.d;
int channels = b.c;
int elempack = b.elempack;
int size = w * h * d * elempack;
// type 2 3 4 20
c.create_like(b, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float a0 = a[0];
const float* ptr = b.channel(q);
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
__m512 _a0_avx512 = _mm512_set1_ps(a0);
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _outp = op.func_pack16(_a0_avx512, _p);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
outptr += 16;
}
#endif // __AVX512F__
__m256 _a0_avx = _mm256_set1_ps(a0);
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _outp = op.func_pack8(_a0_avx, _p);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
outptr += 8;
}
#endif // __AVX__
__m128 _a0 = _mm_set1_ps(a0);
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _outp = op.func_pack4(_a0, _p);
_mm_store_ps(outptr, _outp);
ptr += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(a0, *ptr);
ptr += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x460, %rsp # imm = 0x460
movq %rdi, 0x200(%rsp)
movq %rsi, 0x1f8(%rsp)
movq %rdx, 0x1f0(%rsp)
movq %rcx, 0x1e8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1e0(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1dc(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x1d8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x1d4(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x1d0(%rsp)
movl 0x1e0(%rsp), %eax
imull 0x1dc(%rsp), %eax
imull 0x1d8(%rsp), %eax
imull 0x1d0(%rsp), %eax
movl %eax, 0x1cc(%rsp)
movq 0x1f0(%rsp), %rdi
movq 0x1f8(%rsp), %rsi
movq 0x1e8(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x1f0(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rcx
movq %rcx, 0x60(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x6f(%rsp)
je 0x162f925
movq 0x60(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x6f(%rsp)
movb 0x6f(%rsp), %al
testb $0x1, %al
jne 0x162f92f
jmp 0x162f93f
movl $0xffffff9c, 0x20c(%rsp) # imm = 0xFFFFFF9C
jmp 0x1630564
movl $0x0, 0x1c8(%rsp)
movl 0x1c8(%rsp), %eax
cmpl 0x1d4(%rsp), %eax
jge 0x1630559
movq 0x200(%rsp), %rax
movq %rax, 0x3d0(%rsp)
movq $0x0, 0x3c8(%rsp)
movq 0x3d0(%rsp), %rax
movq (%rax), %rax
movq 0x3c8(%rsp), %rcx
vmovss (%rax,%rcx,4), %xmm0
vmovss %xmm0, 0x1c4(%rsp)
movq 0x1f8(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x170(%rsp), %rdx
movq %rdx, 0x228(%rsp)
movq %rcx, 0x220(%rsp)
movl %eax, 0x21c(%rsp)
movq 0x220(%rsp), %rax
movq %rax, 0x58(%rsp)
movb $0x0, 0x21b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x21c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x170(%rsp), %r10
movq %r10, 0x340(%rsp)
movl %r9d, 0x33c(%rsp)
movl %r8d, 0x338(%rsp)
movl %edi, 0x334(%rsp)
movq %rsi, 0x328(%rsp)
movq %rdx, 0x320(%rsp)
movl %ecx, 0x31c(%rsp)
movq %rax, 0x310(%rsp)
movq 0x340(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x328(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x320(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x31c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x310(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x33c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x338(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x334(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x350(%rsp)
movl $0x10, 0x34c(%rsp)
movq 0x350(%rsp), %rax
movslq 0x34c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x34c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x58(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x198(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x162fb4b
movq 0x58(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1b0(%rsp)
movb $0x1, 0x21b(%rsp)
testb $0x1, 0x21b(%rsp)
jne 0x162fc7a
leaq 0x170(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
movq %rax, 0x3a0(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, 0x48(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x162fc20
movq 0x48(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x39c(%rsp) # imm = 0xFFFFFFFF
movl 0x39c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x398(%rsp)
cmpl $0x1, 0x398(%rsp)
jne 0x162fc20
movq 0x48(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x162fbf1
movq 0x48(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x162fbef
jmp 0x162fc1e
movq 0x48(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3a8(%rsp)
cmpq $0x0, 0x3a8(%rsp)
je 0x162fc1c
movq 0x3a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x162fc1e
jmp 0x162fc20
movq 0x48(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x162fc78
movq %rax, %rdi
callq 0x678a0
jmp 0x162fc7a
leaq 0x170(%rsp), %rax
movq %rax, 0x230(%rsp)
movq 0x230(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x38(%rsp)
leaq 0x170(%rsp), %rax
movq %rax, 0x240(%rsp)
movq 0x240(%rsp), %rax
movq %rax, 0x390(%rsp)
movq 0x390(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x162fd59
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x38c(%rsp) # imm = 0xFFFFFFFF
movl 0x38c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x388(%rsp)
cmpl $0x1, 0x388(%rsp)
jne 0x162fd59
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x162fd2a
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x162fd28
jmp 0x162fd57
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b0(%rsp)
cmpq $0x0, 0x3b0(%rsp)
je 0x162fd55
movq 0x3b0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x162fd57
jmp 0x162fd59
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x162fdb1
movq %rax, %rdi
callq 0x678a0
movq 0x38(%rsp), %rax
movq %rax, 0x1b8(%rsp)
movq 0x1f0(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x120(%rsp), %rdx
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movl %eax, 0x254(%rsp)
movq 0x258(%rsp), %rax
movq %rax, 0x30(%rsp)
movb $0x0, 0x253(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x254(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x120(%rsp), %r10
movq %r10, 0x308(%rsp)
movl %r9d, 0x304(%rsp)
movl %r8d, 0x300(%rsp)
movl %edi, 0x2fc(%rsp)
movq %rsi, 0x2f0(%rsp)
movq %rdx, 0x2e8(%rsp)
movl %ecx, 0x2e4(%rsp)
movq %rax, 0x2d8(%rsp)
movq 0x308(%rsp), %rcx
movq %rcx, 0x28(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2e8(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2e4(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2d8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x304(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x300(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2fc(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x360(%rsp)
movl $0x10, 0x35c(%rsp)
movq 0x360(%rsp), %rax
movslq 0x35c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x35c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x30(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x148(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x162ff6e
movq 0x30(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x160(%rsp)
movb $0x1, 0x253(%rsp)
testb $0x1, 0x253(%rsp)
jne 0x163009d
leaq 0x120(%rsp), %rax
movq %rax, 0x268(%rsp)
movq 0x268(%rsp), %rax
movq %rax, 0x370(%rsp)
movq 0x370(%rsp), %rax
movq %rax, 0x20(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1630043
movq 0x20(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x36c(%rsp) # imm = 0xFFFFFFFF
movl 0x36c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x368(%rsp)
cmpl $0x1, 0x368(%rsp)
jne 0x1630043
movq 0x20(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1630014
movq 0x20(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1630012
jmp 0x1630041
movq 0x20(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3c0(%rsp)
cmpq $0x0, 0x3c0(%rsp)
je 0x163003f
movq 0x3c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1630041
jmp 0x1630043
movq 0x20(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163009b
movq %rax, %rdi
callq 0x678a0
jmp 0x163009d
leaq 0x120(%rsp), %rax
movq %rax, 0x270(%rsp)
movq 0x270(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x120(%rsp), %rax
movq %rax, 0x248(%rsp)
movq 0x248(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163017c
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x37c(%rsp) # imm = 0xFFFFFFFF
movl 0x37c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x378(%rsp)
cmpl $0x1, 0x378(%rsp)
jne 0x163017c
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163014d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x163014b
jmp 0x163017a
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b8(%rsp)
cmpq $0x0, 0x3b8(%rsp)
je 0x1630178
movq 0x3b8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x163017a
jmp 0x163017c
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16301d4
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x168(%rsp)
movl $0x0, 0x11c(%rsp)
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3dc(%rsp)
vmovss 0x3dc(%rsp), %xmm0
vmovss %xmm0, 0xc(%rsp)
vmovss %xmm0, 0x44c(%rsp)
vmovss %xmm0, 0x448(%rsp)
vmovss %xmm0, 0x444(%rsp)
vmovss %xmm0, 0x440(%rsp)
vmovss %xmm0, 0x43c(%rsp)
vmovss %xmm0, 0x438(%rsp)
vmovss %xmm0, 0x434(%rsp)
vmovss %xmm0, 0x430(%rsp)
vmovss 0x444(%rsp), %xmm1
vmovss 0x440(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x448(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x44c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm1 # xmm1 = xmm0[0,1,2],xmm1[0]
vmovss 0x434(%rsp), %xmm2
vmovss 0x430(%rsp), %xmm0
vinsertps $0x10, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
vmovss 0x438(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
vmovss 0x43c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm0, %xmm2 # xmm2 = xmm0[0,1,2],xmm2[0]
vmovaps %xmm2, %xmm0
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl 0x11c(%rsp), %eax
addl $0x7, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x16303bd
movq 0x1b8(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xc0(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0xe0(%rsp), %rsi
leaq 0xc0(%rsp), %rdx
callq 0x1636020
vmovaps %ymm0, 0xa0(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0xa0(%rsp), %ymm0
movq %rax, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
movq 0x2a0(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x11c(%rsp)
jmp 0x16302e6
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3fc(%rsp)
vbroadcastss 0x3fc(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
movl 0x11c(%rsp), %eax
addl $0x3, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x16304c8
movq 0x1b8(%rsp), %rax
movq %rax, 0x2a8(%rsp)
movq 0x2a8(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x80(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0x90(%rsp), %rsi
leaq 0x80(%rsp), %rdx
vzeroupper
callq 0x1636080
vmovaps %xmm0, 0x70(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0x70(%rsp), %xmm0
movq %rax, 0x2c8(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm0
movq 0x2c8(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x11c(%rsp)
jmp 0x16303f4
jmp 0x16304ca
movl 0x11c(%rsp), %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1630541
movq 0x1b8(%rsp), %rdx
leaq 0x1e7(%rsp), %rdi
leaq 0x1c4(%rsp), %rsi
vzeroupper
callq 0x16360c0
movq 0x168(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x11c(%rsp)
jmp 0x16304ca
jmp 0x1630543
movl 0x1c8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c8(%rsp)
jmp 0x162f94a
movl $0x0, 0x20c(%rsp)
movl 0x20c(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,101 |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_avx_functor::binary_op_min>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_7_13_19_29(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
// type 7 13 19 29
c.create_like(a, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float* ptr = a.channel(q);
const float* ptr1 = b.channel(q);
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _p1 = _mm512_loadu_ps(ptr1);
__m512 _outp = op.func_pack16(_p, _p1);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
ptr1 += 16;
outptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _p1 = _mm256_loadu_ps(ptr1);
__m256 _outp = op.func_pack8(_p, _p1);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
ptr1 += 8;
outptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _p1 = _mm_load_ps(ptr1);
__m128 _outp = op.func_pack4(_p, _p1);
_mm_store_ps(outptr, _outp);
ptr += 4;
ptr1 += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(*ptr, *ptr1);
ptr += 1;
ptr1 += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x500, %rsp # imm = 0x500
movq %rdi, 0x268(%rsp)
movq %rsi, 0x260(%rsp)
movq %rdx, 0x258(%rsp)
movq %rcx, 0x250(%rsp)
movq 0x268(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x248(%rsp)
movq 0x268(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x244(%rsp)
movq 0x268(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x240(%rsp)
movq 0x268(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x23c(%rsp)
movq 0x268(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x238(%rsp)
movl 0x248(%rsp), %eax
imull 0x244(%rsp), %eax
imull 0x240(%rsp), %eax
imull 0x238(%rsp), %eax
movl %eax, 0x234(%rsp)
movq 0x258(%rsp), %rdi
movq 0x268(%rsp), %rsi
movq 0x250(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x258(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rcx
movq %rcx, 0x80(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x8f(%rsp)
je 0x16306b1
movq 0x80(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x8f(%rsp)
movb 0x8f(%rsp), %al
testb $0x1, %al
jne 0x16306be
jmp 0x16306ce
movl $0xffffff9c, 0x274(%rsp) # imm = 0xFFFFFF9C
jmp 0x1631639
movl $0x0, 0x230(%rsp)
movl 0x230(%rsp), %eax
cmpl 0x23c(%rsp), %eax
jge 0x163162e
movq 0x268(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x1e0(%rsp), %rdx
movq %rdx, 0x2a8(%rsp)
movq %rcx, 0x2a0(%rsp)
movl %eax, 0x29c(%rsp)
movq 0x2a0(%rsp), %rax
movq %rax, 0x78(%rsp)
movb $0x0, 0x29b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x29c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x1e0(%rsp), %r10
movq %r10, 0x3f0(%rsp)
movl %r9d, 0x3ec(%rsp)
movl %r8d, 0x3e8(%rsp)
movl %edi, 0x3e4(%rsp)
movq %rsi, 0x3d8(%rsp)
movq %rdx, 0x3d0(%rsp)
movl %ecx, 0x3cc(%rsp)
movq %rax, 0x3c0(%rsp)
movq 0x3f0(%rsp), %rcx
movq %rcx, 0x70(%rsp)
movq 0x3d8(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x3d0(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x3cc(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x3c0(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x3ec(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x3e8(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x3e4(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x448(%rsp)
movl $0x10, 0x444(%rsp)
movq 0x448(%rsp), %rax
movslq 0x444(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x444(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x78(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x208(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163089d
movq 0x78(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x220(%rsp)
movb $0x1, 0x29b(%rsp)
testb $0x1, 0x29b(%rsp)
jne 0x16309cc
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2c0(%rsp)
movq 0x2c0(%rsp), %rax
movq %rax, 0x4b8(%rsp)
movq 0x4b8(%rsp), %rax
movq %rax, 0x68(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1630972
movq 0x68(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x4b4(%rsp) # imm = 0xFFFFFFFF
movl 0x4b4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x4b0(%rsp)
cmpl $0x1, 0x4b0(%rsp)
jne 0x1630972
movq 0x68(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1630943
movq 0x68(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1630941
jmp 0x1630970
movq 0x68(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4c0(%rsp)
cmpq $0x0, 0x4c0(%rsp)
je 0x163096e
movq 0x4c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1630970
jmp 0x1630972
movq 0x68(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16309ca
movq %rax, %rdi
callq 0x678a0
jmp 0x16309cc
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2b8(%rsp)
movq 0x2b8(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x58(%rsp)
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rax
movq %rax, 0x498(%rsp)
movq 0x498(%rsp), %rax
movq %rax, 0x60(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1630aab
movq 0x60(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x494(%rsp) # imm = 0xFFFFFFFF
movl 0x494(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x490(%rsp)
cmpl $0x1, 0x490(%rsp)
jne 0x1630aab
movq 0x60(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1630a7c
movq 0x60(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1630a7a
jmp 0x1630aa9
movq 0x60(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4d0(%rsp)
cmpq $0x0, 0x4d0(%rsp)
je 0x1630aa7
movq 0x4d0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1630aa9
jmp 0x1630aab
movq 0x60(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1630b03
movq %rax, %rdi
callq 0x678a0
movq 0x58(%rsp), %rax
movq %rax, 0x228(%rsp)
movq 0x260(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x190(%rsp), %rdx
movq %rdx, 0x290(%rsp)
movq %rcx, 0x288(%rsp)
movl %eax, 0x284(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x50(%rsp)
movb $0x0, 0x283(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x284(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x190(%rsp), %r10
movq %r10, 0x428(%rsp)
movl %r9d, 0x424(%rsp)
movl %r8d, 0x420(%rsp)
movl %edi, 0x41c(%rsp)
movq %rsi, 0x410(%rsp)
movq %rdx, 0x408(%rsp)
movl %ecx, 0x404(%rsp)
movq %rax, 0x3f8(%rsp)
movq 0x428(%rsp), %rcx
movq %rcx, 0x48(%rsp)
movq 0x410(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x408(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x404(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x3f8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x424(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x420(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x41c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x438(%rsp)
movl $0x10, 0x434(%rsp)
movq 0x438(%rsp), %rax
movslq 0x434(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x434(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x50(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x1b8(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1630cc0
movq 0x50(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1d0(%rsp)
movb $0x1, 0x283(%rsp)
testb $0x1, 0x283(%rsp)
jne 0x1630def
leaq 0x190(%rsp), %rax
movq %rax, 0x2c8(%rsp)
movq 0x2c8(%rsp), %rax
movq %rax, 0x4a8(%rsp)
movq 0x4a8(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1630d95
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x4a4(%rsp) # imm = 0xFFFFFFFF
movl 0x4a4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x4a0(%rsp)
cmpl $0x1, 0x4a0(%rsp)
jne 0x1630d95
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1630d66
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1630d64
jmp 0x1630d93
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4c8(%rsp)
cmpq $0x0, 0x4c8(%rsp)
je 0x1630d91
movq 0x4c8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1630d93
jmp 0x1630d95
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1630ded
movq %rax, %rdi
callq 0x678a0
jmp 0x1630def
leaq 0x190(%rsp), %rax
movq %rax, 0x2b0(%rsp)
movq 0x2b0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x30(%rsp)
leaq 0x190(%rsp), %rax
movq %rax, 0x2d8(%rsp)
movq 0x2d8(%rsp), %rax
movq %rax, 0x488(%rsp)
movq 0x488(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1630ece
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x484(%rsp) # imm = 0xFFFFFFFF
movl 0x484(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x480(%rsp)
cmpl $0x1, 0x480(%rsp)
jne 0x1630ece
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1630e9f
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1630e9d
jmp 0x1630ecc
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4d8(%rsp)
cmpq $0x0, 0x4d8(%rsp)
je 0x1630eca
movq 0x4d8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1630ecc
jmp 0x1630ece
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1630f26
movq %rax, %rdi
callq 0x678a0
movq 0x30(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x258(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x140(%rsp), %rdx
movq %rdx, 0x2f8(%rsp)
movq %rcx, 0x2f0(%rsp)
movl %eax, 0x2ec(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0x2eb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x2ec(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x140(%rsp), %r10
movq %r10, 0x3b8(%rsp)
movl %r9d, 0x3b4(%rsp)
movl %r8d, 0x3b0(%rsp)
movl %edi, 0x3ac(%rsp)
movq %rsi, 0x3a0(%rsp)
movq %rdx, 0x398(%rsp)
movl %ecx, 0x394(%rsp)
movq %rax, 0x388(%rsp)
movq 0x3b8(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x398(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x394(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x388(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x3b4(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x3b0(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x3ac(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x458(%rsp)
movl $0x10, 0x454(%rsp)
movq 0x458(%rsp), %rax
movslq 0x454(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x454(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x168(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x16310e3
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x180(%rsp)
movb $0x1, 0x2eb(%rsp)
testb $0x1, 0x2eb(%rsp)
jne 0x1631212
leaq 0x140(%rsp), %rax
movq %rax, 0x300(%rsp)
movq 0x300(%rsp), %rax
movq %rax, 0x468(%rsp)
movq 0x468(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16311b8
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x464(%rsp) # imm = 0xFFFFFFFF
movl 0x464(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x460(%rsp)
cmpl $0x1, 0x460(%rsp)
jne 0x16311b8
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1631189
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1631187
jmp 0x16311b6
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4e8(%rsp)
cmpq $0x0, 0x4e8(%rsp)
je 0x16311b4
movq 0x4e8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16311b6
jmp 0x16311b8
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1631210
movq %rax, %rdi
callq 0x678a0
jmp 0x1631212
leaq 0x140(%rsp), %rax
movq %rax, 0x308(%rsp)
movq 0x308(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x140(%rsp), %rax
movq %rax, 0x2e0(%rsp)
movq 0x2e0(%rsp), %rax
movq %rax, 0x478(%rsp)
movq 0x478(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16312f1
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x474(%rsp) # imm = 0xFFFFFFFF
movl 0x474(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x470(%rsp)
cmpl $0x1, 0x470(%rsp)
jne 0x16312f1
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16312c2
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16312c0
jmp 0x16312ef
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4e0(%rsp)
cmpq $0x0, 0x4e0(%rsp)
je 0x16312ed
movq 0x4e0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16312ef
jmp 0x16312f1
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1631349
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x188(%rsp)
movl $0x0, 0x13c(%rsp)
movl 0x13c(%rsp), %eax
addl $0x7, %eax
cmpl 0x234(%rsp), %eax
jge 0x1631471
movq 0x228(%rsp), %rax
movq %rax, 0x318(%rsp)
movq 0x318(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x100(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x310(%rsp)
movq 0x310(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
leaq 0x24f(%rsp), %rdi
leaq 0x100(%rsp), %rsi
leaq 0xe0(%rsp), %rdx
callq 0x16360f0
vmovaps %ymm0, 0xc0(%rsp)
movq 0x188(%rsp), %rax
vmovaps 0xc0(%rsp), %ymm0
movq %rax, 0x348(%rsp)
vmovaps %ymm0, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm0
movq 0x348(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1631361
jmp 0x1631473
movl 0x13c(%rsp), %eax
addl $0x3, %eax
cmpl 0x234(%rsp), %eax
jge 0x1631586
movq 0x228(%rsp), %rax
movq %rax, 0x358(%rsp)
movq 0x358(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0xb0(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x350(%rsp)
movq 0x350(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0xa0(%rsp)
leaq 0x24f(%rsp), %rdi
leaq 0xb0(%rsp), %rsi
leaq 0xa0(%rsp), %rdx
vzeroupper
callq 0x1636150
vmovaps %xmm0, 0x90(%rsp)
movq 0x188(%rsp), %rax
vmovaps 0x90(%rsp), %xmm0
movq %rax, 0x378(%rsp)
vmovaps %xmm0, 0x360(%rsp)
vmovaps 0x360(%rsp), %xmm0
movq 0x378(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1631473
jmp 0x1631588
movl 0x13c(%rsp), %eax
cmpl 0x234(%rsp), %eax
jge 0x1631616
movq 0x228(%rsp), %rsi
movq 0x1d8(%rsp), %rdx
leaq 0x24f(%rsp), %rdi
vzeroupper
callq 0x1636190
movq 0x188(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1631588
jmp 0x1631618
movl 0x230(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x230(%rsp)
jmp 0x16306d9
movl $0x0, 0x274(%rsp)
movl 0x274(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,102 |
int ncnn::binary_op_6_11_16_25<ncnn::BinaryOp_x86_avx_functor::binary_op_min>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_6_11_16_25(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
// type 6 11 16 25
c.create_like(a, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float* ptr = a.channel(q);
const float b0 = b[0];
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
__m512 _b0_avx512 = _mm512_set1_ps(b0);
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _outp = op.func_pack16(_p, _b0_avx512);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
outptr += 16;
}
#endif // __AVX512F__
__m256 _b0_avx = _mm256_set1_ps(b0);
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _outp = op.func_pack8(_p, _b0_avx);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
outptr += 8;
}
#endif // __AVX__
__m128 _b0 = _mm_set1_ps(b0);
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _outp = op.func_pack4(_p, _b0);
_mm_store_ps(outptr, _outp);
ptr += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(*ptr, b0);
ptr += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x460, %rsp # imm = 0x460
movq %rdi, 0x200(%rsp)
movq %rsi, 0x1f8(%rsp)
movq %rdx, 0x1f0(%rsp)
movq %rcx, 0x1e8(%rsp)
movq 0x200(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1e0(%rsp)
movq 0x200(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1dc(%rsp)
movq 0x200(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x1d8(%rsp)
movq 0x200(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x1d4(%rsp)
movq 0x200(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x1d0(%rsp)
movl 0x1e0(%rsp), %eax
imull 0x1dc(%rsp), %eax
imull 0x1d8(%rsp), %eax
imull 0x1d0(%rsp), %eax
movl %eax, 0x1cc(%rsp)
movq 0x1f0(%rsp), %rdi
movq 0x200(%rsp), %rsi
movq 0x1e8(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x1f0(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rcx
movq %rcx, 0x60(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x6f(%rsp)
je 0x1631775
movq 0x60(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x6f(%rsp)
movb 0x6f(%rsp), %al
testb $0x1, %al
jne 0x163177f
jmp 0x163178f
movl $0xffffff9c, 0x20c(%rsp) # imm = 0xFFFFFF9C
jmp 0x16323b4
movl $0x0, 0x1c8(%rsp)
movl 0x1c8(%rsp), %eax
cmpl 0x1d4(%rsp), %eax
jge 0x16323a9
movq 0x200(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x178(%rsp), %rdx
movq %rdx, 0x228(%rsp)
movq %rcx, 0x220(%rsp)
movl %eax, 0x21c(%rsp)
movq 0x220(%rsp), %rax
movq %rax, 0x58(%rsp)
movb $0x0, 0x21b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x21c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x178(%rsp), %r10
movq %r10, 0x340(%rsp)
movl %r9d, 0x33c(%rsp)
movl %r8d, 0x338(%rsp)
movl %edi, 0x334(%rsp)
movq %rsi, 0x328(%rsp)
movq %rdx, 0x320(%rsp)
movl %ecx, 0x31c(%rsp)
movq %rax, 0x310(%rsp)
movq 0x340(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x328(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x320(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x31c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x310(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x33c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x338(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x334(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x350(%rsp)
movl $0x10, 0x34c(%rsp)
movq 0x350(%rsp), %rax
movslq 0x34c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x34c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x58(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x1a0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163195e
movq 0x58(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1b8(%rsp)
movb $0x1, 0x21b(%rsp)
testb $0x1, 0x21b(%rsp)
jne 0x1631a8d
leaq 0x178(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
movq %rax, 0x3a0(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, 0x48(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1631a33
movq 0x48(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x39c(%rsp) # imm = 0xFFFFFFFF
movl 0x39c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x398(%rsp)
cmpl $0x1, 0x398(%rsp)
jne 0x1631a33
movq 0x48(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1631a04
movq 0x48(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1631a02
jmp 0x1631a31
movq 0x48(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3a8(%rsp)
cmpq $0x0, 0x3a8(%rsp)
je 0x1631a2f
movq 0x3a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1631a31
jmp 0x1631a33
movq 0x48(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1631a8b
movq %rax, %rdi
callq 0x678a0
jmp 0x1631a8d
leaq 0x178(%rsp), %rax
movq %rax, 0x230(%rsp)
movq 0x230(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x38(%rsp)
leaq 0x178(%rsp), %rax
movq %rax, 0x240(%rsp)
movq 0x240(%rsp), %rax
movq %rax, 0x390(%rsp)
movq 0x390(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1631b6c
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x38c(%rsp) # imm = 0xFFFFFFFF
movl 0x38c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x388(%rsp)
cmpl $0x1, 0x388(%rsp)
jne 0x1631b6c
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1631b3d
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1631b3b
jmp 0x1631b6a
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b0(%rsp)
cmpq $0x0, 0x3b0(%rsp)
je 0x1631b68
movq 0x3b0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1631b6a
jmp 0x1631b6c
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1631bc4
movq %rax, %rdi
callq 0x678a0
movq 0x38(%rsp), %rax
movq %rax, 0x1c0(%rsp)
movq 0x1f8(%rsp), %rax
movq %rax, 0x3d0(%rsp)
movq $0x0, 0x3c8(%rsp)
movq 0x3d0(%rsp), %rax
movq (%rax), %rax
movq 0x3c8(%rsp), %rcx
vmovss (%rax,%rcx,4), %xmm0
vmovss %xmm0, 0x174(%rsp)
movq 0x1f0(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x120(%rsp), %rdx
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movl %eax, 0x254(%rsp)
movq 0x258(%rsp), %rax
movq %rax, 0x30(%rsp)
movb $0x0, 0x253(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x254(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x120(%rsp), %r10
movq %r10, 0x308(%rsp)
movl %r9d, 0x304(%rsp)
movl %r8d, 0x300(%rsp)
movl %edi, 0x2fc(%rsp)
movq %rsi, 0x2f0(%rsp)
movq %rdx, 0x2e8(%rsp)
movl %ecx, 0x2e4(%rsp)
movq %rax, 0x2d8(%rsp)
movq 0x308(%rsp), %rcx
movq %rcx, 0x28(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2e8(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2e4(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2d8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x304(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x300(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2fc(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x360(%rsp)
movl $0x10, 0x35c(%rsp)
movq 0x360(%rsp), %rax
movslq 0x35c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x35c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x30(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x148(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1631dbe
movq 0x30(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x160(%rsp)
movb $0x1, 0x253(%rsp)
testb $0x1, 0x253(%rsp)
jne 0x1631eed
leaq 0x120(%rsp), %rax
movq %rax, 0x268(%rsp)
movq 0x268(%rsp), %rax
movq %rax, 0x370(%rsp)
movq 0x370(%rsp), %rax
movq %rax, 0x20(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1631e93
movq 0x20(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x36c(%rsp) # imm = 0xFFFFFFFF
movl 0x36c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x368(%rsp)
cmpl $0x1, 0x368(%rsp)
jne 0x1631e93
movq 0x20(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1631e64
movq 0x20(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1631e62
jmp 0x1631e91
movq 0x20(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3c0(%rsp)
cmpq $0x0, 0x3c0(%rsp)
je 0x1631e8f
movq 0x3c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1631e91
jmp 0x1631e93
movq 0x20(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1631eeb
movq %rax, %rdi
callq 0x678a0
jmp 0x1631eed
leaq 0x120(%rsp), %rax
movq %rax, 0x270(%rsp)
movq 0x270(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x120(%rsp), %rax
movq %rax, 0x248(%rsp)
movq 0x248(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1631fcc
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x37c(%rsp) # imm = 0xFFFFFFFF
movl 0x37c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x378(%rsp)
cmpl $0x1, 0x378(%rsp)
jne 0x1631fcc
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1631f9d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1631f9b
jmp 0x1631fca
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b8(%rsp)
cmpq $0x0, 0x3b8(%rsp)
je 0x1631fc8
movq 0x3b8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1631fca
jmp 0x1631fcc
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1632024
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x168(%rsp)
movl $0x0, 0x11c(%rsp)
vmovss 0x174(%rsp), %xmm0
vmovss %xmm0, 0x3dc(%rsp)
vmovss 0x3dc(%rsp), %xmm0
vmovss %xmm0, 0xc(%rsp)
vmovss %xmm0, 0x44c(%rsp)
vmovss %xmm0, 0x448(%rsp)
vmovss %xmm0, 0x444(%rsp)
vmovss %xmm0, 0x440(%rsp)
vmovss %xmm0, 0x43c(%rsp)
vmovss %xmm0, 0x438(%rsp)
vmovss %xmm0, 0x434(%rsp)
vmovss %xmm0, 0x430(%rsp)
vmovss 0x444(%rsp), %xmm1
vmovss 0x440(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x448(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x44c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm1 # xmm1 = xmm0[0,1,2],xmm1[0]
vmovss 0x434(%rsp), %xmm2
vmovss 0x430(%rsp), %xmm0
vinsertps $0x10, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
vmovss 0x438(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
vmovss 0x43c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm0, %xmm2 # xmm2 = xmm0[0,1,2],xmm2[0]
vmovaps %xmm2, %xmm0
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl 0x11c(%rsp), %eax
addl $0x7, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x163220d
movq 0x1c0(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xc0(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0xc0(%rsp), %rsi
leaq 0xe0(%rsp), %rdx
callq 0x16360f0
vmovaps %ymm0, 0xa0(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0xa0(%rsp), %ymm0
movq %rax, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
movq 0x2a0(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1632136
vmovss 0x174(%rsp), %xmm0
vmovss %xmm0, 0x3fc(%rsp)
vbroadcastss 0x3fc(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
movl 0x11c(%rsp), %eax
addl $0x3, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1632318
movq 0x1c0(%rsp), %rax
movq %rax, 0x2a8(%rsp)
movq 0x2a8(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x80(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0x80(%rsp), %rsi
leaq 0x90(%rsp), %rdx
vzeroupper
callq 0x1636150
vmovaps %xmm0, 0x70(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0x70(%rsp), %xmm0
movq %rax, 0x2c8(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm0
movq 0x2c8(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1632244
jmp 0x163231a
movl 0x11c(%rsp), %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1632391
movq 0x1c0(%rsp), %rsi
leaq 0x1e7(%rsp), %rdi
leaq 0x174(%rsp), %rdx
vzeroupper
callq 0x1636190
movq 0x168(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x11c(%rsp)
jmp 0x163231a
jmp 0x1632393
movl 0x1c8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c8(%rsp)
jmp 0x163179a
movl $0x0, 0x20c(%rsp)
movl 0x20c(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,103 |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_avx_functor::binary_op_min>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_2_3_4_20(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = b.w;
int h = b.h;
int d = b.d;
int channels = b.c;
int elempack = b.elempack;
int size = w * h * d * elempack;
// type 2 3 4 20
c.create_like(b, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float a0 = a[0];
const float* ptr = b.channel(q);
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
__m512 _a0_avx512 = _mm512_set1_ps(a0);
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _outp = op.func_pack16(_a0_avx512, _p);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
outptr += 16;
}
#endif // __AVX512F__
__m256 _a0_avx = _mm256_set1_ps(a0);
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _outp = op.func_pack8(_a0_avx, _p);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
outptr += 8;
}
#endif // __AVX__
__m128 _a0 = _mm_set1_ps(a0);
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _outp = op.func_pack4(_a0, _p);
_mm_store_ps(outptr, _outp);
ptr += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(a0, *ptr);
ptr += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x460, %rsp # imm = 0x460
movq %rdi, 0x200(%rsp)
movq %rsi, 0x1f8(%rsp)
movq %rdx, 0x1f0(%rsp)
movq %rcx, 0x1e8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1e0(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1dc(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x1d8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x1d4(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x1d0(%rsp)
movl 0x1e0(%rsp), %eax
imull 0x1dc(%rsp), %eax
imull 0x1d8(%rsp), %eax
imull 0x1d0(%rsp), %eax
movl %eax, 0x1cc(%rsp)
movq 0x1f0(%rsp), %rdi
movq 0x1f8(%rsp), %rsi
movq 0x1e8(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x1f0(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rcx
movq %rcx, 0x60(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x6f(%rsp)
je 0x16324f5
movq 0x60(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x6f(%rsp)
movb 0x6f(%rsp), %al
testb $0x1, %al
jne 0x16324ff
jmp 0x163250f
movl $0xffffff9c, 0x20c(%rsp) # imm = 0xFFFFFF9C
jmp 0x1633134
movl $0x0, 0x1c8(%rsp)
movl 0x1c8(%rsp), %eax
cmpl 0x1d4(%rsp), %eax
jge 0x1633129
movq 0x200(%rsp), %rax
movq %rax, 0x3d0(%rsp)
movq $0x0, 0x3c8(%rsp)
movq 0x3d0(%rsp), %rax
movq (%rax), %rax
movq 0x3c8(%rsp), %rcx
vmovss (%rax,%rcx,4), %xmm0
vmovss %xmm0, 0x1c4(%rsp)
movq 0x1f8(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x170(%rsp), %rdx
movq %rdx, 0x228(%rsp)
movq %rcx, 0x220(%rsp)
movl %eax, 0x21c(%rsp)
movq 0x220(%rsp), %rax
movq %rax, 0x58(%rsp)
movb $0x0, 0x21b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x21c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x170(%rsp), %r10
movq %r10, 0x340(%rsp)
movl %r9d, 0x33c(%rsp)
movl %r8d, 0x338(%rsp)
movl %edi, 0x334(%rsp)
movq %rsi, 0x328(%rsp)
movq %rdx, 0x320(%rsp)
movl %ecx, 0x31c(%rsp)
movq %rax, 0x310(%rsp)
movq 0x340(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x328(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x320(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x31c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x310(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x33c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x338(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x334(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x350(%rsp)
movl $0x10, 0x34c(%rsp)
movq 0x350(%rsp), %rax
movslq 0x34c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x34c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x58(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x198(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163271b
movq 0x58(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1b0(%rsp)
movb $0x1, 0x21b(%rsp)
testb $0x1, 0x21b(%rsp)
jne 0x163284a
leaq 0x170(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
movq %rax, 0x3a0(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, 0x48(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16327f0
movq 0x48(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x39c(%rsp) # imm = 0xFFFFFFFF
movl 0x39c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x398(%rsp)
cmpl $0x1, 0x398(%rsp)
jne 0x16327f0
movq 0x48(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16327c1
movq 0x48(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16327bf
jmp 0x16327ee
movq 0x48(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3a8(%rsp)
cmpq $0x0, 0x3a8(%rsp)
je 0x16327ec
movq 0x3a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16327ee
jmp 0x16327f0
movq 0x48(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1632848
movq %rax, %rdi
callq 0x678a0
jmp 0x163284a
leaq 0x170(%rsp), %rax
movq %rax, 0x230(%rsp)
movq 0x230(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x38(%rsp)
leaq 0x170(%rsp), %rax
movq %rax, 0x240(%rsp)
movq 0x240(%rsp), %rax
movq %rax, 0x390(%rsp)
movq 0x390(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1632929
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x38c(%rsp) # imm = 0xFFFFFFFF
movl 0x38c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x388(%rsp)
cmpl $0x1, 0x388(%rsp)
jne 0x1632929
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16328fa
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16328f8
jmp 0x1632927
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b0(%rsp)
cmpq $0x0, 0x3b0(%rsp)
je 0x1632925
movq 0x3b0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1632927
jmp 0x1632929
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1632981
movq %rax, %rdi
callq 0x678a0
movq 0x38(%rsp), %rax
movq %rax, 0x1b8(%rsp)
movq 0x1f0(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x120(%rsp), %rdx
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movl %eax, 0x254(%rsp)
movq 0x258(%rsp), %rax
movq %rax, 0x30(%rsp)
movb $0x0, 0x253(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x254(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x120(%rsp), %r10
movq %r10, 0x308(%rsp)
movl %r9d, 0x304(%rsp)
movl %r8d, 0x300(%rsp)
movl %edi, 0x2fc(%rsp)
movq %rsi, 0x2f0(%rsp)
movq %rdx, 0x2e8(%rsp)
movl %ecx, 0x2e4(%rsp)
movq %rax, 0x2d8(%rsp)
movq 0x308(%rsp), %rcx
movq %rcx, 0x28(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2e8(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2e4(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2d8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x304(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x300(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2fc(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x360(%rsp)
movl $0x10, 0x35c(%rsp)
movq 0x360(%rsp), %rax
movslq 0x35c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x35c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x30(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x148(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1632b3e
movq 0x30(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x160(%rsp)
movb $0x1, 0x253(%rsp)
testb $0x1, 0x253(%rsp)
jne 0x1632c6d
leaq 0x120(%rsp), %rax
movq %rax, 0x268(%rsp)
movq 0x268(%rsp), %rax
movq %rax, 0x370(%rsp)
movq 0x370(%rsp), %rax
movq %rax, 0x20(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1632c13
movq 0x20(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x36c(%rsp) # imm = 0xFFFFFFFF
movl 0x36c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x368(%rsp)
cmpl $0x1, 0x368(%rsp)
jne 0x1632c13
movq 0x20(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1632be4
movq 0x20(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1632be2
jmp 0x1632c11
movq 0x20(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3c0(%rsp)
cmpq $0x0, 0x3c0(%rsp)
je 0x1632c0f
movq 0x3c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1632c11
jmp 0x1632c13
movq 0x20(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1632c6b
movq %rax, %rdi
callq 0x678a0
jmp 0x1632c6d
leaq 0x120(%rsp), %rax
movq %rax, 0x270(%rsp)
movq 0x270(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x120(%rsp), %rax
movq %rax, 0x248(%rsp)
movq 0x248(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1632d4c
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x37c(%rsp) # imm = 0xFFFFFFFF
movl 0x37c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x378(%rsp)
cmpl $0x1, 0x378(%rsp)
jne 0x1632d4c
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1632d1d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1632d1b
jmp 0x1632d4a
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b8(%rsp)
cmpq $0x0, 0x3b8(%rsp)
je 0x1632d48
movq 0x3b8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1632d4a
jmp 0x1632d4c
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1632da4
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x168(%rsp)
movl $0x0, 0x11c(%rsp)
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3dc(%rsp)
vmovss 0x3dc(%rsp), %xmm0
vmovss %xmm0, 0xc(%rsp)
vmovss %xmm0, 0x44c(%rsp)
vmovss %xmm0, 0x448(%rsp)
vmovss %xmm0, 0x444(%rsp)
vmovss %xmm0, 0x440(%rsp)
vmovss %xmm0, 0x43c(%rsp)
vmovss %xmm0, 0x438(%rsp)
vmovss %xmm0, 0x434(%rsp)
vmovss %xmm0, 0x430(%rsp)
vmovss 0x444(%rsp), %xmm1
vmovss 0x440(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x448(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x44c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm1 # xmm1 = xmm0[0,1,2],xmm1[0]
vmovss 0x434(%rsp), %xmm2
vmovss 0x430(%rsp), %xmm0
vinsertps $0x10, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
vmovss 0x438(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
vmovss 0x43c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm0, %xmm2 # xmm2 = xmm0[0,1,2],xmm2[0]
vmovaps %xmm2, %xmm0
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl 0x11c(%rsp), %eax
addl $0x7, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1632f8d
movq 0x1b8(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xc0(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0xe0(%rsp), %rsi
leaq 0xc0(%rsp), %rdx
callq 0x16360f0
vmovaps %ymm0, 0xa0(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0xa0(%rsp), %ymm0
movq %rax, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
movq 0x2a0(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1632eb6
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3fc(%rsp)
vbroadcastss 0x3fc(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
movl 0x11c(%rsp), %eax
addl $0x3, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1633098
movq 0x1b8(%rsp), %rax
movq %rax, 0x2a8(%rsp)
movq 0x2a8(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x80(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0x90(%rsp), %rsi
leaq 0x80(%rsp), %rdx
vzeroupper
callq 0x1636150
vmovaps %xmm0, 0x70(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0x70(%rsp), %xmm0
movq %rax, 0x2c8(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm0
movq 0x2c8(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1632fc4
jmp 0x163309a
movl 0x11c(%rsp), %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1633111
movq 0x1b8(%rsp), %rdx
leaq 0x1e7(%rsp), %rdi
leaq 0x1c4(%rsp), %rsi
vzeroupper
callq 0x1636190
movq 0x168(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x11c(%rsp)
jmp 0x163309a
jmp 0x1633113
movl 0x1c8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c8(%rsp)
jmp 0x163251a
movl $0x0, 0x20c(%rsp)
movl 0x20c(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,104 |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_avx_functor::binary_op_pow>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_7_13_19_29(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
// type 7 13 19 29
c.create_like(a, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float* ptr = a.channel(q);
const float* ptr1 = b.channel(q);
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _p1 = _mm512_loadu_ps(ptr1);
__m512 _outp = op.func_pack16(_p, _p1);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
ptr1 += 16;
outptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _p1 = _mm256_loadu_ps(ptr1);
__m256 _outp = op.func_pack8(_p, _p1);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
ptr1 += 8;
outptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _p1 = _mm_load_ps(ptr1);
__m128 _outp = op.func_pack4(_p, _p1);
_mm_store_ps(outptr, _outp);
ptr += 4;
ptr1 += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(*ptr, *ptr1);
ptr += 1;
ptr1 += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x500, %rsp # imm = 0x500
movq %rdi, 0x268(%rsp)
movq %rsi, 0x260(%rsp)
movq %rdx, 0x258(%rsp)
movq %rcx, 0x250(%rsp)
movq 0x268(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x248(%rsp)
movq 0x268(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x244(%rsp)
movq 0x268(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x240(%rsp)
movq 0x268(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x23c(%rsp)
movq 0x268(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x238(%rsp)
movl 0x248(%rsp), %eax
imull 0x244(%rsp), %eax
imull 0x240(%rsp), %eax
imull 0x238(%rsp), %eax
movl %eax, 0x234(%rsp)
movq 0x258(%rsp), %rdi
movq 0x268(%rsp), %rsi
movq 0x250(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x258(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rcx
movq %rcx, 0x80(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x8f(%rsp)
je 0x1633281
movq 0x80(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x8f(%rsp)
movb 0x8f(%rsp), %al
testb $0x1, %al
jne 0x163328e
jmp 0x163329e
movl $0xffffff9c, 0x274(%rsp) # imm = 0xFFFFFF9C
jmp 0x1634209
movl $0x0, 0x230(%rsp)
movl 0x230(%rsp), %eax
cmpl 0x23c(%rsp), %eax
jge 0x16341fe
movq 0x268(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x1e0(%rsp), %rdx
movq %rdx, 0x2a8(%rsp)
movq %rcx, 0x2a0(%rsp)
movl %eax, 0x29c(%rsp)
movq 0x2a0(%rsp), %rax
movq %rax, 0x78(%rsp)
movb $0x0, 0x29b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x29c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x1e0(%rsp), %r10
movq %r10, 0x3f0(%rsp)
movl %r9d, 0x3ec(%rsp)
movl %r8d, 0x3e8(%rsp)
movl %edi, 0x3e4(%rsp)
movq %rsi, 0x3d8(%rsp)
movq %rdx, 0x3d0(%rsp)
movl %ecx, 0x3cc(%rsp)
movq %rax, 0x3c0(%rsp)
movq 0x3f0(%rsp), %rcx
movq %rcx, 0x70(%rsp)
movq 0x3d8(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x3d0(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x3cc(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x3c0(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x3ec(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x3e8(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x3e4(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x448(%rsp)
movl $0x10, 0x444(%rsp)
movq 0x448(%rsp), %rax
movslq 0x444(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x444(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x78(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x208(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163346d
movq 0x78(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x220(%rsp)
movb $0x1, 0x29b(%rsp)
testb $0x1, 0x29b(%rsp)
jne 0x163359c
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2c0(%rsp)
movq 0x2c0(%rsp), %rax
movq %rax, 0x4b8(%rsp)
movq 0x4b8(%rsp), %rax
movq %rax, 0x68(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1633542
movq 0x68(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x4b4(%rsp) # imm = 0xFFFFFFFF
movl 0x4b4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x4b0(%rsp)
cmpl $0x1, 0x4b0(%rsp)
jne 0x1633542
movq 0x68(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1633513
movq 0x68(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1633511
jmp 0x1633540
movq 0x68(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4c0(%rsp)
cmpq $0x0, 0x4c0(%rsp)
je 0x163353e
movq 0x4c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633540
jmp 0x1633542
movq 0x68(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163359a
movq %rax, %rdi
callq 0x678a0
jmp 0x163359c
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2b8(%rsp)
movq 0x2b8(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x58(%rsp)
leaq 0x1e0(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rax
movq %rax, 0x498(%rsp)
movq 0x498(%rsp), %rax
movq %rax, 0x60(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163367b
movq 0x60(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x494(%rsp) # imm = 0xFFFFFFFF
movl 0x494(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x490(%rsp)
cmpl $0x1, 0x490(%rsp)
jne 0x163367b
movq 0x60(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163364c
movq 0x60(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x163364a
jmp 0x1633679
movq 0x60(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4d0(%rsp)
cmpq $0x0, 0x4d0(%rsp)
je 0x1633677
movq 0x4d0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633679
jmp 0x163367b
movq 0x60(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16336d3
movq %rax, %rdi
callq 0x678a0
movq 0x58(%rsp), %rax
movq %rax, 0x228(%rsp)
movq 0x260(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x190(%rsp), %rdx
movq %rdx, 0x290(%rsp)
movq %rcx, 0x288(%rsp)
movl %eax, 0x284(%rsp)
movq 0x288(%rsp), %rax
movq %rax, 0x50(%rsp)
movb $0x0, 0x283(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x284(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x190(%rsp), %r10
movq %r10, 0x428(%rsp)
movl %r9d, 0x424(%rsp)
movl %r8d, 0x420(%rsp)
movl %edi, 0x41c(%rsp)
movq %rsi, 0x410(%rsp)
movq %rdx, 0x408(%rsp)
movl %ecx, 0x404(%rsp)
movq %rax, 0x3f8(%rsp)
movq 0x428(%rsp), %rcx
movq %rcx, 0x48(%rsp)
movq 0x410(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x408(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x404(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x3f8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x424(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x420(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x41c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x438(%rsp)
movl $0x10, 0x434(%rsp)
movq 0x438(%rsp), %rax
movslq 0x434(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x434(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x50(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x1b8(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1633890
movq 0x50(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1d0(%rsp)
movb $0x1, 0x283(%rsp)
testb $0x1, 0x283(%rsp)
jne 0x16339bf
leaq 0x190(%rsp), %rax
movq %rax, 0x2c8(%rsp)
movq 0x2c8(%rsp), %rax
movq %rax, 0x4a8(%rsp)
movq 0x4a8(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1633965
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x4a4(%rsp) # imm = 0xFFFFFFFF
movl 0x4a4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x4a0(%rsp)
cmpl $0x1, 0x4a0(%rsp)
jne 0x1633965
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1633936
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1633934
jmp 0x1633963
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4c8(%rsp)
cmpq $0x0, 0x4c8(%rsp)
je 0x1633961
movq 0x4c8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633963
jmp 0x1633965
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16339bd
movq %rax, %rdi
callq 0x678a0
jmp 0x16339bf
leaq 0x190(%rsp), %rax
movq %rax, 0x2b0(%rsp)
movq 0x2b0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x30(%rsp)
leaq 0x190(%rsp), %rax
movq %rax, 0x2d8(%rsp)
movq 0x2d8(%rsp), %rax
movq %rax, 0x488(%rsp)
movq 0x488(%rsp), %rax
movq %rax, 0x38(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1633a9e
movq 0x38(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x484(%rsp) # imm = 0xFFFFFFFF
movl 0x484(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x480(%rsp)
cmpl $0x1, 0x480(%rsp)
jne 0x1633a9e
movq 0x38(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1633a6f
movq 0x38(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1633a6d
jmp 0x1633a9c
movq 0x38(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4d8(%rsp)
cmpq $0x0, 0x4d8(%rsp)
je 0x1633a9a
movq 0x4d8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633a9c
jmp 0x1633a9e
movq 0x38(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1633af6
movq %rax, %rdi
callq 0x678a0
movq 0x30(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x258(%rsp), %rcx
movl 0x230(%rsp), %eax
leaq 0x140(%rsp), %rdx
movq %rdx, 0x2f8(%rsp)
movq %rcx, 0x2f0(%rsp)
movl %eax, 0x2ec(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0x2eb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x2ec(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x140(%rsp), %r10
movq %r10, 0x3b8(%rsp)
movl %r9d, 0x3b4(%rsp)
movl %r8d, 0x3b0(%rsp)
movl %edi, 0x3ac(%rsp)
movq %rsi, 0x3a0(%rsp)
movq %rdx, 0x398(%rsp)
movl %ecx, 0x394(%rsp)
movq %rax, 0x388(%rsp)
movq 0x3b8(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x398(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x394(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x388(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x3b4(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x3b0(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x3ac(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x458(%rsp)
movl $0x10, 0x454(%rsp)
movq 0x458(%rsp), %rax
movslq 0x454(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x454(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x168(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x1633cb3
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x180(%rsp)
movb $0x1, 0x2eb(%rsp)
testb $0x1, 0x2eb(%rsp)
jne 0x1633de2
leaq 0x140(%rsp), %rax
movq %rax, 0x300(%rsp)
movq 0x300(%rsp), %rax
movq %rax, 0x468(%rsp)
movq 0x468(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1633d88
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x464(%rsp) # imm = 0xFFFFFFFF
movl 0x464(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x460(%rsp)
cmpl $0x1, 0x460(%rsp)
jne 0x1633d88
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1633d59
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1633d57
jmp 0x1633d86
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4e8(%rsp)
cmpq $0x0, 0x4e8(%rsp)
je 0x1633d84
movq 0x4e8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633d86
jmp 0x1633d88
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1633de0
movq %rax, %rdi
callq 0x678a0
jmp 0x1633de2
leaq 0x140(%rsp), %rax
movq %rax, 0x308(%rsp)
movq 0x308(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x140(%rsp), %rax
movq %rax, 0x2e0(%rsp)
movq 0x2e0(%rsp), %rax
movq %rax, 0x478(%rsp)
movq 0x478(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1633ec1
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x474(%rsp) # imm = 0xFFFFFFFF
movl 0x474(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x470(%rsp)
cmpl $0x1, 0x470(%rsp)
jne 0x1633ec1
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1633e92
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1633e90
jmp 0x1633ebf
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x4e0(%rsp)
cmpq $0x0, 0x4e0(%rsp)
je 0x1633ebd
movq 0x4e0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1633ebf
jmp 0x1633ec1
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1633f19
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x188(%rsp)
movl $0x0, 0x13c(%rsp)
movl 0x13c(%rsp), %eax
addl $0x7, %eax
cmpl 0x234(%rsp), %eax
jge 0x1634041
movq 0x228(%rsp), %rax
movq %rax, 0x318(%rsp)
movq 0x318(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x100(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x310(%rsp)
movq 0x310(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
leaq 0x24f(%rsp), %rdi
leaq 0x100(%rsp), %rsi
leaq 0xe0(%rsp), %rdx
callq 0x16361c0
vmovaps %ymm0, 0xc0(%rsp)
movq 0x188(%rsp), %rax
vmovaps 0xc0(%rsp), %ymm0
movq %rax, 0x348(%rsp)
vmovaps %ymm0, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm0
movq 0x348(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1633f31
jmp 0x1634043
movl 0x13c(%rsp), %eax
addl $0x3, %eax
cmpl 0x234(%rsp), %eax
jge 0x1634156
movq 0x228(%rsp), %rax
movq %rax, 0x358(%rsp)
movq 0x358(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0xb0(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x350(%rsp)
movq 0x350(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0xa0(%rsp)
leaq 0x24f(%rsp), %rdi
leaq 0xb0(%rsp), %rsi
leaq 0xa0(%rsp), %rdx
vzeroupper
callq 0x1637b00
vmovaps %xmm0, 0x90(%rsp)
movq 0x188(%rsp), %rax
vmovaps 0x90(%rsp), %xmm0
movq %rax, 0x378(%rsp)
vmovaps %xmm0, 0x360(%rsp)
vmovaps 0x360(%rsp), %xmm0
movq 0x378(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1634043
jmp 0x1634158
movl 0x13c(%rsp), %eax
cmpl 0x234(%rsp), %eax
jge 0x16341e6
movq 0x228(%rsp), %rsi
movq 0x1d8(%rsp), %rdx
leaq 0x24f(%rsp), %rdi
vzeroupper
callq 0x1638fe0
movq 0x188(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x228(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x228(%rsp)
movq 0x1d8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1d8(%rsp)
movq 0x188(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x188(%rsp)
movl 0x13c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x13c(%rsp)
jmp 0x1634158
jmp 0x16341e8
movl 0x230(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x230(%rsp)
jmp 0x16332a9
movl $0x0, 0x274(%rsp)
movl 0x274(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,105 |
int ncnn::binary_op_6_11_16_25<ncnn::BinaryOp_x86_avx_functor::binary_op_pow>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_6_11_16_25(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
// type 6 11 16 25
c.create_like(a, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float* ptr = a.channel(q);
const float b0 = b[0];
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
__m512 _b0_avx512 = _mm512_set1_ps(b0);
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _outp = op.func_pack16(_p, _b0_avx512);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
outptr += 16;
}
#endif // __AVX512F__
__m256 _b0_avx = _mm256_set1_ps(b0);
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _outp = op.func_pack8(_p, _b0_avx);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
outptr += 8;
}
#endif // __AVX__
__m128 _b0 = _mm_set1_ps(b0);
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _outp = op.func_pack4(_p, _b0);
_mm_store_ps(outptr, _outp);
ptr += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(*ptr, b0);
ptr += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x460, %rsp # imm = 0x460
movq %rdi, 0x200(%rsp)
movq %rsi, 0x1f8(%rsp)
movq %rdx, 0x1f0(%rsp)
movq %rcx, 0x1e8(%rsp)
movq 0x200(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1e0(%rsp)
movq 0x200(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1dc(%rsp)
movq 0x200(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x1d8(%rsp)
movq 0x200(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x1d4(%rsp)
movq 0x200(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x1d0(%rsp)
movl 0x1e0(%rsp), %eax
imull 0x1dc(%rsp), %eax
imull 0x1d8(%rsp), %eax
imull 0x1d0(%rsp), %eax
movl %eax, 0x1cc(%rsp)
movq 0x1f0(%rsp), %rdi
movq 0x200(%rsp), %rsi
movq 0x1e8(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x1f0(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rcx
movq %rcx, 0x60(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x6f(%rsp)
je 0x1634345
movq 0x60(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x6f(%rsp)
movb 0x6f(%rsp), %al
testb $0x1, %al
jne 0x163434f
jmp 0x163435f
movl $0xffffff9c, 0x20c(%rsp) # imm = 0xFFFFFF9C
jmp 0x1634f84
movl $0x0, 0x1c8(%rsp)
movl 0x1c8(%rsp), %eax
cmpl 0x1d4(%rsp), %eax
jge 0x1634f79
movq 0x200(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x178(%rsp), %rdx
movq %rdx, 0x228(%rsp)
movq %rcx, 0x220(%rsp)
movl %eax, 0x21c(%rsp)
movq 0x220(%rsp), %rax
movq %rax, 0x58(%rsp)
movb $0x0, 0x21b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x21c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x178(%rsp), %r10
movq %r10, 0x340(%rsp)
movl %r9d, 0x33c(%rsp)
movl %r8d, 0x338(%rsp)
movl %edi, 0x334(%rsp)
movq %rsi, 0x328(%rsp)
movq %rdx, 0x320(%rsp)
movl %ecx, 0x31c(%rsp)
movq %rax, 0x310(%rsp)
movq 0x340(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x328(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x320(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x31c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x310(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x33c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x338(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x334(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x350(%rsp)
movl $0x10, 0x34c(%rsp)
movq 0x350(%rsp), %rax
movslq 0x34c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x34c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x58(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x1a0(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163452e
movq 0x58(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1b8(%rsp)
movb $0x1, 0x21b(%rsp)
testb $0x1, 0x21b(%rsp)
jne 0x163465d
leaq 0x178(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
movq %rax, 0x3a0(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, 0x48(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1634603
movq 0x48(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x39c(%rsp) # imm = 0xFFFFFFFF
movl 0x39c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x398(%rsp)
cmpl $0x1, 0x398(%rsp)
jne 0x1634603
movq 0x48(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16345d4
movq 0x48(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16345d2
jmp 0x1634601
movq 0x48(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3a8(%rsp)
cmpq $0x0, 0x3a8(%rsp)
je 0x16345ff
movq 0x3a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1634601
jmp 0x1634603
movq 0x48(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163465b
movq %rax, %rdi
callq 0x678a0
jmp 0x163465d
leaq 0x178(%rsp), %rax
movq %rax, 0x230(%rsp)
movq 0x230(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x38(%rsp)
leaq 0x178(%rsp), %rax
movq %rax, 0x240(%rsp)
movq 0x240(%rsp), %rax
movq %rax, 0x390(%rsp)
movq 0x390(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163473c
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x38c(%rsp) # imm = 0xFFFFFFFF
movl 0x38c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x388(%rsp)
cmpl $0x1, 0x388(%rsp)
jne 0x163473c
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163470d
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x163470b
jmp 0x163473a
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b0(%rsp)
cmpq $0x0, 0x3b0(%rsp)
je 0x1634738
movq 0x3b0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x163473a
jmp 0x163473c
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1634794
movq %rax, %rdi
callq 0x678a0
movq 0x38(%rsp), %rax
movq %rax, 0x1c0(%rsp)
movq 0x1f8(%rsp), %rax
movq %rax, 0x3d0(%rsp)
movq $0x0, 0x3c8(%rsp)
movq 0x3d0(%rsp), %rax
movq (%rax), %rax
movq 0x3c8(%rsp), %rcx
vmovss (%rax,%rcx,4), %xmm0
vmovss %xmm0, 0x174(%rsp)
movq 0x1f0(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x120(%rsp), %rdx
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movl %eax, 0x254(%rsp)
movq 0x258(%rsp), %rax
movq %rax, 0x30(%rsp)
movb $0x0, 0x253(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x254(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x120(%rsp), %r10
movq %r10, 0x308(%rsp)
movl %r9d, 0x304(%rsp)
movl %r8d, 0x300(%rsp)
movl %edi, 0x2fc(%rsp)
movq %rsi, 0x2f0(%rsp)
movq %rdx, 0x2e8(%rsp)
movl %ecx, 0x2e4(%rsp)
movq %rax, 0x2d8(%rsp)
movq 0x308(%rsp), %rcx
movq %rcx, 0x28(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2e8(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2e4(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2d8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x304(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x300(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2fc(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x360(%rsp)
movl $0x10, 0x35c(%rsp)
movq 0x360(%rsp), %rax
movslq 0x35c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x35c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x30(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x148(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163498e
movq 0x30(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x160(%rsp)
movb $0x1, 0x253(%rsp)
testb $0x1, 0x253(%rsp)
jne 0x1634abd
leaq 0x120(%rsp), %rax
movq %rax, 0x268(%rsp)
movq 0x268(%rsp), %rax
movq %rax, 0x370(%rsp)
movq 0x370(%rsp), %rax
movq %rax, 0x20(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1634a63
movq 0x20(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x36c(%rsp) # imm = 0xFFFFFFFF
movl 0x36c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x368(%rsp)
cmpl $0x1, 0x368(%rsp)
jne 0x1634a63
movq 0x20(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1634a34
movq 0x20(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1634a32
jmp 0x1634a61
movq 0x20(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3c0(%rsp)
cmpq $0x0, 0x3c0(%rsp)
je 0x1634a5f
movq 0x3c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1634a61
jmp 0x1634a63
movq 0x20(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1634abb
movq %rax, %rdi
callq 0x678a0
jmp 0x1634abd
leaq 0x120(%rsp), %rax
movq %rax, 0x270(%rsp)
movq 0x270(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x120(%rsp), %rax
movq %rax, 0x248(%rsp)
movq 0x248(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1634b9c
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x37c(%rsp) # imm = 0xFFFFFFFF
movl 0x37c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x378(%rsp)
cmpl $0x1, 0x378(%rsp)
jne 0x1634b9c
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1634b6d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x1634b6b
jmp 0x1634b9a
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b8(%rsp)
cmpq $0x0, 0x3b8(%rsp)
je 0x1634b98
movq 0x3b8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x1634b9a
jmp 0x1634b9c
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1634bf4
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x168(%rsp)
movl $0x0, 0x11c(%rsp)
vmovss 0x174(%rsp), %xmm0
vmovss %xmm0, 0x3dc(%rsp)
vmovss 0x3dc(%rsp), %xmm0
vmovss %xmm0, 0xc(%rsp)
vmovss %xmm0, 0x44c(%rsp)
vmovss %xmm0, 0x448(%rsp)
vmovss %xmm0, 0x444(%rsp)
vmovss %xmm0, 0x440(%rsp)
vmovss %xmm0, 0x43c(%rsp)
vmovss %xmm0, 0x438(%rsp)
vmovss %xmm0, 0x434(%rsp)
vmovss %xmm0, 0x430(%rsp)
vmovss 0x444(%rsp), %xmm1
vmovss 0x440(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x448(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x44c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm1 # xmm1 = xmm0[0,1,2],xmm1[0]
vmovss 0x434(%rsp), %xmm2
vmovss 0x430(%rsp), %xmm0
vinsertps $0x10, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
vmovss 0x438(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
vmovss 0x43c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm0, %xmm2 # xmm2 = xmm0[0,1,2],xmm2[0]
vmovaps %xmm2, %xmm0
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl 0x11c(%rsp), %eax
addl $0x7, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1634ddd
movq 0x1c0(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xc0(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0xc0(%rsp), %rsi
leaq 0xe0(%rsp), %rdx
callq 0x16361c0
vmovaps %ymm0, 0xa0(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0xa0(%rsp), %ymm0
movq %rax, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
movq 0x2a0(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1634d06
vmovss 0x174(%rsp), %xmm0
vmovss %xmm0, 0x3fc(%rsp)
vbroadcastss 0x3fc(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
movl 0x11c(%rsp), %eax
addl $0x3, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1634ee8
movq 0x1c0(%rsp), %rax
movq %rax, 0x2a8(%rsp)
movq 0x2a8(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x80(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0x80(%rsp), %rsi
leaq 0x90(%rsp), %rdx
vzeroupper
callq 0x1637b00
vmovaps %xmm0, 0x70(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0x70(%rsp), %xmm0
movq %rax, 0x2c8(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm0
movq 0x2c8(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1634e14
jmp 0x1634eea
movl 0x11c(%rsp), %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1634f61
movq 0x1c0(%rsp), %rsi
leaq 0x1e7(%rsp), %rdi
leaq 0x174(%rsp), %rdx
vzeroupper
callq 0x1638fe0
movq 0x168(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x1c0(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1c0(%rsp)
movq 0x168(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1634eea
jmp 0x1634f63
movl 0x1c8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c8(%rsp)
jmp 0x163436a
movl $0x0, 0x20c(%rsp)
movl 0x20c(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,106 |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_avx_functor::binary_op_pow>(ncnn::Mat const&, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&)
|
static int binary_op_2_3_4_20(const Mat& a, const Mat& b, Mat& c, const Option& opt)
{
Op op;
int w = b.w;
int h = b.h;
int d = b.d;
int channels = b.c;
int elempack = b.elempack;
int size = w * h * d * elempack;
// type 2 3 4 20
c.create_like(b, opt.blob_allocator);
if (c.empty())
return -100;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
const float a0 = a[0];
const float* ptr = b.channel(q);
float* outptr = c.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
__m512 _a0_avx512 = _mm512_set1_ps(a0);
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
__m512 _outp = op.func_pack16(_a0_avx512, _p);
_mm512_storeu_ps(outptr, _outp);
ptr += 16;
outptr += 16;
}
#endif // __AVX512F__
__m256 _a0_avx = _mm256_set1_ps(a0);
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
__m256 _outp = op.func_pack8(_a0_avx, _p);
_mm256_storeu_ps(outptr, _outp);
ptr += 8;
outptr += 8;
}
#endif // __AVX__
__m128 _a0 = _mm_set1_ps(a0);
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
__m128 _outp = op.func_pack4(_a0, _p);
_mm_store_ps(outptr, _outp);
ptr += 4;
outptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*outptr = op.func(a0, *ptr);
ptr += 1;
outptr += 1;
}
}
return 0;
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x460, %rsp # imm = 0x460
movq %rdi, 0x200(%rsp)
movq %rsi, 0x1f8(%rsp)
movq %rdx, 0x1f0(%rsp)
movq %rcx, 0x1e8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1e0(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1dc(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0x1d8(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x1d4(%rsp)
movq 0x1f8(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0x1d0(%rsp)
movl 0x1e0(%rsp), %eax
imull 0x1dc(%rsp), %eax
imull 0x1d8(%rsp), %eax
imull 0x1d0(%rsp), %eax
movl %eax, 0x1cc(%rsp)
movq 0x1f0(%rsp), %rdi
movq 0x1f8(%rsp), %rsi
movq 0x1e8(%rsp), %rax
movq 0x8(%rax), %rdx
callq 0x6fe40
movq 0x1f0(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rcx
movq %rcx, 0x60(%rsp)
movb $0x1, %al
cmpq $0x0, (%rcx)
movb %al, 0x6f(%rsp)
je 0x16350c5
movq 0x60(%rsp), %rax
movq %rax, 0x2d0(%rsp)
movq 0x2d0(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
cmpq $0x0, %rax
sete %al
movb %al, 0x6f(%rsp)
movb 0x6f(%rsp), %al
testb $0x1, %al
jne 0x16350cf
jmp 0x16350df
movl $0xffffff9c, 0x20c(%rsp) # imm = 0xFFFFFF9C
jmp 0x1635d04
movl $0x0, 0x1c8(%rsp)
movl 0x1c8(%rsp), %eax
cmpl 0x1d4(%rsp), %eax
jge 0x1635cf9
movq 0x200(%rsp), %rax
movq %rax, 0x3d0(%rsp)
movq $0x0, 0x3c8(%rsp)
movq 0x3d0(%rsp), %rax
movq (%rax), %rax
movq 0x3c8(%rsp), %rcx
vmovss (%rax,%rcx,4), %xmm0
vmovss %xmm0, 0x1c4(%rsp)
movq 0x1f8(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x170(%rsp), %rdx
movq %rdx, 0x228(%rsp)
movq %rcx, 0x220(%rsp)
movl %eax, 0x21c(%rsp)
movq 0x220(%rsp), %rax
movq %rax, 0x58(%rsp)
movb $0x0, 0x21b(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x21c(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x170(%rsp), %r10
movq %r10, 0x340(%rsp)
movl %r9d, 0x33c(%rsp)
movl %r8d, 0x338(%rsp)
movl %edi, 0x334(%rsp)
movq %rsi, 0x328(%rsp)
movq %rdx, 0x320(%rsp)
movl %ecx, 0x31c(%rsp)
movq %rax, 0x310(%rsp)
movq 0x340(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x328(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x320(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x31c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x310(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x33c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x338(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x334(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x350(%rsp)
movl $0x10, 0x34c(%rsp)
movq 0x350(%rsp), %rax
movslq 0x34c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x34c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x58(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x198(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x16352eb
movq 0x58(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x1b0(%rsp)
movb $0x1, 0x21b(%rsp)
testb $0x1, 0x21b(%rsp)
jne 0x163541a
leaq 0x170(%rsp), %rax
movq %rax, 0x238(%rsp)
movq 0x238(%rsp), %rax
movq %rax, 0x3a0(%rsp)
movq 0x3a0(%rsp), %rax
movq %rax, 0x48(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16353c0
movq 0x48(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x39c(%rsp) # imm = 0xFFFFFFFF
movl 0x39c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x398(%rsp)
cmpl $0x1, 0x398(%rsp)
jne 0x16353c0
movq 0x48(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1635391
movq 0x48(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x163538f
jmp 0x16353be
movq 0x48(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3a8(%rsp)
cmpq $0x0, 0x3a8(%rsp)
je 0x16353bc
movq 0x3a8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16353be
jmp 0x16353c0
movq 0x48(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1635418
movq %rax, %rdi
callq 0x678a0
jmp 0x163541a
leaq 0x170(%rsp), %rax
movq %rax, 0x230(%rsp)
movq 0x230(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x38(%rsp)
leaq 0x170(%rsp), %rax
movq %rax, 0x240(%rsp)
movq 0x240(%rsp), %rax
movq %rax, 0x390(%rsp)
movq 0x390(%rsp), %rax
movq %rax, 0x40(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16354f9
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x38c(%rsp) # imm = 0xFFFFFFFF
movl 0x38c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x388(%rsp)
cmpl $0x1, 0x388(%rsp)
jne 0x16354f9
movq 0x40(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16354ca
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16354c8
jmp 0x16354f7
movq 0x40(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b0(%rsp)
cmpq $0x0, 0x3b0(%rsp)
je 0x16354f5
movq 0x3b0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16354f7
jmp 0x16354f9
movq 0x40(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1635551
movq %rax, %rdi
callq 0x678a0
movq 0x38(%rsp), %rax
movq %rax, 0x1b8(%rsp)
movq 0x1f0(%rsp), %rcx
movl 0x1c8(%rsp), %eax
leaq 0x120(%rsp), %rdx
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movl %eax, 0x254(%rsp)
movq 0x258(%rsp), %rax
movq %rax, 0x30(%rsp)
movb $0x0, 0x253(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x254(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x120(%rsp), %r10
movq %r10, 0x308(%rsp)
movl %r9d, 0x304(%rsp)
movl %r8d, 0x300(%rsp)
movl %edi, 0x2fc(%rsp)
movq %rsi, 0x2f0(%rsp)
movq %rdx, 0x2e8(%rsp)
movl %ecx, 0x2e4(%rsp)
movq %rax, 0x2d8(%rsp)
movq 0x308(%rsp), %rcx
movq %rcx, 0x28(%rsp)
movq 0x2f0(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x2e8(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x2e4(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x2d8(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x304(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x300(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x2fc(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x360(%rsp)
movl $0x10, 0x35c(%rsp)
movq 0x360(%rsp), %rax
movslq 0x35c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x35c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x30(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x148(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163570e
movq 0x30(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x160(%rsp)
movb $0x1, 0x253(%rsp)
testb $0x1, 0x253(%rsp)
jne 0x163583d
leaq 0x120(%rsp), %rax
movq %rax, 0x268(%rsp)
movq 0x268(%rsp), %rax
movq %rax, 0x370(%rsp)
movq 0x370(%rsp), %rax
movq %rax, 0x20(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x16357e3
movq 0x20(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x36c(%rsp) # imm = 0xFFFFFFFF
movl 0x36c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x368(%rsp)
cmpl $0x1, 0x368(%rsp)
jne 0x16357e3
movq 0x20(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16357b4
movq 0x20(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16357b2
jmp 0x16357e1
movq 0x20(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3c0(%rsp)
cmpq $0x0, 0x3c0(%rsp)
je 0x16357df
movq 0x3c0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x16357e1
jmp 0x16357e3
movq 0x20(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163583b
movq %rax, %rdi
callq 0x678a0
jmp 0x163583d
leaq 0x120(%rsp), %rax
movq %rax, 0x270(%rsp)
movq 0x270(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x120(%rsp), %rax
movq %rax, 0x248(%rsp)
movq 0x248(%rsp), %rax
movq %rax, 0x380(%rsp)
movq 0x380(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163591c
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x37c(%rsp) # imm = 0xFFFFFFFF
movl 0x37c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x378(%rsp)
cmpl $0x1, 0x378(%rsp)
jne 0x163591c
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x16358ed
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x16358eb
jmp 0x163591a
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x3b8(%rsp)
cmpq $0x0, 0x3b8(%rsp)
je 0x1635918
movq 0x3b8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x163591a
jmp 0x163591c
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1635974
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x168(%rsp)
movl $0x0, 0x11c(%rsp)
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3dc(%rsp)
vmovss 0x3dc(%rsp), %xmm0
vmovss %xmm0, 0xc(%rsp)
vmovss %xmm0, 0x44c(%rsp)
vmovss %xmm0, 0x448(%rsp)
vmovss %xmm0, 0x444(%rsp)
vmovss %xmm0, 0x440(%rsp)
vmovss %xmm0, 0x43c(%rsp)
vmovss %xmm0, 0x438(%rsp)
vmovss %xmm0, 0x434(%rsp)
vmovss %xmm0, 0x430(%rsp)
vmovss 0x444(%rsp), %xmm1
vmovss 0x440(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x448(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x44c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm1 # xmm1 = xmm0[0,1,2],xmm1[0]
vmovss 0x434(%rsp), %xmm2
vmovss 0x430(%rsp), %xmm0
vinsertps $0x10, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm2[0],xmm0[2,3]
vmovss 0x438(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm2[0],xmm0[3]
vmovss 0x43c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm0, %xmm2 # xmm2 = xmm0[0,1,2],xmm2[0]
vmovaps %xmm2, %xmm0
vinsertf128 $0x1, %xmm1, %ymm0, %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl 0x11c(%rsp), %eax
addl $0x7, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1635b5d
movq 0x1b8(%rsp), %rax
movq %rax, 0x278(%rsp)
movq 0x278(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0xc0(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0xe0(%rsp), %rsi
leaq 0xc0(%rsp), %rdx
callq 0x16361c0
vmovaps %ymm0, 0xa0(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0xa0(%rsp), %ymm0
movq %rax, 0x2a0(%rsp)
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
movq 0x2a0(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x20, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1635a86
vmovss 0x1c4(%rsp), %xmm0
vmovss %xmm0, 0x3fc(%rsp)
vbroadcastss 0x3fc(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %xmm0
vmovaps %xmm0, 0x90(%rsp)
movl 0x11c(%rsp), %eax
addl $0x3, %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1635c68
movq 0x1b8(%rsp), %rax
movq %rax, 0x2a8(%rsp)
movq 0x2a8(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x80(%rsp)
leaq 0x1e7(%rsp), %rdi
leaq 0x90(%rsp), %rsi
leaq 0x80(%rsp), %rdx
vzeroupper
callq 0x1637b00
vmovaps %xmm0, 0x70(%rsp)
movq 0x168(%rsp), %rax
vmovaps 0x70(%rsp), %xmm0
movq %rax, 0x2c8(%rsp)
vmovaps %xmm0, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm0
movq 0x2c8(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1635b94
jmp 0x1635c6a
movl 0x11c(%rsp), %eax
cmpl 0x1cc(%rsp), %eax
jge 0x1635ce1
movq 0x1b8(%rsp), %rdx
leaq 0x1e7(%rsp), %rdi
leaq 0x1c4(%rsp), %rsi
vzeroupper
callq 0x1638fe0
movq 0x168(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0x1b8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x1b8(%rsp)
movq 0x168(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x168(%rsp)
movl 0x11c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x11c(%rsp)
jmp 0x1635c6a
jmp 0x1635ce3
movl 0x1c8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c8(%rsp)
jmp 0x16350ea
movl $0x0, 0x20c(%rsp)
movl 0x20c(%rsp), %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,107 |
ncnn::BinaryOp_x86_avx_functor::binary_op_add::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_add_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vaddps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,108 |
ncnn::BinaryOp_x86_avx_functor::binary_op_add::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_add_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vaddps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,109 |
ncnn::BinaryOp_x86_avx_functor::binary_op_add::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return x + y;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x18(%rsp), %rax
vaddss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,110 |
ncnn::BinaryOp_x86_avx_functor::binary_op_sub::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_sub_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vsubps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,111 |
ncnn::BinaryOp_x86_avx_functor::binary_op_sub::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_sub_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vsubps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,112 |
ncnn::BinaryOp_x86_avx_functor::binary_op_sub::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return x - y;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x18(%rsp), %rax
vsubss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,113 |
ncnn::BinaryOp_x86_avx_functor::binary_op_mul::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_mul_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vmulps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,114 |
ncnn::BinaryOp_x86_avx_functor::binary_op_mul::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_mul_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vmulps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,115 |
ncnn::BinaryOp_x86_avx_functor::binary_op_mul::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return x * y;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x18(%rsp), %rax
vmulss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,116 |
ncnn::BinaryOp_x86_avx_functor::binary_op_div::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_div_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vdivps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,117 |
ncnn::BinaryOp_x86_avx_functor::binary_op_div::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_div_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vdivps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,118 |
ncnn::BinaryOp_x86_avx_functor::binary_op_div::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return x / y;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x10(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x18(%rsp), %rax
vdivss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,119 |
ncnn::BinaryOp_x86_avx_functor::binary_op_max::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_max_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vmovaps 0x20(%rsp), %ymm1
vmaxps %ymm1, %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,120 |
ncnn::BinaryOp_x86_avx_functor::binary_op_max::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_max_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vmovaps -0x28(%rsp), %xmm1
vmaxps %xmm1, %xmm0, %xmm0
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,121 |
ncnn::BinaryOp_x86_avx_functor::binary_op_max::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return std::max(x, y);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq %rdx, (%rsp)
movq 0x8(%rsp), %rdi
movq (%rsp), %rsi
callq 0x670b0
vmovss (%rax), %xmm0
addq $0x18, %rsp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,122 |
ncnn::BinaryOp_x86_avx_functor::binary_op_min::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_min_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vmovaps 0x20(%rsp), %ymm1
vminps %ymm1, %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,123 |
ncnn::BinaryOp_x86_avx_functor::binary_op_min::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_min_ps(x, y);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vmovaps -0x28(%rsp), %xmm1
vminps %xmm1, %xmm0, %xmm0
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,124 |
ncnn::BinaryOp_x86_avx_functor::binary_op_min::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return std::min(x, y);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq %rdx, (%rsp)
movq 0x8(%rsp), %rdi
movq (%rsp), %rsi
callq 0x670f0
vmovss (%rax), %xmm0
addq $0x18, %rsp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,125 |
ncnn::BinaryOp_x86_avx_functor::binary_op_pow::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return pow256_ps(x, y);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x1ee0, %rsp # imm = 0x1EE0
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %ymm7
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x7d68e7(%rip), %ymm2 # 0x1e0cb00
vmovaps %ymm2, 0x3e0(%rsp)
vmovaps 0x420(%rsp), %ymm0
vxorps %xmm5, %xmm5, %xmm5
vmovaps %ymm5, 0x680(%rsp)
vmovaps 0x680(%rsp), %ymm1
vcmpleps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x3c0(%rsp)
vmovaps 0x420(%rsp), %ymm0
vmovaps %ymm0, 0x660(%rsp)
vmovaps 0x7d6a37(%rip), %ymm0 # 0x1e0cca0
vmovaps %ymm0, 0x640(%rsp)
vmovaps 0x660(%rsp), %ymm0
vmovaps 0x640(%rsp), %ymm1
vmaxps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x420(%rsp), %ymm0
vmovaps %ymm0, 0x1cc0(%rsp)
vmovaps 0x1cc0(%rsp), %ymm0
vmovaps %ymm0, 0x1c60(%rsp)
movl $0x17, 0x1c5c(%rsp)
vmovaps 0x1c60(%rsp), %ymm0
vmovaps %ymm0, 0x1be0(%rsp)
vmovdqa 0x1be0(%rsp), %xmm0
vmovdqa %xmm0, 0x1c40(%rsp)
vmovdqa 0x1bf0(%rsp), %xmm0
vmovdqa %xmm0, 0x1c30(%rsp)
vmovdqa 0x1c40(%rsp), %xmm0
movl 0x1c5c(%rsp), %eax
vmovdqa %xmm0, 0x1cb0(%rsp)
movl %eax, 0x1cac(%rsp)
vmovdqa 0x1cb0(%rsp), %xmm0
vmovd 0x1cac(%rsp), %xmm1
vpsrld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1c40(%rsp)
vmovdqa 0x1c30(%rsp), %xmm0
movl 0x1c5c(%rsp), %eax
vmovdqa %xmm0, 0x1c90(%rsp)
movl %eax, 0x1c8c(%rsp)
vmovdqa 0x1c90(%rsp), %xmm0
vmovd 0x1c8c(%rsp), %xmm1
vpsrld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1c30(%rsp)
vmovdqa 0x1c40(%rsp), %xmm0
vmovdqa %xmm0, 0x1bc0(%rsp)
vmovdqa 0x1c30(%rsp), %xmm0
vmovdqa %xmm0, 0x1bd0(%rsp)
vmovaps 0x1bc0(%rsp), %ymm0
vmovaps %ymm0, 0x1c00(%rsp)
vmovaps 0x1c00(%rsp), %ymm0
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x420(%rsp), %ymm0
vmovaps %ymm0, 0x1640(%rsp)
vmovaps 0x7d68ea(%rip), %ymm0 # 0x1e0ccc0
vmovaps %ymm0, 0x1620(%rsp)
vmovaps 0x1640(%rsp), %ymm0
vmovaps 0x1620(%rsp), %ymm1
vandps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x420(%rsp), %ymm0
vmovaps %ymm0, 0x1d40(%rsp)
vmovaps 0x7d6768(%rip), %ymm1 # 0x1e0cb80
vmovaps %ymm1, 0x1d20(%rsp)
vmovaps 0x1d40(%rsp), %ymm0
vmovaps 0x1d20(%rsp), %ymm3
vorps %ymm3, %ymm0, %ymm0
vmovaps %ymm0, 0x420(%rsp)
vmovaps 0x400(%rsp), %ymm0
vmovaps %ymm0, 0x1e40(%rsp)
vmovaps 0x7d6826(%rip), %ymm0 # 0x1e0cc80
vmovaps %ymm0, 0x1e20(%rsp)
vmovaps 0x1e40(%rsp), %ymm3
vmovaps %ymm3, 0x1da0(%rsp)
vmovdqa 0x1da0(%rsp), %xmm3
vmovdqa %xmm3, 0x1e10(%rsp)
vmovdqa 0x1db0(%rsp), %xmm3
vmovdqa %xmm3, 0x1e00(%rsp)
vmovaps 0x1e20(%rsp), %ymm3
vmovaps %ymm3, 0x1d80(%rsp)
vmovdqa 0x1d80(%rsp), %xmm3
vmovdqa %xmm3, 0x1df0(%rsp)
vmovdqa 0x1d90(%rsp), %xmm3
vmovdqa %xmm3, 0x1de0(%rsp)
vmovdqa 0x1e10(%rsp), %xmm4
vmovdqa 0x1df0(%rsp), %xmm3
vmovdqa %xmm4, 0x1e90(%rsp)
vmovdqa %xmm3, 0x1e80(%rsp)
vmovdqa 0x1e90(%rsp), %xmm3
vmovdqa 0x1e80(%rsp), %xmm4
vpsubd %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0x1e10(%rsp)
vmovdqa 0x1e00(%rsp), %xmm4
vmovdqa 0x1de0(%rsp), %xmm3
vmovdqa %xmm4, 0x1e70(%rsp)
vmovdqa %xmm3, 0x1e60(%rsp)
vmovdqa 0x1e70(%rsp), %xmm3
vmovdqa 0x1e60(%rsp), %xmm4
vpsubd %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0x1e00(%rsp)
vmovdqa 0x1e10(%rsp), %xmm3
vmovdqa %xmm3, 0x1d60(%rsp)
vmovdqa 0x1e00(%rsp), %xmm3
vmovdqa %xmm3, 0x1d70(%rsp)
vmovaps 0x1d60(%rsp), %ymm3
vmovaps %ymm3, 0x1dc0(%rsp)
vmovaps 0x1dc0(%rsp), %ymm3
vmovaps %ymm3, 0x400(%rsp)
vmovaps 0x400(%rsp), %ymm3
vmovaps %ymm3, 0x1ea0(%rsp)
vcvtdq2ps 0x1ea0(%rsp), %ymm3
vmovaps %ymm3, 0x3a0(%rsp)
vmovaps 0x3a0(%rsp), %ymm4
vmovaps 0x3e0(%rsp), %ymm3
vmovaps %ymm4, 0x4e0(%rsp)
vmovaps %ymm3, 0x4c0(%rsp)
vmovaps 0x4e0(%rsp), %ymm3
vmovaps 0x4c0(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x3a0(%rsp)
vmovaps 0x420(%rsp), %ymm3
vmovaps 0x7d66cb(%rip), %ymm4 # 0x1e0cce0
vcmpltps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x380(%rsp)
vmovaps 0x420(%rsp), %ymm4
vmovaps 0x380(%rsp), %ymm3
vmovaps %ymm4, 0x1600(%rsp)
vmovaps %ymm3, 0x15e0(%rsp)
vmovaps 0x1600(%rsp), %ymm3
vmovaps 0x15e0(%rsp), %ymm4
vandps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x360(%rsp)
vmovaps 0x420(%rsp), %ymm4
vmovaps 0x3e0(%rsp), %ymm3
vmovaps %ymm4, 0x560(%rsp)
vmovaps %ymm3, 0x540(%rsp)
vmovaps 0x560(%rsp), %ymm3
vmovaps 0x540(%rsp), %ymm4
vsubps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x420(%rsp)
vmovaps 0x3a0(%rsp), %ymm4
vmovaps 0x3e0(%rsp), %ymm6
vmovaps 0x380(%rsp), %ymm3
vmovaps %ymm6, 0x15c0(%rsp)
vmovaps %ymm3, 0x15a0(%rsp)
vmovaps 0x15c0(%rsp), %ymm3
vmovaps 0x15a0(%rsp), %ymm6
vandps %ymm6, %ymm3, %ymm3
vmovaps %ymm4, 0x520(%rsp)
vmovaps %ymm3, 0x500(%rsp)
vmovaps 0x520(%rsp), %ymm3
vmovaps 0x500(%rsp), %ymm4
vsubps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x3a0(%rsp)
vmovaps 0x420(%rsp), %ymm4
vmovaps 0x360(%rsp), %ymm3
vmovaps %ymm4, 0x4a0(%rsp)
vmovaps %ymm3, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %ymm3
vmovaps 0x480(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x420(%rsp)
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm3, 0x620(%rsp)
vmovaps %ymm3, 0x600(%rsp)
vmovaps 0x620(%rsp), %ymm3
vmovaps 0x600(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x340(%rsp)
vmovaps 0x7d655e(%rip), %ymm3 # 0x1e0cd00
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xee0(%rsp)
vmovaps %ymm3, 0xec0(%rsp)
vmovaps 0x7d6549(%rip), %ymm3 # 0x1e0cd20
vmovaps %ymm3, 0xea0(%rsp)
vmovaps 0xee0(%rsp), %ymm4
vmovaps 0xec0(%rsp), %ymm3
vmovaps %ymm4, 0xf60(%rsp)
vmovaps %ymm3, 0xf40(%rsp)
vmovaps 0xf60(%rsp), %ymm3
vmovaps 0xf40(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xea0(%rsp), %ymm3
vmovaps %ymm4, 0xf20(%rsp)
vmovaps %ymm3, 0xf00(%rsp)
vmovaps 0xf20(%rsp), %ymm3
vmovaps 0xf00(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xe00(%rsp)
vmovaps %ymm3, 0xde0(%rsp)
vmovaps 0x7d64c0(%rip), %ymm3 # 0x1e0cd40
vmovaps %ymm3, 0xdc0(%rsp)
vmovaps 0xe00(%rsp), %ymm4
vmovaps 0xde0(%rsp), %ymm3
vmovaps %ymm4, 0xe80(%rsp)
vmovaps %ymm3, 0xe60(%rsp)
vmovaps 0xe80(%rsp), %ymm3
vmovaps 0xe60(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xdc0(%rsp), %ymm3
vmovaps %ymm4, 0xe40(%rsp)
vmovaps %ymm3, 0xe20(%rsp)
vmovaps 0xe40(%rsp), %ymm3
vmovaps 0xe20(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xd20(%rsp)
vmovaps %ymm3, 0xd00(%rsp)
vmovaps 0x7d6437(%rip), %ymm3 # 0x1e0cd60
vmovaps %ymm3, 0xce0(%rsp)
vmovaps 0xd20(%rsp), %ymm4
vmovaps 0xd00(%rsp), %ymm3
vmovaps %ymm4, 0xda0(%rsp)
vmovaps %ymm3, 0xd80(%rsp)
vmovaps 0xda0(%rsp), %ymm3
vmovaps 0xd80(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xce0(%rsp), %ymm3
vmovaps %ymm4, 0xd60(%rsp)
vmovaps %ymm3, 0xd40(%rsp)
vmovaps 0xd60(%rsp), %ymm3
vmovaps 0xd40(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xc40(%rsp)
vmovaps %ymm3, 0xc20(%rsp)
vmovaps 0x7d63ae(%rip), %ymm3 # 0x1e0cd80
vmovaps %ymm3, 0xc00(%rsp)
vmovaps 0xc40(%rsp), %ymm4
vmovaps 0xc20(%rsp), %ymm3
vmovaps %ymm4, 0xcc0(%rsp)
vmovaps %ymm3, 0xca0(%rsp)
vmovaps 0xcc0(%rsp), %ymm3
vmovaps 0xca0(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xc00(%rsp), %ymm3
vmovaps %ymm4, 0xc80(%rsp)
vmovaps %ymm3, 0xc60(%rsp)
vmovaps 0xc80(%rsp), %ymm3
vmovaps 0xc60(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xb60(%rsp)
vmovaps %ymm3, 0xb40(%rsp)
vmovaps 0x7d6325(%rip), %ymm3 # 0x1e0cda0
vmovaps %ymm3, 0xb20(%rsp)
vmovaps 0xb60(%rsp), %ymm4
vmovaps 0xb40(%rsp), %ymm3
vmovaps %ymm4, 0xbe0(%rsp)
vmovaps %ymm3, 0xbc0(%rsp)
vmovaps 0xbe0(%rsp), %ymm3
vmovaps 0xbc0(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xb20(%rsp), %ymm3
vmovaps %ymm4, 0xba0(%rsp)
vmovaps %ymm3, 0xb80(%rsp)
vmovaps 0xba0(%rsp), %ymm3
vmovaps 0xb80(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0xa80(%rsp)
vmovaps %ymm3, 0xa60(%rsp)
vmovaps 0x7d629c(%rip), %ymm3 # 0x1e0cdc0
vmovaps %ymm3, 0xa40(%rsp)
vmovaps 0xa80(%rsp), %ymm4
vmovaps 0xa60(%rsp), %ymm3
vmovaps %ymm4, 0xb00(%rsp)
vmovaps %ymm3, 0xae0(%rsp)
vmovaps 0xb00(%rsp), %ymm3
vmovaps 0xae0(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0xa40(%rsp), %ymm3
vmovaps %ymm4, 0xac0(%rsp)
vmovaps %ymm3, 0xaa0(%rsp)
vmovaps 0xac0(%rsp), %ymm3
vmovaps 0xaa0(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0x9a0(%rsp)
vmovaps %ymm3, 0x980(%rsp)
vmovaps 0x7d6213(%rip), %ymm3 # 0x1e0cde0
vmovaps %ymm3, 0x960(%rsp)
vmovaps 0x9a0(%rsp), %ymm4
vmovaps 0x980(%rsp), %ymm3
vmovaps %ymm4, 0xa20(%rsp)
vmovaps %ymm3, 0xa00(%rsp)
vmovaps 0xa20(%rsp), %ymm3
vmovaps 0xa00(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0x960(%rsp), %ymm3
vmovaps %ymm4, 0x9e0(%rsp)
vmovaps %ymm3, 0x9c0(%rsp)
vmovaps 0x9e0(%rsp), %ymm3
vmovaps 0x9c0(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0x8c0(%rsp)
vmovaps %ymm3, 0x8a0(%rsp)
vmovaps 0x7d618a(%rip), %ymm3 # 0x1e0ce00
vmovaps %ymm3, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %ymm4
vmovaps 0x8a0(%rsp), %ymm3
vmovaps %ymm4, 0x940(%rsp)
vmovaps %ymm3, 0x920(%rsp)
vmovaps 0x940(%rsp), %ymm3
vmovaps 0x920(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm4
vmovaps 0x880(%rsp), %ymm3
vmovaps %ymm4, 0x900(%rsp)
vmovaps %ymm3, 0x8e0(%rsp)
vmovaps 0x900(%rsp), %ymm3
vmovaps 0x8e0(%rsp), %ymm4
vaddps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm3
vmovaps %ymm4, 0x5e0(%rsp)
vmovaps %ymm3, 0x5c0(%rsp)
vmovaps 0x5e0(%rsp), %ymm3
vmovaps 0x5c0(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm4
vmovaps 0x340(%rsp), %ymm3
vmovaps %ymm4, 0x5a0(%rsp)
vmovaps %ymm3, 0x580(%rsp)
vmovaps 0x5a0(%rsp), %ymm3
vmovaps 0x580(%rsp), %ymm4
vmulps %ymm4, %ymm3, %ymm3
vmovaps %ymm3, 0x320(%rsp)
vmovaps 0x3a0(%rsp), %ymm3
vmovaps 0x320(%rsp), %ymm4
vmovaps %ymm3, 0x7e0(%rsp)
vmovaps 0x7d5e24(%rip), %ymm3 # 0x1e0cbc0
vmovaps %ymm3, 0x7c0(%rsp)
vmovaps %ymm4, 0x7a0(%rsp)
vmovaps 0x7e0(%rsp), %ymm6
vmovaps 0x7c0(%rsp), %ymm4
vmovaps %ymm6, 0x860(%rsp)
vmovaps %ymm4, 0x840(%rsp)
vmovaps 0x860(%rsp), %ymm4
vmovaps 0x840(%rsp), %ymm6
vmulps %ymm6, %ymm4, %ymm6
vmovaps 0x7a0(%rsp), %ymm4
vmovaps %ymm6, 0x820(%rsp)
vmovaps %ymm4, 0x800(%rsp)
vmovaps 0x820(%rsp), %ymm4
vmovaps 0x800(%rsp), %ymm6
vaddps %ymm6, %ymm4, %ymm4
vmovaps %ymm4, 0x320(%rsp)
vmovaps 0x340(%rsp), %ymm6
vmovaps 0x320(%rsp), %ymm4
vmovaps %ymm6, 0x16e0(%rsp)
vmovaps %ymm1, 0x16c0(%rsp)
vmovaps %ymm4, 0x16a0(%rsp)
vmovaps 0x16a0(%rsp), %ymm6
vmovaps 0x16e0(%rsp), %ymm8
vmovaps 0x16c0(%rsp), %ymm4
vmovaps %ymm8, 0x1760(%rsp)
vmovaps %ymm4, 0x1740(%rsp)
vmovaps 0x1760(%rsp), %ymm4
vmovaps 0x1740(%rsp), %ymm8
vmulps %ymm4, %ymm8, %ymm4
vmovaps %ymm6, 0x1720(%rsp)
vmovaps %ymm4, 0x1700(%rsp)
vmovaps 0x1720(%rsp), %ymm4
vmovaps 0x1700(%rsp), %ymm6
vsubps %ymm6, %ymm4, %ymm4
vmovaps %ymm4, 0x320(%rsp)
vmovaps 0x420(%rsp), %ymm6
vmovaps 0x320(%rsp), %ymm4
vmovaps %ymm6, 0x460(%rsp)
vmovaps %ymm4, 0x440(%rsp)
vmovaps 0x460(%rsp), %ymm4
vmovaps 0x440(%rsp), %ymm6
vaddps %ymm6, %ymm4, %ymm4
vmovaps %ymm4, 0x420(%rsp)
vmovaps 0x3a0(%rsp), %ymm4
vmovaps 0x420(%rsp), %ymm6
vmovaps %ymm4, 0x700(%rsp)
vmovaps 0x7d5c77(%rip), %ymm4 # 0x1e0cba0
vmovaps %ymm4, 0x6e0(%rsp)
vmovaps %ymm6, 0x6c0(%rsp)
vmovaps 0x700(%rsp), %ymm8
vmovaps 0x6e0(%rsp), %ymm6
vmovaps %ymm8, 0x780(%rsp)
vmovaps %ymm6, 0x760(%rsp)
vmovaps 0x780(%rsp), %ymm6
vmovaps 0x760(%rsp), %ymm8
vmulps %ymm6, %ymm8, %ymm8
vmovaps 0x6c0(%rsp), %ymm6
vmovaps %ymm8, 0x740(%rsp)
vmovaps %ymm6, 0x720(%rsp)
vmovaps 0x740(%rsp), %ymm6
vmovaps 0x720(%rsp), %ymm8
vaddps %ymm6, %ymm8, %ymm6
vmovaps %ymm6, 0x420(%rsp)
vmovaps 0x420(%rsp), %ymm8
vmovaps 0x3c0(%rsp), %ymm6
vmovaps %ymm8, 0x1d00(%rsp)
vmovaps %ymm6, 0x1ce0(%rsp)
vmovaps 0x1d00(%rsp), %ymm6
vmovaps 0x1ce0(%rsp), %ymm8
vorps %ymm6, %ymm8, %ymm6
vmovaps %ymm6, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm6
vmovaps %ymm7, 0x80(%rsp)
vmovaps %ymm6, 0x60(%rsp)
vmovaps 0x80(%rsp), %ymm6
vmovaps 0x60(%rsp), %ymm7
vmulps %ymm7, %ymm6, %ymm6
vmovaps %ymm6, 0x180(%rsp)
vmovaps %ymm5, 0x6a0(%rsp)
vmovaps 0x6a0(%rsp), %ymm5
vmovaps %ymm5, 0x160(%rsp)
vmovaps %ymm2, 0x100(%rsp)
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm2, 0x300(%rsp)
vmovaps 0x7d5abc(%rip), %ymm2 # 0x1e0cb20
vmovaps %ymm2, 0x2e0(%rsp)
vmovaps 0x300(%rsp), %ymm2
vmovaps 0x2e0(%rsp), %ymm5
vminps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0x180(%rsp)
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm2, 0x2c0(%rsp)
vmovaps 0x7d5a9a(%rip), %ymm2 # 0x1e0cb40
vmovaps %ymm2, 0x2a0(%rsp)
vmovaps 0x2c0(%rsp), %ymm2
vmovaps 0x2a0(%rsp), %ymm5
vmaxps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0x180(%rsp)
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm2, 0x1500(%rsp)
vmovaps 0x7d5a78(%rip), %ymm2 # 0x1e0cb60
vmovaps %ymm2, 0x14e0(%rsp)
vmovaps %ymm1, 0x14c0(%rsp)
vmovaps 0x1500(%rsp), %ymm5
vmovaps 0x14e0(%rsp), %ymm2
vmovaps %ymm5, 0x1580(%rsp)
vmovaps %ymm2, 0x1560(%rsp)
vmovaps 0x1580(%rsp), %ymm2
vmovaps 0x1560(%rsp), %ymm5
vmulps %ymm5, %ymm2, %ymm5
vmovaps 0x14c0(%rsp), %ymm2
vmovaps %ymm5, 0x1540(%rsp)
vmovaps %ymm2, 0x1520(%rsp)
vmovaps 0x1540(%rsp), %ymm2
vmovaps 0x1520(%rsp), %ymm5
vaddps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0x140(%rsp)
vmovaps 0x140(%rsp), %ymm2
vroundps $0x1, %ymm2, %ymm2
vmovaps %ymm2, 0x160(%rsp)
vmovaps 0x160(%rsp), %ymm5
vmovaps 0x140(%rsp), %ymm2
vcmpltps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm5
vmovaps 0x100(%rsp), %ymm2
vmovaps %ymm5, 0x1680(%rsp)
vmovaps %ymm2, 0x1660(%rsp)
vmovaps 0x1680(%rsp), %ymm2
vmovaps 0x1660(%rsp), %ymm5
vandps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0xe0(%rsp)
vmovaps 0x160(%rsp), %ymm5
vmovaps 0xe0(%rsp), %ymm2
vmovaps %ymm5, 0x200(%rsp)
vmovaps %ymm2, 0x1e0(%rsp)
vmovaps 0x200(%rsp), %ymm2
vmovaps 0x1e0(%rsp), %ymm5
vsubps %ymm5, %ymm2, %ymm2
vmovaps %ymm2, 0x140(%rsp)
vmovaps 0x140(%rsp), %ymm5
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm5, 0x18a0(%rsp)
vmovaps %ymm4, 0x1880(%rsp)
vmovaps %ymm2, 0x1860(%rsp)
vmovaps 0x1860(%rsp), %ymm4
vmovaps 0x18a0(%rsp), %ymm5
vmovaps 0x1880(%rsp), %ymm2
vmovaps %ymm5, 0x1920(%rsp)
vmovaps %ymm2, 0x1900(%rsp)
vmovaps 0x1920(%rsp), %ymm2
vmovaps 0x1900(%rsp), %ymm5
vmulps %ymm5, %ymm2, %ymm2
vmovaps %ymm4, 0x18e0(%rsp)
vmovaps %ymm2, 0x18c0(%rsp)
vmovaps 0x18e0(%rsp), %ymm2
vmovaps 0x18c0(%rsp), %ymm4
vsubps %ymm4, %ymm2, %ymm2
vmovaps %ymm2, 0x180(%rsp)
vmovaps 0x140(%rsp), %ymm4
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm4, 0x17c0(%rsp)
vmovaps %ymm3, 0x17a0(%rsp)
vmovaps %ymm2, 0x1780(%rsp)
vmovaps 0x1780(%rsp), %ymm3
vmovaps 0x17c0(%rsp), %ymm4
vmovaps 0x17a0(%rsp), %ymm2
vmovaps %ymm4, 0x1840(%rsp)
vmovaps %ymm2, 0x1820(%rsp)
vmovaps 0x1840(%rsp), %ymm2
vmovaps 0x1820(%rsp), %ymm4
vmulps %ymm4, %ymm2, %ymm2
vmovaps %ymm3, 0x1800(%rsp)
vmovaps %ymm2, 0x17e0(%rsp)
vmovaps 0x1800(%rsp), %ymm2
vmovaps 0x17e0(%rsp), %ymm3
vsubps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0x180(%rsp)
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm2, 0x280(%rsp)
vmovaps %ymm2, 0x260(%rsp)
vmovaps 0x280(%rsp), %ymm2
vmovaps 0x260(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0x160(%rsp)
vmovaps 0x7d5830(%rip), %ymm2 # 0x1e0cbe0
vmovaps %ymm2, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm3, 0x1420(%rsp)
vmovaps %ymm2, 0x1400(%rsp)
vmovaps 0x7d581b(%rip), %ymm2 # 0x1e0cc00
vmovaps %ymm2, 0x13e0(%rsp)
vmovaps 0x1420(%rsp), %ymm3
vmovaps 0x1400(%rsp), %ymm2
vmovaps %ymm3, 0x14a0(%rsp)
vmovaps %ymm2, 0x1480(%rsp)
vmovaps 0x14a0(%rsp), %ymm2
vmovaps 0x1480(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm3
vmovaps 0x13e0(%rsp), %ymm2
vmovaps %ymm3, 0x1460(%rsp)
vmovaps %ymm2, 0x1440(%rsp)
vmovaps 0x1460(%rsp), %ymm2
vmovaps 0x1440(%rsp), %ymm3
vaddps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm3, 0x1340(%rsp)
vmovaps %ymm2, 0x1320(%rsp)
vmovaps 0x7d5792(%rip), %ymm2 # 0x1e0cc20
vmovaps %ymm2, 0x1300(%rsp)
vmovaps 0x1340(%rsp), %ymm3
vmovaps 0x1320(%rsp), %ymm2
vmovaps %ymm3, 0x13c0(%rsp)
vmovaps %ymm2, 0x13a0(%rsp)
vmovaps 0x13c0(%rsp), %ymm2
vmovaps 0x13a0(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm3
vmovaps 0x1300(%rsp), %ymm2
vmovaps %ymm3, 0x1380(%rsp)
vmovaps %ymm2, 0x1360(%rsp)
vmovaps 0x1380(%rsp), %ymm2
vmovaps 0x1360(%rsp), %ymm3
vaddps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm3, 0x1260(%rsp)
vmovaps %ymm2, 0x1240(%rsp)
vmovaps 0x7d5709(%rip), %ymm2 # 0x1e0cc40
vmovaps %ymm2, 0x1220(%rsp)
vmovaps 0x1260(%rsp), %ymm3
vmovaps 0x1240(%rsp), %ymm2
vmovaps %ymm3, 0x12e0(%rsp)
vmovaps %ymm2, 0x12c0(%rsp)
vmovaps 0x12e0(%rsp), %ymm2
vmovaps 0x12c0(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm3
vmovaps 0x1220(%rsp), %ymm2
vmovaps %ymm3, 0x12a0(%rsp)
vmovaps %ymm2, 0x1280(%rsp)
vmovaps 0x12a0(%rsp), %ymm2
vmovaps 0x1280(%rsp), %ymm3
vaddps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm3, 0x1180(%rsp)
vmovaps %ymm2, 0x1160(%rsp)
vmovaps 0x7d5680(%rip), %ymm2 # 0x1e0cc60
vmovaps %ymm2, 0x1140(%rsp)
vmovaps 0x1180(%rsp), %ymm3
vmovaps 0x1160(%rsp), %ymm2
vmovaps %ymm3, 0x1200(%rsp)
vmovaps %ymm2, 0x11e0(%rsp)
vmovaps 0x1200(%rsp), %ymm2
vmovaps 0x11e0(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm3
vmovaps 0x1140(%rsp), %ymm2
vmovaps %ymm3, 0x11c0(%rsp)
vmovaps %ymm2, 0x11a0(%rsp)
vmovaps 0x11c0(%rsp), %ymm2
vmovaps 0x11a0(%rsp), %ymm3
vaddps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x180(%rsp), %ymm2
vmovaps %ymm3, 0x10a0(%rsp)
vmovaps %ymm2, 0x1080(%rsp)
vmovaps %ymm1, 0x1060(%rsp)
vmovaps 0x10a0(%rsp), %ymm2
vmovaps 0x1080(%rsp), %ymm1
vmovaps %ymm2, 0x1120(%rsp)
vmovaps %ymm1, 0x1100(%rsp)
vmovaps 0x1120(%rsp), %ymm1
vmovaps 0x1100(%rsp), %ymm2
vmulps %ymm2, %ymm1, %ymm2
vmovaps 0x1060(%rsp), %ymm1
vmovaps %ymm2, 0x10e0(%rsp)
vmovaps %ymm1, 0x10c0(%rsp)
vmovaps 0x10e0(%rsp), %ymm1
vmovaps 0x10c0(%rsp), %ymm2
vaddps %ymm2, %ymm1, %ymm1
vmovaps %ymm1, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm3
vmovaps 0x160(%rsp), %ymm2
vmovaps 0x180(%rsp), %ymm1
vmovaps %ymm3, 0xfc0(%rsp)
vmovaps %ymm2, 0xfa0(%rsp)
vmovaps %ymm1, 0xf80(%rsp)
vmovaps 0xfc0(%rsp), %ymm2
vmovaps 0xfa0(%rsp), %ymm1
vmovaps %ymm2, 0x1040(%rsp)
vmovaps %ymm1, 0x1020(%rsp)
vmovaps 0x1040(%rsp), %ymm1
vmovaps 0x1020(%rsp), %ymm2
vmulps %ymm2, %ymm1, %ymm2
vmovaps 0xf80(%rsp), %ymm1
vmovaps %ymm2, 0x1000(%rsp)
vmovaps %ymm1, 0xfe0(%rsp)
vmovaps 0x1000(%rsp), %ymm1
vmovaps 0xfe0(%rsp), %ymm2
vaddps %ymm2, %ymm1, %ymm1
vmovaps %ymm1, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm2
vmovaps 0x100(%rsp), %ymm1
vmovaps %ymm2, 0x1c0(%rsp)
vmovaps %ymm1, 0x1a0(%rsp)
vmovaps 0x1c0(%rsp), %ymm1
vmovaps 0x1a0(%rsp), %ymm2
vaddps %ymm2, %ymm1, %ymm1
vmovaps %ymm1, 0xc0(%rsp)
vmovaps 0x140(%rsp), %ymm1
vmovaps %ymm1, 0x1940(%rsp)
vcvttps2dq 0x1940(%rsp), %ymm1
vmovaps %ymm1, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm1
vmovaps %ymm1, 0x1a40(%rsp)
vmovaps %ymm0, 0x1a20(%rsp)
vmovdqa 0x1a40(%rsp), %ymm0
vmovdqa %ymm0, 0x19a0(%rsp)
vmovdqa 0x19a0(%rsp), %xmm0
vmovdqa %xmm0, 0x1a10(%rsp)
vmovdqa 0x19b0(%rsp), %xmm0
vmovdqa %xmm0, 0x1a00(%rsp)
vmovdqa 0x1a20(%rsp), %ymm0
vmovdqa %ymm0, 0x1980(%rsp)
vmovdqa 0x1980(%rsp), %xmm0
vmovdqa %xmm0, 0x19f0(%rsp)
vmovdqa 0x1990(%rsp), %xmm0
vmovdqa %xmm0, 0x19e0(%rsp)
vmovdqa 0x1a10(%rsp), %xmm1
vmovdqa 0x19f0(%rsp), %xmm0
vmovdqa %xmm1, 0x1a90(%rsp)
vmovdqa %xmm0, 0x1a80(%rsp)
vmovdqa 0x1a90(%rsp), %xmm0
vmovdqa 0x1a80(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1a10(%rsp)
vmovdqa 0x1a00(%rsp), %xmm1
vmovdqa 0x19e0(%rsp), %xmm0
vmovdqa %xmm1, 0x1a70(%rsp)
vmovdqa %xmm0, 0x1a60(%rsp)
vmovdqa 0x1a70(%rsp), %xmm0
vmovdqa 0x1a60(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1a00(%rsp)
vmovdqa 0x1a10(%rsp), %xmm0
vmovdqa %xmm0, 0x1960(%rsp)
vmovdqa 0x1a00(%rsp), %xmm0
vmovdqa %xmm0, 0x1970(%rsp)
vmovdqa 0x1960(%rsp), %ymm0
vmovdqa %ymm0, 0x19c0(%rsp)
vmovdqa 0x19c0(%rsp), %ymm0
vmovdqa %ymm0, 0x120(%rsp)
vmovdqa 0x120(%rsp), %ymm0
vmovdqa %ymm0, 0x1b40(%rsp)
movl $0x17, 0x1b3c(%rsp)
vmovdqa 0x1b40(%rsp), %ymm0
vmovdqa %ymm0, 0x1ac0(%rsp)
vmovdqa 0x1ac0(%rsp), %xmm0
vmovdqa %xmm0, 0x1b20(%rsp)
vmovdqa 0x1ad0(%rsp), %xmm0
vmovdqa %xmm0, 0x1b10(%rsp)
vmovdqa 0x1b20(%rsp), %xmm0
movl 0x1b3c(%rsp), %eax
vmovdqa %xmm0, 0x1b90(%rsp)
movl %eax, 0x1b8c(%rsp)
vmovdqa 0x1b90(%rsp), %xmm0
movl 0x1b8c(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1b20(%rsp)
vmovdqa 0x1b10(%rsp), %xmm0
movl 0x1b3c(%rsp), %eax
vmovdqa %xmm0, 0x1b70(%rsp)
movl %eax, 0x1b6c(%rsp)
vmovdqa 0x1b70(%rsp), %xmm0
movl 0x1b6c(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1b10(%rsp)
vmovdqa 0x1b20(%rsp), %xmm0
vmovdqa %xmm0, 0x1aa0(%rsp)
vmovdqa 0x1b10(%rsp), %xmm0
vmovdqa %xmm0, 0x1ab0(%rsp)
vmovdqa 0x1aa0(%rsp), %ymm0
vmovdqa %ymm0, 0x1ae0(%rsp)
vmovdqa 0x1ae0(%rsp), %ymm0
vmovdqa %ymm0, 0x120(%rsp)
vmovdqa 0x120(%rsp), %ymm0
vmovdqa %ymm0, 0x1ba0(%rsp)
vmovdqa 0x1ba0(%rsp), %ymm0
vmovaps %ymm0, 0xa0(%rsp)
vmovaps 0xc0(%rsp), %ymm1
vmovaps 0xa0(%rsp), %ymm0
vmovaps %ymm1, 0x240(%rsp)
vmovaps %ymm0, 0x220(%rsp)
vmovaps 0x240(%rsp), %ymm0
vmulps 0x220(%rsp), %ymm0, %ymm0
vmovaps %ymm0, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,126 |
ncnn::BinaryOp_x86_avx_functor::binary_op_pow::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return pow_ps(x, y);
}
|
subq $0xd88, %rsp # imm = 0xD88
movq %rdi, -0x68(%rsp)
movq %rsi, -0x70(%rsp)
movq %rdx, -0x78(%rsp)
movq -0x70(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x78(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x50(%rsp)
vmovaps %xmm0, -0x60(%rsp)
vmovaps -0x60(%rsp), %xmm7
vmovaps -0x50(%rsp), %xmm0
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x7d3eef(%rip), %xmm2 # 0x1e0ba40
vmovaps %xmm2, 0x1c0(%rsp)
vmovaps 0x1e0(%rsp), %xmm1
vxorps %xmm5, %xmm5, %xmm5
vmovaps %xmm5, 0x310(%rsp)
vmovaps 0x310(%rsp), %xmm0
vmovaps %xmm1, 0xcc0(%rsp)
vmovaps %xmm0, 0xcb0(%rsp)
vmovaps 0xcc0(%rsp), %xmm0
vmovaps 0xcb0(%rsp), %xmm1
vcmpleps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x1b0(%rsp)
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm0, 0x300(%rsp)
vmovaps 0x7d3f4b(%rip), %xmm0 # 0x1e0bb10
vmovaps %xmm0, 0x2f0(%rsp)
vmovaps 0x300(%rsp), %xmm0
vmovaps 0x2f0(%rsp), %xmm1
vmaxps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm0, 0xcf0(%rsp)
vmovaps 0xcf0(%rsp), %xmm0
vmovaps %xmm0, 0xce0(%rsp)
movl $0x17, 0xcdc(%rsp)
vmovdqa 0xce0(%rsp), %xmm0
vmovd 0xcdc(%rsp), %xmm1
vpsrld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1d0(%rsp)
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm0, 0x3e0(%rsp)
vmovaps 0x7d3ecb(%rip), %xmm0 # 0x1e0bb20
vmovaps %xmm0, 0x3d0(%rsp)
vmovdqa 0x3e0(%rsp), %xmm0
vmovdqa 0x3d0(%rsp), %xmm1
vpand %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm0
vmovaps %xmm0, 0xd30(%rsp)
vmovaps 0x7d3de9(%rip), %xmm1 # 0x1e0ba80
vmovaps %xmm1, 0xd20(%rsp)
vmovdqa 0xd30(%rsp), %xmm0
vmovdqa 0xd20(%rsp), %xmm3
vpor %xmm3, %xmm0, %xmm0
vmovdqa %xmm0, 0x1e0(%rsp)
vmovdqa 0x1d0(%rsp), %xmm0
vmovdqa %xmm0, 0xd50(%rsp)
vmovdqa 0x7d3e27(%rip), %xmm0 # 0x1e0bb00
vmovdqa %xmm0, 0xd40(%rsp)
vmovdqa 0xd50(%rsp), %xmm3
vmovdqa 0xd40(%rsp), %xmm4
vpsubd %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0x1d0(%rsp)
vmovdqa 0x1d0(%rsp), %xmm3
vmovdqa %xmm3, 0x350(%rsp)
vcvtdq2ps 0x350(%rsp), %xmm3
vmovaps %xmm3, 0x1a0(%rsp)
vmovaps 0x1a0(%rsp), %xmm4
vmovaps 0x1c0(%rsp), %xmm3
vmovaps %xmm4, 0x240(%rsp)
vmovaps %xmm3, 0x230(%rsp)
vmovaps 0x240(%rsp), %xmm3
vmovaps 0x230(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x1a0(%rsp)
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm3, 0xd70(%rsp)
vmovaps 0x7d3dae(%rip), %xmm3 # 0x1e0bb30
vmovaps %xmm3, 0xd60(%rsp)
vmovaps 0xd70(%rsp), %xmm3
vmovaps 0xd60(%rsp), %xmm4
vcmpltps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x190(%rsp)
vmovaps 0x1e0(%rsp), %xmm4
vmovaps 0x190(%rsp), %xmm3
vmovaps %xmm4, 0x3c0(%rsp)
vmovaps %xmm3, 0x3b0(%rsp)
vmovdqa 0x3c0(%rsp), %xmm3
vmovdqa 0x3b0(%rsp), %xmm4
vpand %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0x180(%rsp)
vmovaps 0x1e0(%rsp), %xmm4
vmovaps 0x1c0(%rsp), %xmm3
vmovaps %xmm4, 0x280(%rsp)
vmovaps %xmm3, 0x270(%rsp)
vmovaps 0x280(%rsp), %xmm3
vmovaps 0x270(%rsp), %xmm4
vsubps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x1e0(%rsp)
vmovaps 0x1a0(%rsp), %xmm4
vmovaps 0x1c0(%rsp), %xmm6
vmovaps 0x190(%rsp), %xmm3
vmovaps %xmm6, 0x3a0(%rsp)
vmovaps %xmm3, 0x390(%rsp)
vmovdqa 0x3a0(%rsp), %xmm3
vmovdqa 0x390(%rsp), %xmm6
vpand %xmm6, %xmm3, %xmm3
vmovaps %xmm4, 0x260(%rsp)
vmovdqa %xmm3, 0x250(%rsp)
vmovaps 0x260(%rsp), %xmm3
vmovaps 0x250(%rsp), %xmm4
vsubps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x1a0(%rsp)
vmovaps 0x1e0(%rsp), %xmm4
vmovaps 0x180(%rsp), %xmm3
vmovaps %xmm4, 0x220(%rsp)
vmovaps %xmm3, 0x210(%rsp)
vmovaps 0x220(%rsp), %xmm3
vmovaps 0x210(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm3, 0x2e0(%rsp)
vmovaps %xmm3, 0x2d0(%rsp)
vmovaps 0x2e0(%rsp), %xmm3
vmovaps 0x2d0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x170(%rsp)
vmovaps 0x7d3c16(%rip), %xmm3 # 0x1e0bb40
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x970(%rsp)
vmovaps %xmm3, 0x960(%rsp)
vmovaps 0x7d3bf1(%rip), %xmm3 # 0x1e0bb50
vmovaps %xmm3, 0x950(%rsp)
vmovaps 0x970(%rsp), %xmm4
vmovaps 0x960(%rsp), %xmm3
vmovaps %xmm4, 0x9b0(%rsp)
vmovaps %xmm3, 0x9a0(%rsp)
vmovaps 0x9b0(%rsp), %xmm3
vmovaps 0x9a0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x950(%rsp), %xmm3
vmovaps %xmm4, 0x990(%rsp)
vmovaps %xmm3, 0x980(%rsp)
vmovaps 0x990(%rsp), %xmm3
vmovaps 0x980(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x900(%rsp)
vmovaps %xmm3, 0x8f0(%rsp)
vmovaps 0x7d3b58(%rip), %xmm3 # 0x1e0bb60
vmovaps %xmm3, 0x8e0(%rsp)
vmovaps 0x900(%rsp), %xmm4
vmovaps 0x8f0(%rsp), %xmm3
vmovaps %xmm4, 0x940(%rsp)
vmovaps %xmm3, 0x930(%rsp)
vmovaps 0x940(%rsp), %xmm3
vmovaps 0x930(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x8e0(%rsp), %xmm3
vmovaps %xmm4, 0x920(%rsp)
vmovaps %xmm3, 0x910(%rsp)
vmovaps 0x920(%rsp), %xmm3
vmovaps 0x910(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x890(%rsp)
vmovaps %xmm3, 0x880(%rsp)
vmovaps 0x7d3abf(%rip), %xmm3 # 0x1e0bb70
vmovaps %xmm3, 0x870(%rsp)
vmovaps 0x890(%rsp), %xmm4
vmovaps 0x880(%rsp), %xmm3
vmovaps %xmm4, 0x8d0(%rsp)
vmovaps %xmm3, 0x8c0(%rsp)
vmovaps 0x8d0(%rsp), %xmm3
vmovaps 0x8c0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x870(%rsp), %xmm3
vmovaps %xmm4, 0x8b0(%rsp)
vmovaps %xmm3, 0x8a0(%rsp)
vmovaps 0x8b0(%rsp), %xmm3
vmovaps 0x8a0(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x820(%rsp)
vmovaps %xmm3, 0x810(%rsp)
vmovaps 0x7d3a26(%rip), %xmm3 # 0x1e0bb80
vmovaps %xmm3, 0x800(%rsp)
vmovaps 0x820(%rsp), %xmm4
vmovaps 0x810(%rsp), %xmm3
vmovaps %xmm4, 0x860(%rsp)
vmovaps %xmm3, 0x850(%rsp)
vmovaps 0x860(%rsp), %xmm3
vmovaps 0x850(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x800(%rsp), %xmm3
vmovaps %xmm4, 0x840(%rsp)
vmovaps %xmm3, 0x830(%rsp)
vmovaps 0x840(%rsp), %xmm3
vmovaps 0x830(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x7b0(%rsp)
vmovaps %xmm3, 0x7a0(%rsp)
vmovaps 0x7d398d(%rip), %xmm3 # 0x1e0bb90
vmovaps %xmm3, 0x790(%rsp)
vmovaps 0x7b0(%rsp), %xmm4
vmovaps 0x7a0(%rsp), %xmm3
vmovaps %xmm4, 0x7f0(%rsp)
vmovaps %xmm3, 0x7e0(%rsp)
vmovaps 0x7f0(%rsp), %xmm3
vmovaps 0x7e0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x790(%rsp), %xmm3
vmovaps %xmm4, 0x7d0(%rsp)
vmovaps %xmm3, 0x7c0(%rsp)
vmovaps 0x7d0(%rsp), %xmm3
vmovaps 0x7c0(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x740(%rsp)
vmovaps %xmm3, 0x730(%rsp)
vmovaps 0x7d38f4(%rip), %xmm3 # 0x1e0bba0
vmovaps %xmm3, 0x720(%rsp)
vmovaps 0x740(%rsp), %xmm4
vmovaps 0x730(%rsp), %xmm3
vmovaps %xmm4, 0x780(%rsp)
vmovaps %xmm3, 0x770(%rsp)
vmovaps 0x780(%rsp), %xmm3
vmovaps 0x770(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x720(%rsp), %xmm3
vmovaps %xmm4, 0x760(%rsp)
vmovaps %xmm3, 0x750(%rsp)
vmovaps 0x760(%rsp), %xmm3
vmovaps 0x750(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x6d0(%rsp)
vmovaps %xmm3, 0x6c0(%rsp)
vmovaps 0x7d385b(%rip), %xmm3 # 0x1e0bbb0
vmovaps %xmm3, 0x6b0(%rsp)
vmovaps 0x6d0(%rsp), %xmm4
vmovaps 0x6c0(%rsp), %xmm3
vmovaps %xmm4, 0x710(%rsp)
vmovaps %xmm3, 0x700(%rsp)
vmovaps 0x710(%rsp), %xmm3
vmovaps 0x700(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x6b0(%rsp), %xmm3
vmovaps %xmm4, 0x6f0(%rsp)
vmovaps %xmm3, 0x6e0(%rsp)
vmovaps 0x6f0(%rsp), %xmm3
vmovaps 0x6e0(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x660(%rsp)
vmovaps %xmm3, 0x650(%rsp)
vmovaps 0x7d37c2(%rip), %xmm3 # 0x1e0bbc0
vmovaps %xmm3, 0x640(%rsp)
vmovaps 0x660(%rsp), %xmm4
vmovaps 0x650(%rsp), %xmm3
vmovaps %xmm4, 0x6a0(%rsp)
vmovaps %xmm3, 0x690(%rsp)
vmovaps 0x6a0(%rsp), %xmm3
vmovaps 0x690(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm4
vmovaps 0x640(%rsp), %xmm3
vmovaps %xmm4, 0x680(%rsp)
vmovaps %xmm3, 0x670(%rsp)
vmovaps 0x680(%rsp), %xmm3
vmovaps 0x670(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm3
vmovaps %xmm4, 0x2c0(%rsp)
vmovaps %xmm3, 0x2b0(%rsp)
vmovaps 0x2c0(%rsp), %xmm3
vmovaps 0x2b0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm4
vmovaps 0x170(%rsp), %xmm3
vmovaps %xmm4, 0x2a0(%rsp)
vmovaps %xmm3, 0x290(%rsp)
vmovaps 0x2a0(%rsp), %xmm3
vmovaps 0x290(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x160(%rsp)
vmovaps 0x1a0(%rsp), %xmm3
vmovaps 0x160(%rsp), %xmm4
vmovaps %xmm3, 0x5f0(%rsp)
vmovaps 0x7d357c(%rip), %xmm3 # 0x1e0baa0
vmovaps %xmm3, 0x5e0(%rsp)
vmovaps %xmm4, 0x5d0(%rsp)
vmovaps 0x5f0(%rsp), %xmm6
vmovaps 0x5e0(%rsp), %xmm4
vmovaps %xmm6, 0x630(%rsp)
vmovaps %xmm4, 0x620(%rsp)
vmovaps 0x630(%rsp), %xmm4
vmovaps 0x620(%rsp), %xmm6
vmulps %xmm6, %xmm4, %xmm6
vmovaps 0x5d0(%rsp), %xmm4
vmovaps %xmm6, 0x610(%rsp)
vmovaps %xmm4, 0x600(%rsp)
vmovaps 0x610(%rsp), %xmm4
vmovaps 0x600(%rsp), %xmm6
vaddps %xmm6, %xmm4, %xmm4
vmovaps %xmm4, 0x160(%rsp)
vmovaps 0x170(%rsp), %xmm6
vmovaps 0x160(%rsp), %xmm4
vmovaps %xmm6, 0x430(%rsp)
vmovaps %xmm1, 0x420(%rsp)
vmovaps %xmm4, 0x410(%rsp)
vmovaps 0x410(%rsp), %xmm6
vmovaps 0x430(%rsp), %xmm8
vmovaps 0x420(%rsp), %xmm4
vmovaps %xmm8, 0x470(%rsp)
vmovaps %xmm4, 0x460(%rsp)
vmovaps 0x470(%rsp), %xmm4
vmovaps 0x460(%rsp), %xmm8
vmulps %xmm4, %xmm8, %xmm4
vmovaps %xmm6, 0x450(%rsp)
vmovaps %xmm4, 0x440(%rsp)
vmovaps 0x450(%rsp), %xmm4
vmovaps 0x440(%rsp), %xmm6
vsubps %xmm6, %xmm4, %xmm4
vmovaps %xmm4, 0x160(%rsp)
vmovaps 0x1e0(%rsp), %xmm6
vmovaps 0x160(%rsp), %xmm4
vmovaps %xmm6, 0x200(%rsp)
vmovaps %xmm4, 0x1f0(%rsp)
vmovaps 0x200(%rsp), %xmm4
vmovaps 0x1f0(%rsp), %xmm6
vaddps %xmm6, %xmm4, %xmm4
vmovaps %xmm4, 0x1e0(%rsp)
vmovaps 0x1a0(%rsp), %xmm4
vmovaps 0x1e0(%rsp), %xmm6
vmovaps %xmm4, 0x580(%rsp)
vmovaps 0x7d33df(%rip), %xmm4 # 0x1e0ba90
vmovaps %xmm4, 0x570(%rsp)
vmovaps %xmm6, 0x560(%rsp)
vmovaps 0x580(%rsp), %xmm8
vmovaps 0x570(%rsp), %xmm6
vmovaps %xmm8, 0x5c0(%rsp)
vmovaps %xmm6, 0x5b0(%rsp)
vmovaps 0x5c0(%rsp), %xmm6
vmovaps 0x5b0(%rsp), %xmm8
vmulps %xmm6, %xmm8, %xmm8
vmovaps 0x560(%rsp), %xmm6
vmovaps %xmm8, 0x5a0(%rsp)
vmovaps %xmm6, 0x590(%rsp)
vmovaps 0x5a0(%rsp), %xmm6
vmovaps 0x590(%rsp), %xmm8
vaddps %xmm6, %xmm8, %xmm6
vmovaps %xmm6, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm8
vmovaps 0x1b0(%rsp), %xmm6
vmovaps %xmm8, 0xd10(%rsp)
vmovaps %xmm6, 0xd00(%rsp)
vmovdqa 0xd10(%rsp), %xmm6
vmovdqa 0xd00(%rsp), %xmm8
vpor %xmm6, %xmm8, %xmm6
vmovdqa %xmm6, 0x1e0(%rsp)
vmovaps 0x1e0(%rsp), %xmm6
vmovaps %xmm7, -0x30(%rsp)
vmovaps %xmm6, -0x40(%rsp)
vmovaps -0x30(%rsp), %xmm6
vmovaps -0x40(%rsp), %xmm7
vmulps %xmm7, %xmm6, %xmm6
vmovaps %xmm6, 0x50(%rsp)
vmovaps %xmm5, 0x320(%rsp)
vmovaps 0x320(%rsp), %xmm5
vmovaps %xmm5, 0x40(%rsp)
vmovaps %xmm2, 0x10(%rsp)
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm2, 0x150(%rsp)
vmovaps 0x7d3276(%rip), %xmm2 # 0x1e0ba50
vmovaps %xmm2, 0x140(%rsp)
vmovaps 0x150(%rsp), %xmm2
vmovaps 0x140(%rsp), %xmm5
vminps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, 0x50(%rsp)
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm2, 0x130(%rsp)
vmovaps 0x7d324a(%rip), %xmm2 # 0x1e0ba60
vmovaps %xmm2, 0x120(%rsp)
vmovaps 0x130(%rsp), %xmm2
vmovaps 0x120(%rsp), %xmm5
vmaxps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, 0x50(%rsp)
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm2, 0x110(%rsp)
vmovaps 0x7d321e(%rip), %xmm2 # 0x1e0ba70
vmovaps %xmm2, 0x100(%rsp)
vmovaps 0x110(%rsp), %xmm2
vmovaps 0x100(%rsp), %xmm5
vmulps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm2
vmovaps %xmm2, 0x90(%rsp)
vmovaps %xmm1, 0x80(%rsp)
vmovaps 0x90(%rsp), %xmm2
vmovaps 0x80(%rsp), %xmm5
vaddps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm2
vmovaps %xmm2, 0x340(%rsp)
vcvttps2dq 0x340(%rsp), %xmm2
vmovdqa %xmm2, 0x20(%rsp)
vmovdqa 0x20(%rsp), %xmm2
vmovdqa %xmm2, 0x360(%rsp)
vcvtdq2ps 0x360(%rsp), %xmm2
vmovaps %xmm2, 0x40(%rsp)
vmovaps 0x40(%rsp), %xmm5
vmovaps 0x30(%rsp), %xmm2
vmovaps %xmm5, 0x380(%rsp)
vmovaps %xmm2, 0x370(%rsp)
vmovaps 0x370(%rsp), %xmm2
vmovaps 0x380(%rsp), %xmm5
vcmpltps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, (%rsp)
vmovaps (%rsp), %xmm5
vmovaps 0x10(%rsp), %xmm2
vmovaps %xmm5, 0x400(%rsp)
vmovaps %xmm2, 0x3f0(%rsp)
vmovdqa 0x400(%rsp), %xmm2
vmovdqa 0x3f0(%rsp), %xmm5
vpand %xmm5, %xmm2, %xmm2
vmovdqa %xmm2, (%rsp)
vmovaps 0x40(%rsp), %xmm5
vmovaps (%rsp), %xmm2
vmovaps %xmm5, 0xb0(%rsp)
vmovaps %xmm2, 0xa0(%rsp)
vmovaps 0xb0(%rsp), %xmm2
vmovaps 0xa0(%rsp), %xmm5
vsubps %xmm5, %xmm2, %xmm2
vmovaps %xmm2, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm5
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm5, 0x510(%rsp)
vmovaps %xmm4, 0x500(%rsp)
vmovaps %xmm2, 0x4f0(%rsp)
vmovaps 0x4f0(%rsp), %xmm4
vmovaps 0x510(%rsp), %xmm5
vmovaps 0x500(%rsp), %xmm2
vmovaps %xmm5, 0x550(%rsp)
vmovaps %xmm2, 0x540(%rsp)
vmovaps 0x550(%rsp), %xmm2
vmovaps 0x540(%rsp), %xmm5
vmulps %xmm5, %xmm2, %xmm2
vmovaps %xmm4, 0x530(%rsp)
vmovaps %xmm2, 0x520(%rsp)
vmovaps 0x530(%rsp), %xmm2
vmovaps 0x520(%rsp), %xmm4
vsubps %xmm4, %xmm2, %xmm2
vmovaps %xmm2, 0x50(%rsp)
vmovaps 0x30(%rsp), %xmm4
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm4, 0x4a0(%rsp)
vmovaps %xmm3, 0x490(%rsp)
vmovaps %xmm2, 0x480(%rsp)
vmovaps 0x480(%rsp), %xmm3
vmovaps 0x4a0(%rsp), %xmm4
vmovaps 0x490(%rsp), %xmm2
vmovaps %xmm4, 0x4e0(%rsp)
vmovaps %xmm2, 0x4d0(%rsp)
vmovaps 0x4e0(%rsp), %xmm2
vmovaps 0x4d0(%rsp), %xmm4
vmulps %xmm4, %xmm2, %xmm2
vmovaps %xmm3, 0x4c0(%rsp)
vmovaps %xmm2, 0x4b0(%rsp)
vmovaps 0x4c0(%rsp), %xmm2
vmovaps 0x4b0(%rsp), %xmm3
vsubps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, 0x50(%rsp)
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm2, 0xf0(%rsp)
vmovaps %xmm2, 0xe0(%rsp)
vmovaps 0xf0(%rsp), %xmm2
vmovaps 0xe0(%rsp), %xmm3
vmulps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, 0x40(%rsp)
vmovaps 0x7d2fb2(%rip), %xmm2 # 0x1e0bab0
vmovaps %xmm2, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm3, 0xc10(%rsp)
vmovaps %xmm2, 0xc00(%rsp)
vmovaps 0x7d2f96(%rip), %xmm2 # 0x1e0bac0
vmovaps %xmm2, 0xbf0(%rsp)
vmovaps 0xc10(%rsp), %xmm3
vmovaps 0xc00(%rsp), %xmm2
vmovaps %xmm3, 0xc50(%rsp)
vmovaps %xmm2, 0xc40(%rsp)
vmovaps 0xc50(%rsp), %xmm2
vmovaps 0xc40(%rsp), %xmm3
vmulps %xmm3, %xmm2, %xmm3
vmovaps 0xbf0(%rsp), %xmm2
vmovaps %xmm3, 0xc30(%rsp)
vmovaps %xmm2, 0xc20(%rsp)
vmovaps 0xc30(%rsp), %xmm2
vmovaps 0xc20(%rsp), %xmm3
vaddps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm3, 0xba0(%rsp)
vmovaps %xmm2, 0xb90(%rsp)
vmovaps 0x7d2f06(%rip), %xmm2 # 0x1e0bad0
vmovaps %xmm2, 0xb80(%rsp)
vmovaps 0xba0(%rsp), %xmm3
vmovaps 0xb90(%rsp), %xmm2
vmovaps %xmm3, 0xbe0(%rsp)
vmovaps %xmm2, 0xbd0(%rsp)
vmovaps 0xbe0(%rsp), %xmm2
vmovaps 0xbd0(%rsp), %xmm3
vmulps %xmm3, %xmm2, %xmm3
vmovaps 0xb80(%rsp), %xmm2
vmovaps %xmm3, 0xbc0(%rsp)
vmovaps %xmm2, 0xbb0(%rsp)
vmovaps 0xbc0(%rsp), %xmm2
vmovaps 0xbb0(%rsp), %xmm3
vaddps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm3, 0xb30(%rsp)
vmovaps %xmm2, 0xb20(%rsp)
vmovaps 0x7d2e76(%rip), %xmm2 # 0x1e0bae0
vmovaps %xmm2, 0xb10(%rsp)
vmovaps 0xb30(%rsp), %xmm3
vmovaps 0xb20(%rsp), %xmm2
vmovaps %xmm3, 0xb70(%rsp)
vmovaps %xmm2, 0xb60(%rsp)
vmovaps 0xb70(%rsp), %xmm2
vmovaps 0xb60(%rsp), %xmm3
vmulps %xmm3, %xmm2, %xmm3
vmovaps 0xb10(%rsp), %xmm2
vmovaps %xmm3, 0xb50(%rsp)
vmovaps %xmm2, 0xb40(%rsp)
vmovaps 0xb50(%rsp), %xmm2
vmovaps 0xb40(%rsp), %xmm3
vaddps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm3, 0xac0(%rsp)
vmovaps %xmm2, 0xab0(%rsp)
vmovaps 0x7d2de6(%rip), %xmm2 # 0x1e0baf0
vmovaps %xmm2, 0xaa0(%rsp)
vmovaps 0xac0(%rsp), %xmm3
vmovaps 0xab0(%rsp), %xmm2
vmovaps %xmm3, 0xb00(%rsp)
vmovaps %xmm2, 0xaf0(%rsp)
vmovaps 0xb00(%rsp), %xmm2
vmovaps 0xaf0(%rsp), %xmm3
vmulps %xmm3, %xmm2, %xmm3
vmovaps 0xaa0(%rsp), %xmm2
vmovaps %xmm3, 0xae0(%rsp)
vmovaps %xmm2, 0xad0(%rsp)
vmovaps 0xae0(%rsp), %xmm2
vmovaps 0xad0(%rsp), %xmm3
vaddps %xmm3, %xmm2, %xmm2
vmovaps %xmm2, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x50(%rsp), %xmm2
vmovaps %xmm3, 0xa50(%rsp)
vmovaps %xmm2, 0xa40(%rsp)
vmovaps %xmm1, 0xa30(%rsp)
vmovaps 0xa50(%rsp), %xmm2
vmovaps 0xa40(%rsp), %xmm1
vmovaps %xmm2, 0xa90(%rsp)
vmovaps %xmm1, 0xa80(%rsp)
vmovaps 0xa90(%rsp), %xmm1
vmovaps 0xa80(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm2
vmovaps 0xa30(%rsp), %xmm1
vmovaps %xmm2, 0xa70(%rsp)
vmovaps %xmm1, 0xa60(%rsp)
vmovaps 0xa70(%rsp), %xmm1
vmovaps 0xa60(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm3
vmovaps 0x40(%rsp), %xmm2
vmovaps 0x50(%rsp), %xmm1
vmovaps %xmm3, 0x9e0(%rsp)
vmovaps %xmm2, 0x9d0(%rsp)
vmovaps %xmm1, 0x9c0(%rsp)
vmovaps 0x9e0(%rsp), %xmm2
vmovaps 0x9d0(%rsp), %xmm1
vmovaps %xmm2, 0xa20(%rsp)
vmovaps %xmm1, 0xa10(%rsp)
vmovaps 0xa20(%rsp), %xmm1
vmovaps 0xa10(%rsp), %xmm2
vmulps %xmm2, %xmm1, %xmm2
vmovaps 0x9c0(%rsp), %xmm1
vmovaps %xmm2, 0xa00(%rsp)
vmovaps %xmm1, 0x9f0(%rsp)
vmovaps 0xa00(%rsp), %xmm1
vmovaps 0x9f0(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm2
vmovaps 0x10(%rsp), %xmm1
vmovaps %xmm2, 0x70(%rsp)
vmovaps %xmm1, 0x60(%rsp)
vmovaps 0x70(%rsp), %xmm1
vmovaps 0x60(%rsp), %xmm2
vaddps %xmm2, %xmm1, %xmm1
vmovaps %xmm1, -0x10(%rsp)
vmovaps 0x30(%rsp), %xmm1
vmovaps %xmm1, 0x330(%rsp)
vcvttps2dq 0x330(%rsp), %xmm1
vmovdqa %xmm1, 0x20(%rsp)
vmovdqa 0x20(%rsp), %xmm1
vmovdqa %xmm1, 0xc70(%rsp)
vmovdqa %xmm0, 0xc60(%rsp)
vmovdqa 0xc70(%rsp), %xmm0
vmovdqa 0xc60(%rsp), %xmm1
vpaddd %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x20(%rsp)
vmovdqa 0x20(%rsp), %xmm0
vmovdqa %xmm0, 0xc90(%rsp)
movl $0x17, 0xc8c(%rsp)
vmovdqa 0xc90(%rsp), %xmm0
movl 0xc8c(%rsp), %eax
vmovd %eax, %xmm1
vpslld %xmm1, %xmm0, %xmm0
vmovdqa %xmm0, 0x20(%rsp)
vmovdqa 0x20(%rsp), %xmm0
vmovdqa %xmm0, 0xca0(%rsp)
vmovdqa 0xca0(%rsp), %xmm0
vmovaps %xmm0, -0x20(%rsp)
vmovaps -0x10(%rsp), %xmm1
vmovaps -0x20(%rsp), %xmm0
vmovaps %xmm1, 0xd0(%rsp)
vmovaps %xmm0, 0xc0(%rsp)
vmovaps 0xd0(%rsp), %xmm0
vmulps 0xc0(%rsp), %xmm0, %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps -0x10(%rsp), %xmm0
addq $0xd88, %rsp # imm = 0xD88
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,127 |
ncnn::BinaryOp_x86_avx_functor::binary_op_pow::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return (float)pow(x, y);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq %rdx, (%rsp)
movq 0x8(%rsp), %rax
vmovss (%rax), %xmm0
movq (%rsp), %rax
vmovss (%rax), %xmm1
callq 0xa14a10
addq $0x18, %rsp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,128 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rsub::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_sub_ps(y, x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vsubps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,129 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rsub::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_sub_ps(y, x);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vsubps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,130 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rsub::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return y - x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x18(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x10(%rsp), %rax
vsubss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,131 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rdiv::func_pack8(float vector[8] const&, float vector[8] const&) const
|
__m256 func_pack8(const __m256& x, const __m256& y) const
{
return _mm256_div_ps(y, x);
}
|
pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovaps (%rax), %ymm1
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm1, 0x40(%rsp)
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %ymm0
vdivps 0x20(%rsp), %ymm0, %ymm0
movq %rbp, %rsp
popq %rbp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,132 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rdiv::func_pack4(float vector[4] const&, float vector[4] const&) const
|
__m128 func_pack4(const __m128& x, const __m128& y) const
{
return _mm_div_ps(y, x);
}
|
movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
movq %rdx, -0x40(%rsp)
movq -0x40(%rsp), %rax
vmovaps (%rax), %xmm1
movq -0x38(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm1, -0x18(%rsp)
vmovaps %xmm0, -0x28(%rsp)
vmovaps -0x18(%rsp), %xmm0
vdivps -0x28(%rsp), %xmm0, %xmm0
retq
nopw (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,133 |
ncnn::BinaryOp_x86_avx_functor::binary_op_rdiv::func(float const&, float const&) const
|
float func(const float& x, const float& y) const
{
return y / x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x18(%rsp), %rax
vmovss (%rax), %xmm0
movq -0x10(%rsp), %rax
vdivss (%rax), %xmm0, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp
|
2,113,134 |
ncnn::UnaryOp::UnaryOp()
|
UnaryOp::UnaryOp()
{
one_blob_only = true;
support_inplace = true;
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq 0x10(%rsp), %rdi
movq %rdi, 0x8(%rsp)
callq 0xb6b80
movq 0x8(%rsp), %rax
leaq 0x89883c(%rip), %rcx # 0x1ed19f0
addq $0x10, %rcx
movq %rcx, (%rax)
movb $0x1, 0x8(%rax)
movb $0x1, 0x9(%rax)
addq $0x18, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,135 |
ncnn::UnaryOp::load_param(ncnn::ParamDict const&)
|
int UnaryOp::load_param(const ParamDict& pd)
{
op_type = pd.get(0, 0);
return 0;
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x10(%rsp), %rax
movq %rax, (%rsp)
movq 0x8(%rsp), %rdi
xorl %edx, %edx
movl %edx, %esi
callq 0xaace0
movl %eax, %ecx
movq (%rsp), %rax
movl %ecx, 0xd0(%rax)
xorl %eax, %eax
addq $0x18, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,136 |
ncnn::UnaryOp::forward_inplace(ncnn::Mat&, ncnn::Option const&) const
|
int UnaryOp::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
if (op_type == Operation_ABS)
return unary_op_inplace<unary_op_abs>(bottom_top_blob, opt);
if (op_type == Operation_NEG)
return unary_op_inplace<unary_op_neg>(bottom_top_blob, opt);
if (op_type == Operation_FLOOR)
return unary_op_inplace<unary_op_floor>(bottom_top_blob, opt);
if (op_type == Operation_CEIL)
return unary_op_inplace<unary_op_ceil>(bottom_top_blob, opt);
if (op_type == Operation_SQUARE)
return unary_op_inplace<unary_op_square>(bottom_top_blob, opt);
if (op_type == Operation_SQRT)
return unary_op_inplace<unary_op_sqrt>(bottom_top_blob, opt);
if (op_type == Operation_RSQRT)
return unary_op_inplace<unary_op_rsqrt>(bottom_top_blob, opt);
if (op_type == Operation_EXP)
return unary_op_inplace<unary_op_exp>(bottom_top_blob, opt);
if (op_type == Operation_LOG)
return unary_op_inplace<unary_op_log>(bottom_top_blob, opt);
if (op_type == Operation_SIN)
return unary_op_inplace<unary_op_sin>(bottom_top_blob, opt);
if (op_type == Operation_COS)
return unary_op_inplace<unary_op_cos>(bottom_top_blob, opt);
if (op_type == Operation_TAN)
return unary_op_inplace<unary_op_tan>(bottom_top_blob, opt);
if (op_type == Operation_ASIN)
return unary_op_inplace<unary_op_asin>(bottom_top_blob, opt);
if (op_type == Operation_ACOS)
return unary_op_inplace<unary_op_acos>(bottom_top_blob, opt);
if (op_type == Operation_ATAN)
return unary_op_inplace<unary_op_atan>(bottom_top_blob, opt);
if (op_type == Operation_RECIPROCAL)
return unary_op_inplace<unary_op_reciprocal>(bottom_top_blob, opt);
if (op_type == Operation_TANH)
return unary_op_inplace<unary_op_tanh>(bottom_top_blob, opt);
return 0;
}
|
subq $0x28, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x18(%rsp), %rax
movq %rax, (%rsp)
cmpl $0x0, 0xd0(%rax)
jne 0x163924d
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16394b0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x1, 0xd0(%rax)
jne 0x1639272
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639560
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x2, 0xd0(%rax)
jne 0x1639297
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639610
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x3, 0xd0(%rax)
jne 0x16392bc
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16396c0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x4, 0xd0(%rax)
jne 0x16392e1
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639770
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x5, 0xd0(%rax)
jne 0x1639306
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639820
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x6, 0xd0(%rax)
jne 0x163932b
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16398d0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x7, 0xd0(%rax)
jne 0x1639350
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639980
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x8, 0xd0(%rax)
jne 0x1639375
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639a30
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x9, 0xd0(%rax)
jne 0x163939a
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639ae0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xa, 0xd0(%rax)
jne 0x16393bf
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639b90
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xb, 0xd0(%rax)
jne 0x16393e4
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639c40
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xc, 0xd0(%rax)
jne 0x1639409
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639cf0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xd, 0xd0(%rax)
jne 0x163942b
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639da0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xe, 0xd0(%rax)
jne 0x163944d
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639e50
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0xf, 0xd0(%rax)
jne 0x163946f
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639f00
movl %eax, 0x24(%rsp)
jmp 0x1639499
movq (%rsp), %rax
cmpl $0x10, 0xd0(%rax)
jne 0x1639491
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1639fb0
movl %eax, 0x24(%rsp)
jmp 0x1639499
movl $0x0, 0x24(%rsp)
movl 0x24(%rsp), %eax
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,137 |
int ncnn::unary_op_inplace<ncnn::unary_op_abs>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639554
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a090
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x16394e5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,138 |
int ncnn::unary_op_inplace<ncnn::unary_op_neg>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639604
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a0c0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639595
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,139 |
int ncnn::unary_op_inplace<ncnn::unary_op_floor>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x16396b4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a0f0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639645
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,140 |
int ncnn::unary_op_inplace<ncnn::unary_op_ceil>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639764
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a120
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x16396f5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,141 |
int ncnn::unary_op_inplace<ncnn::unary_op_square>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639814
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a150
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x16397a5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,142 |
int ncnn::unary_op_inplace<ncnn::unary_op_sqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x16398c4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a170
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639855
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,143 |
int ncnn::unary_op_inplace<ncnn::unary_op_rsqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639974
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a1a0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639905
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,144 |
int ncnn::unary_op_inplace<ncnn::unary_op_exp>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639a24
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a1d0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x16399b5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,145 |
int ncnn::unary_op_inplace<ncnn::unary_op_log>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639ad4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a200
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639a65
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,146 |
int ncnn::unary_op_inplace<ncnn::unary_op_sin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639b84
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a230
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639b15
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,147 |
int ncnn::unary_op_inplace<ncnn::unary_op_cos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639c34
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a280
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639bc5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,148 |
int ncnn::unary_op_inplace<ncnn::unary_op_tan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639ce4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a2d0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639c75
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,149 |
int ncnn::unary_op_inplace<ncnn::unary_op_asin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639d94
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a320
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639d25
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,150 |
int ncnn::unary_op_inplace<ncnn::unary_op_acos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639e44
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a370
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639dd5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,151 |
int ncnn::unary_op_inplace<ncnn::unary_op_atan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639ef4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a3c0
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639e85
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,152 |
int ncnn::unary_op_inplace<ncnn::unary_op_reciprocal>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x1639fa4
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a410
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639f35
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,153 |
int ncnn::unary_op_inplace<ncnn::unary_op_tanh>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int size = static_cast<int>(a.total());
#pragma omp parallel for num_threads(opt.num_threads)
for (int i = 0; i < size; i++)
{
a[i] = op(a[i]);
}
return 0;
}
|
subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x20(%rsp), %rcx
movq 0x40(%rcx), %rax
movslq 0x38(%rcx), %rcx
imulq %rcx, %rax
movl %eax, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x8(%rsp), %eax
jge 0x163a054
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x40(%rsp)
movq %rax, 0x38(%rsp)
movq 0x40(%rsp), %rax
movq (%rax), %rsi
movq 0x38(%rsp), %rax
shlq $0x2, %rax
addq %rax, %rsi
leaq 0xf(%rsp), %rdi
callq 0x163a430
movq 0x18(%rsp), %rcx
movslq 0x4(%rsp), %rax
movq %rcx, 0x30(%rsp)
movq %rax, 0x28(%rsp)
movq 0x30(%rsp), %rax
movq (%rax), %rax
movq 0x28(%rsp), %rcx
movss %xmm0, (%rax,%rcx,4)
movl 0x4(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x4(%rsp)
jmp 0x1639fe5
xorl %eax, %eax
addq $0x48, %rsp
retq
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,154 |
ncnn::unary_op_abs::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)fabs(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x102e2f0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,155 |
ncnn::unary_op_neg::operator()(float const&) const
|
float operator()(const float& x) const
{
return -x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss (%rax), %xmm0
movd %xmm0, %eax
xorl $0x80000000, %eax # imm = 0x80000000
movd %eax, %xmm0
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,156 |
ncnn::unary_op_floor::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)floor(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x8e720
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,157 |
ncnn::unary_op_ceil::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)ceil(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x1058980
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,158 |
ncnn::unary_op_square::operator()(float const&) const
|
float operator()(const float& x) const
{
return x * x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss (%rax), %xmm0
movq -0x10(%rsp), %rax
mulss (%rax), %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,159 |
ncnn::unary_op_sqrt::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)sqrt(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x671b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,160 |
ncnn::unary_op_rsqrt::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)(1.f / sqrt(x));
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x671b0
movaps %xmm0, %xmm1
movss 0x7cee3d(%rip), %xmm0 # 0x1e09004
divss %xmm1, %xmm0
addq $0x18, %rsp
retq
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,161 |
ncnn::unary_op_exp::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)exp(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x100cf0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,162 |
ncnn::unary_op_log::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)log(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x100cd0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,163 |
ncnn::unary_op_sin::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)sin(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a260
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,164 |
std::sin(float)
|
inline _GLIBCXX_CONSTEXPR float
sin(float __x)
{ return __builtin_sinf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f450
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,165 |
ncnn::unary_op_cos::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)cos(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a2b0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,166 |
std::cos(float)
|
inline _GLIBCXX_CONSTEXPR float
cos(float __x)
{ return __builtin_cosf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f540
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,167 |
ncnn::unary_op_tan::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)tan(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a300
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,168 |
std::tan(float)
|
inline _GLIBCXX_CONSTEXPR float
tan(float __x)
{ return __builtin_tanf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f560
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,169 |
ncnn::unary_op_asin::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)asin(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a350
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,170 |
std::asin(float)
|
inline _GLIBCXX_CONSTEXPR float
asin(float __x)
{ return __builtin_asinf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f2b0
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,171 |
ncnn::unary_op_acos::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)acos(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a3a0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,172 |
std::acos(float)
|
inline _GLIBCXX_CONSTEXPR float
acos(float __x)
{ return __builtin_acosf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f0d0
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,173 |
ncnn::unary_op_atan::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)atan(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x163a3f0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,174 |
std::atan(float)
|
inline _GLIBCXX_CONSTEXPR float
atan(float __x)
{ return __builtin_atanf(__x); }
|
pushq %rax
movss %xmm0, 0x4(%rsp)
movss 0x4(%rsp), %xmm0
callq 0x5f600
popq %rax
retq
nopw %cs:(%rax,%rax)
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
/usr/lib/gcc/x86_64-linux-gnu/11/../../../../include/c++/11/cmath
|
2,113,175 |
ncnn::unary_op_reciprocal::operator()(float const&) const
|
float operator()(const float& x) const
{
return 1.f / x;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
movss 0x7cebdd(%rip), %xmm0 # 0x1e09004
divss (%rax), %xmm0
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,176 |
ncnn::unary_op_tanh::operator()(float const&) const
|
float operator()(const float& x) const
{
return (float)tanh(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x140cc0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/unaryop.cpp
|
2,113,177 |
ncnn::UnaryOp_x86::UnaryOp_x86()
|
UnaryOp_x86::UnaryOp_x86()
{
#if __SSE2__
support_packing = true;
#endif // __SSE2__
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x8(%rsp), %rax
movq -0x10(%rsp), %rcx
movq (%rcx), %rdx
movq %rdx, (%rax)
movq 0x8(%rcx), %rdx
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movq %rdx, (%rax,%rcx)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movb $0x1, 0xb(%rax,%rcx)
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,178 |
ncnn::UnaryOp_x86::UnaryOp_x86()
|
UnaryOp_x86::UnaryOp_x86()
{
#if __SSE2__
support_packing = true;
#endif // __SSE2__
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq 0x10(%rsp), %rdi
movq %rdi, 0x8(%rsp)
addq $0x8, %rdi
callq 0x1639190
movq 0x8(%rsp), %rax
leaq 0x8975a0(%rip), %rcx # 0x1ed1a68
addq $0x18, %rcx
movq %rcx, (%rax)
leaq 0x897592(%rip), %rcx # 0x1ed1a68
addq $0x88, %rcx
movq %rcx, 0x8(%rax)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
movb $0x1, 0xb(%rax,%rcx)
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,179 |
ncnn::UnaryOp_x86::forward_inplace(ncnn::Mat&, ncnn::Option const&) const
|
int UnaryOp_x86::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_functor;
if (op_type == Operation_ABS)
return unary_op_inplace<unary_op_abs>(bottom_top_blob, opt);
if (op_type == Operation_NEG)
return unary_op_inplace<unary_op_neg>(bottom_top_blob, opt);
if (op_type == Operation_FLOOR)
return unary_op_inplace<unary_op_floor>(bottom_top_blob, opt);
if (op_type == Operation_CEIL)
return unary_op_inplace<unary_op_ceil>(bottom_top_blob, opt);
if (op_type == Operation_SQUARE)
return unary_op_inplace<unary_op_square>(bottom_top_blob, opt);
if (op_type == Operation_SQRT)
return unary_op_inplace<unary_op_sqrt>(bottom_top_blob, opt);
if (op_type == Operation_RSQRT)
return unary_op_inplace<unary_op_rsqrt>(bottom_top_blob, opt);
if (op_type == Operation_EXP)
return unary_op_inplace<unary_op_exp>(bottom_top_blob, opt);
if (op_type == Operation_LOG)
return unary_op_inplace<unary_op_log>(bottom_top_blob, opt);
if (op_type == Operation_SIN)
return unary_op_inplace<unary_op_sin>(bottom_top_blob, opt);
if (op_type == Operation_COS)
return unary_op_inplace<unary_op_cos>(bottom_top_blob, opt);
if (op_type == Operation_TAN)
return unary_op_inplace<unary_op_tan>(bottom_top_blob, opt);
if (op_type == Operation_ASIN)
return unary_op_inplace<unary_op_asin>(bottom_top_blob, opt);
if (op_type == Operation_ACOS)
return unary_op_inplace<unary_op_acos>(bottom_top_blob, opt);
if (op_type == Operation_ATAN)
return unary_op_inplace<unary_op_atan>(bottom_top_blob, opt);
if (op_type == Operation_RECIPROCAL)
return unary_op_inplace<unary_op_reciprocal>(bottom_top_blob, opt);
if (op_type == Operation_TANH)
return unary_op_inplace<unary_op_tanh>(bottom_top_blob, opt);
return 0;
}
|
subq $0x28, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rdx, 0x8(%rsp)
movq 0x18(%rsp), %rax
movq %rax, (%rsp)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x0, 0xd0(%rax,%rcx)
jne 0x163a545
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163a820
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x1, 0xd0(%rax,%rcx)
jne 0x163a572
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163ae00
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x2, 0xd0(%rax,%rcx)
jne 0x163a59f
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163b3e0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x3, 0xd0(%rax,%rcx)
jne 0x163a5cc
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163b9c0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x4, 0xd0(%rax,%rcx)
jne 0x163a5f9
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163bfa0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x5, 0xd0(%rax,%rcx)
jne 0x163a626
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163c580
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x6, 0xd0(%rax,%rcx)
jne 0x163a653
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163cb60
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x7, 0xd0(%rax,%rcx)
jne 0x163a680
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163d140
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x8, 0xd0(%rax,%rcx)
jne 0x163a6ad
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163d720
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x9, 0xd0(%rax,%rcx)
jne 0x163a6da
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163dd00
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xa, 0xd0(%rax,%rcx)
jne 0x163a707
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163e2e0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xb, 0xd0(%rax,%rcx)
jne 0x163a734
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163e8c0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xc, 0xd0(%rax,%rcx)
jne 0x163a761
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163eea0
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xd, 0xd0(%rax,%rcx)
jne 0x163a78e
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163f480
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xe, 0xd0(%rax,%rcx)
jne 0x163a7b8
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x163fa60
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xf, 0xd0(%rax,%rcx)
jne 0x163a7e2
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1640040
movl %eax, 0x24(%rsp)
jmp 0x163a814
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x10, 0xd0(%rax,%rcx)
jne 0x163a80c
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1640620
movl %eax, 0x24(%rsp)
jmp 0x163a814
movl $0x0, 0x24(%rsp)
movl 0x24(%rsp), %eax
addq $0x28, %rsp
retq
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,180 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_abs>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x1b8, %rsp # imm = 0x1B8
movq %rdi, 0xc8(%rsp)
movq %rsi, 0xc0(%rsp)
movq 0xc8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb8(%rsp)
movq 0xc8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xb4(%rsp)
movq 0xc8(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc8(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa8(%rsp)
movl 0xb8(%rsp), %eax
imull 0xb4(%rsp), %eax
imull 0xb0(%rsp), %eax
imull 0xa8(%rsp), %eax
movl %eax, 0xa4(%rsp)
movl $0x0, 0xa0(%rsp)
movl 0xa0(%rsp), %eax
cmpl 0xac(%rsp), %eax
jge 0x163adeb
movq 0xc8(%rsp), %rcx
movl 0xa0(%rsp), %eax
leaq 0x50(%rsp), %rdx
movq %rdx, 0xe0(%rsp)
movq %rcx, 0xd8(%rsp)
movl %eax, 0xd4(%rsp)
movq 0xd8(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xd3(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xd4(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x50(%rsp), %r10
movq %r10, 0x158(%rsp)
movl %r9d, 0x154(%rsp)
movl %r8d, 0x150(%rsp)
movl %edi, 0x14c(%rsp)
movq %rsi, 0x140(%rsp)
movq %rdx, 0x138(%rsp)
movl %ecx, 0x134(%rsp)
movq %rax, 0x128(%rsp)
movq 0x158(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x140(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x138(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x134(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x128(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x154(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x150(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x168(%rsp)
movl $0x10, 0x164(%rsp)
movq 0x168(%rsp), %rax
movslq 0x164(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x164(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x78(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163aa7d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x90(%rsp)
movb $0x1, 0xd3(%rsp)
testb $0x1, 0xd3(%rsp)
jne 0x163aba3
leaq 0x50(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x198(%rsp)
movq 0x198(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163ab49
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x194(%rsp) # imm = 0xFFFFFFFF
movl 0x194(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x190(%rsp)
cmpl $0x1, 0x190(%rsp)
jne 0x163ab49
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163ab1d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163ab1b
jmp 0x163ab47
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1a0(%rsp)
cmpq $0x0, 0x1a0(%rsp)
je 0x163ab45
movq 0x1a0(%rsp), %rdi
callq 0x5f480
jmp 0x163ab47
jmp 0x163ab49
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163aba1
movq %rax, %rdi
callq 0x678a0
jmp 0x163aba3
leaq 0x50(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x50(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movq %rax, 0x188(%rsp)
movq 0x188(%rsp), %rax
movq %rax, 0x8(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163ac76
movq 0x8(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x184(%rsp) # imm = 0xFFFFFFFF
movl 0x184(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x180(%rsp)
cmpl $0x1, 0x180(%rsp)
jne 0x163ac76
movq 0x8(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163ac4a
movq 0x8(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163ac48
jmp 0x163ac74
movq 0x8(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1a8(%rsp)
cmpq $0x0, 0x1a8(%rsp)
je 0x163ac72
movq 0x1a8(%rsp), %rdi
callq 0x5f480
jmp 0x163ac74
jmp 0x163ac76
movq 0x8(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163acce
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0x98(%rsp)
movl $0x0, 0x40(%rsp)
movl 0x40(%rsp), %eax
addl $0x3, %eax
cmpl 0xa4(%rsp), %eax
jge 0x163ad82
movq 0x98(%rsp), %rax
movq %rax, 0x108(%rsp)
movq 0x108(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xbf(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1640c30
movaps %xmm0, 0x30(%rsp)
movq 0x98(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x120(%rsp)
movaps %xmm0, 0x110(%rsp)
movaps 0x110(%rsp), %xmm0
movq 0x120(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x98(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x98(%rsp)
movl 0x40(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x40(%rsp)
jmp 0x163ace3
jmp 0x163ad84
movl 0x40(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163add3
movq 0x98(%rsp), %rsi
leaq 0xbf(%rsp), %rdi
callq 0x1640ca0
movq 0x98(%rsp), %rax
movss %xmm0, (%rax)
movq 0x98(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x98(%rsp)
movl 0x40(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x40(%rsp)
jmp 0x163ad84
jmp 0x163add5
movl 0xa0(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xa0(%rsp)
jmp 0x163a8c2
xorl %eax, %eax
addq $0x1b8, %rsp # imm = 0x1B8
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,181 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_neg>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163b3cb
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163b05d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163b183
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163b129
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163b129
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163b0fd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163b0fb
jmp 0x163b127
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163b125
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163b127
jmp 0x163b129
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163b181
movq %rax, %rdi
callq 0x678a0
jmp 0x163b183
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163b256
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163b256
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163b22a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163b228
jmp 0x163b254
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163b252
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163b254
jmp 0x163b256
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163b2ae
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163b362
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1640cd0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163b2c3
jmp 0x163b364
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163b3b3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1640d10
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163b364
jmp 0x163b3b5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163aea2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,182 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_floor>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163b9ab
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163b63d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163b763
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163b709
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163b709
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163b6dd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163b6db
jmp 0x163b707
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163b705
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163b707
jmp 0x163b709
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163b761
movq %rax, %rdi
callq 0x678a0
jmp 0x163b763
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163b836
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163b836
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163b80a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163b808
jmp 0x163b834
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163b832
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163b834
jmp 0x163b836
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163b88e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163b942
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1640d40
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163b8a3
jmp 0x163b944
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163b993
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1641000
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163b944
jmp 0x163b995
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163b482
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,183 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_ceil>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163bf8b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163bc1d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163bd43
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163bce9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163bce9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163bcbd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163bcbb
jmp 0x163bce7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163bce5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163bce7
jmp 0x163bce9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163bd41
movq %rax, %rdi
callq 0x678a0
jmp 0x163bd43
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163be16
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163be16
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163bdea
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163bde8
jmp 0x163be14
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163be12
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163be14
jmp 0x163be16
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163be6e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163bf22
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1641030
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163be83
jmp 0x163bf24
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163bf73
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1641340
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163bf24
jmp 0x163bf75
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163ba62
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,184 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_square>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163c56b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163c1fd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163c323
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163c2c9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163c2c9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163c29d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163c29b
jmp 0x163c2c7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163c2c5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163c2c7
jmp 0x163c2c9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163c321
movq %rax, %rdi
callq 0x678a0
jmp 0x163c323
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163c3f6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163c3f6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163c3ca
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163c3c8
jmp 0x163c3f4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163c3f2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163c3f4
jmp 0x163c3f6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163c44e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163c502
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1641370
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163c463
jmp 0x163c504
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163c553
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x16413a0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163c504
jmp 0x163c555
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163c042
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,185 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_sqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163cb4b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163c7dd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163c903
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163c8a9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163c8a9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163c87d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163c87b
jmp 0x163c8a7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163c8a5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163c8a7
jmp 0x163c8a9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163c901
movq %rax, %rdi
callq 0x678a0
jmp 0x163c903
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163c9d6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163c9d6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163c9aa
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163c9a8
jmp 0x163c9d4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163c9d2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163c9d4
jmp 0x163c9d6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163ca2e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163cae2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x16413c0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163ca43
jmp 0x163cae4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163cb33
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1641420
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163cae4
jmp 0x163cb35
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163c622
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,186 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_rsqrt>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163d12b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163cdbd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163cee3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163ce89
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163ce89
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163ce5d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163ce5b
jmp 0x163ce87
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163ce85
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163ce87
jmp 0x163ce89
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163cee1
movq %rax, %rdi
callq 0x678a0
jmp 0x163cee3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163cfb6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163cfb6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163cf8a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163cf88
jmp 0x163cfb4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163cfb2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163cfb4
jmp 0x163cfb6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163d00e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163d0c2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1641450
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163d023
jmp 0x163d0c4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163d113
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1641470
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163d0c4
jmp 0x163d115
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163cc02
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,187 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_exp>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163d70b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163d39d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163d4c3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163d469
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163d469
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163d43d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163d43b
jmp 0x163d467
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163d465
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163d467
jmp 0x163d469
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163d4c1
movq %rax, %rdi
callq 0x678a0
jmp 0x163d4c3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163d596
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163d596
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163d56a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163d568
jmp 0x163d594
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163d592
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163d594
jmp 0x163d596
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163d5ee
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163d6a2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x16414a0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163d603
jmp 0x163d6a4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163d6f3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1641bf0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163d6a4
jmp 0x163d6f5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163d1e2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,188 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_log>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163dceb
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163d97d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163daa3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163da49
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163da49
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163da1d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163da1b
jmp 0x163da47
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163da45
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163da47
jmp 0x163da49
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163daa1
movq %rax, %rdi
callq 0x678a0
jmp 0x163daa3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163db76
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163db76
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163db4a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163db48
jmp 0x163db74
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163db72
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163db74
jmp 0x163db76
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163dbce
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163dc82
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1641c20
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163dbe3
jmp 0x163dc84
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163dcd3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1642600
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163dc84
jmp 0x163dcd5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163d7c2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,189 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_sin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163e2cb
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163df5d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163e083
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163e029
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163e029
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163dffd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163dffb
jmp 0x163e027
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163e025
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163e027
jmp 0x163e029
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163e081
movq %rax, %rdi
callq 0x678a0
jmp 0x163e083
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163e156
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163e156
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163e12a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163e128
jmp 0x163e154
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163e152
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163e154
jmp 0x163e156
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163e1ae
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163e262
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1642630
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163e1c3
jmp 0x163e264
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163e2b3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1642fe0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163e264
jmp 0x163e2b5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163dda2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,190 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_cos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163e8ab
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163e53d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163e663
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163e609
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163e609
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163e5dd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163e5db
jmp 0x163e607
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163e605
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163e607
jmp 0x163e609
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163e661
movq %rax, %rdi
callq 0x678a0
jmp 0x163e663
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163e736
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163e736
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163e70a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163e708
jmp 0x163e734
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163e732
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163e734
jmp 0x163e736
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163e78e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163e842
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1643010
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163e7a3
jmp 0x163e844
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163e893
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1643970
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163e844
jmp 0x163e895
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163e382
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,191 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_tan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163ee8b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163eb1d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163ec43
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163ebe9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163ebe9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163ebbd
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163ebbb
jmp 0x163ebe7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163ebe5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163ebe7
jmp 0x163ebe9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163ec41
movq %rax, %rdi
callq 0x678a0
jmp 0x163ec43
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163ed16
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163ed16
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163ecea
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163ece8
jmp 0x163ed14
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163ed12
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163ed14
jmp 0x163ed16
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163ed6e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163ee22
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x16439a0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163ed83
jmp 0x163ee24
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163ee73
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1644720
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163ee24
jmp 0x163ee75
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163e962
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,192 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_asin>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163f46b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163f0fd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163f223
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163f1c9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163f1c9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163f19d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163f19b
jmp 0x163f1c7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163f1c5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163f1c7
jmp 0x163f1c9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163f221
movq %rax, %rdi
callq 0x678a0
jmp 0x163f223
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163f2f6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163f2f6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163f2ca
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163f2c8
jmp 0x163f2f4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163f2f2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163f2f4
jmp 0x163f2f6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163f34e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163f402
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1644750
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163f363
jmp 0x163f404
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163f453
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x16447e0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163f404
jmp 0x163f455
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163ef42
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,193 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_acos>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x163fa4b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163f6dd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163f803
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163f7a9
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163f7a9
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163f77d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163f77b
jmp 0x163f7a7
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163f7a5
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163f7a7
jmp 0x163f7a9
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163f801
movq %rax, %rdi
callq 0x678a0
jmp 0x163f803
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163f8d6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163f8d6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163f8aa
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163f8a8
jmp 0x163f8d4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163f8d2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163f8d4
jmp 0x163f8d6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163f92e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163f9e2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1644810
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163f943
jmp 0x163f9e4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x163fa33
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x16448a0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163f9e4
jmp 0x163fa35
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163f522
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,194 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_atan>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x164002b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x163fcbd
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x163fde3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163fd89
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x163fd89
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163fd5d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163fd5b
jmp 0x163fd87
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x163fd85
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x163fd87
jmp 0x163fd89
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163fde1
movq %rax, %rdi
callq 0x678a0
jmp 0x163fde3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x163feb6
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x163feb6
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x163fe8a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x163fe88
jmp 0x163feb4
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x163feb2
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x163feb4
jmp 0x163feb6
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x163ff0e
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x163ffc2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x16448d0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x163ff23
jmp 0x163ffc4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x1640013
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1644960
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x163ffc4
jmp 0x1640015
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x163fb02
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,195 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_reciprocal>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x164060b
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164029d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x16403c3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1640369
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x1640369
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164033d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x164033b
jmp 0x1640367
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x1640365
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x1640367
jmp 0x1640369
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16403c1
movq %rax, %rdi
callq 0x678a0
jmp 0x16403c3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1640496
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x1640496
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164046a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x1640468
jmp 0x1640494
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x1640492
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x1640494
jmp 0x1640496
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16404ee
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x16405a2
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x1644990
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x1640503
jmp 0x16405a4
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x16405f3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x16449c0
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x16405a4
jmp 0x16405f5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x16400e2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,196 |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_functor::unary_op_tanh>(ncnn::Mat&, ncnn::Option const&)
|
static int unary_op_inplace(Mat& a, const Option& opt)
{
Op op;
int w = a.w;
int h = a.h;
int d = a.d;
int channels = a.c;
int elempack = a.elempack;
int size = w * h * d * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr = a.channel(q);
int i = 0;
#if __SSE2__
#if __AVX__
#if __AVX512F__
for (; i + 15 < size; i += 16)
{
__m512 _p = _mm512_loadu_ps(ptr);
_p = op.func_pack16(_p);
_mm512_storeu_ps(ptr, _p);
ptr += 16;
}
#endif // __AVX512F__
for (; i + 7 < size; i += 8)
{
__m256 _p = _mm256_loadu_ps(ptr);
_p = op.func_pack8(_p);
_mm256_storeu_ps(ptr, _p);
ptr += 8;
}
#endif // __AVX__
for (; i + 3 < size; i += 4)
{
__m128 _p = _mm_load_ps(ptr);
_p = op.func_pack4(_p);
_mm_store_ps(ptr, _p);
ptr += 4;
}
#endif // __SSE2__
for (; i < size; i++)
{
*ptr = op.func(*ptr);
ptr++;
}
}
return 0;
}
|
subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xac(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xa8(%rsp)
movq 0xc0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xa0(%rsp)
movl 0xb0(%rsp), %eax
imull 0xac(%rsp), %eax
imull 0xa8(%rsp), %eax
imull 0xa0(%rsp), %eax
movl %eax, 0x9c(%rsp)
movl $0x0, 0x98(%rsp)
movl 0x98(%rsp), %eax
cmpl 0xa4(%rsp), %eax
jge 0x1640beb
movq 0xc0(%rsp), %rcx
movl 0x98(%rsp), %eax
leaq 0x48(%rsp), %rdx
movq %rdx, 0xd8(%rsp)
movq %rcx, 0xd0(%rsp)
movl %eax, 0xcc(%rsp)
movq 0xd0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xcb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xcc(%rsp), %rdx
imulq %rdx, %rcx
imulq 0x10(%rax), %rcx
addq %rcx, %rsi
movq 0x10(%rax), %rdx
movl 0x18(%rax), %ecx
movq 0x20(%rax), %rax
leaq 0x48(%rsp), %r10
movq %r10, 0x150(%rsp)
movl %r9d, 0x14c(%rsp)
movl %r8d, 0x148(%rsp)
movl %edi, 0x144(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movl %ecx, 0x12c(%rsp)
movq %rax, 0x120(%rsp)
movq 0x150(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x138(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x130(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x12c(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x120(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x14c(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x148(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x144(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x160(%rsp)
movl $0x10, 0x15c(%rsp)
movq 0x160(%rsp), %rax
movslq 0x15c(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x15c(%rsp), %edx
movslq %edx, %rdx
andq %rdx, %rax
xorl %edx, %edx
divq 0x10(%rcx)
movq %rax, %rdx
movq 0x28(%rsp), %rax
movq %rdx, 0x40(%rcx)
movl 0x28(%rax), %ecx
subl $0x1, %ecx
movl %ecx, 0x70(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x164087d
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0x88(%rsp)
movb $0x1, 0xcb(%rsp)
testb $0x1, 0xcb(%rsp)
jne 0x16409a3
leaq 0x48(%rsp), %rax
movq %rax, 0xe8(%rsp)
movq 0xe8(%rsp), %rax
movq %rax, 0x180(%rsp)
movq 0x180(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1640949
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x17c(%rsp) # imm = 0xFFFFFFFF
movl 0x17c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x178(%rsp)
cmpl $0x1, 0x178(%rsp)
jne 0x1640949
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x164091d
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x164091b
jmp 0x1640947
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x188(%rsp)
cmpq $0x0, 0x188(%rsp)
je 0x1640945
movq 0x188(%rsp), %rdi
callq 0x5f480
jmp 0x1640947
jmp 0x1640949
movq 0x18(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x16409a1
movq %rax, %rdi
callq 0x678a0
jmp 0x16409a3
leaq 0x48(%rsp), %rax
movq %rax, 0xe0(%rsp)
movq 0xe0(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x48(%rsp), %rax
movq %rax, 0xf0(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x170(%rsp)
movq 0x170(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x1640a76
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x16c(%rsp) # imm = 0xFFFFFFFF
movl 0x16c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x168(%rsp)
cmpl $0x1, 0x168(%rsp)
jne 0x1640a76
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x1640a4a
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
callq *%rax
jmp 0x1640a48
jmp 0x1640a74
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x190(%rsp)
cmpq $0x0, 0x190(%rsp)
je 0x1640a72
movq 0x190(%rsp), %rdi
callq 0x5f480
jmp 0x1640a74
jmp 0x1640a76
movq 0x10(%rsp), %rax
movq $0x0, (%rax)
movq $0x0, 0x10(%rax)
movl $0x0, 0x18(%rax)
movl $0x0, 0x28(%rax)
movl $0x0, 0x2c(%rax)
movl $0x0, 0x30(%rax)
movl $0x0, 0x34(%rax)
movl $0x0, 0x38(%rax)
movq $0x0, 0x40(%rax)
movq $0x0, 0x8(%rax)
jmp 0x1640ace
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0x90(%rsp)
movl $0x0, 0x44(%rsp)
movl 0x44(%rsp), %eax
addl $0x3, %eax
cmpl 0x9c(%rsp), %eax
jge 0x1640b82
movq 0x90(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0xf8(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, 0x30(%rsp)
leaq 0xb7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
callq 0x16449e0
movaps %xmm0, 0x30(%rsp)
movq 0x90(%rsp), %rax
movaps 0x30(%rsp), %xmm0
movq %rax, 0x118(%rsp)
movaps %xmm0, 0x100(%rsp)
movaps 0x100(%rsp), %xmm0
movq 0x118(%rsp), %rax
movaps %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x10, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x44(%rsp)
jmp 0x1640ae3
jmp 0x1640b84
movl 0x44(%rsp), %eax
cmpl 0x9c(%rsp), %eax
jge 0x1640bd3
movq 0x90(%rsp), %rsi
leaq 0xb7(%rsp), %rdi
callq 0x1645380
movq 0x90(%rsp), %rax
movss %xmm0, (%rax)
movq 0x90(%rsp), %rax
addq $0x4, %rax
movq %rax, 0x90(%rsp)
movl 0x44(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x44(%rsp)
jmp 0x1640b84
jmp 0x1640bd5
movl 0x98(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x98(%rsp)
jmp 0x16406c2
xorl %eax, %eax
addq $0x198, %rsp # imm = 0x198
retq
nopw %cs:(%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,197 |
virtual thunk to ncnn::UnaryOp_x86::forward_inplace(ncnn::Mat&, ncnn::Option const&) const
|
int UnaryOp_x86::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_functor;
if (op_type == Operation_ABS)
return unary_op_inplace<unary_op_abs>(bottom_top_blob, opt);
if (op_type == Operation_NEG)
return unary_op_inplace<unary_op_neg>(bottom_top_blob, opt);
if (op_type == Operation_FLOOR)
return unary_op_inplace<unary_op_floor>(bottom_top_blob, opt);
if (op_type == Operation_CEIL)
return unary_op_inplace<unary_op_ceil>(bottom_top_blob, opt);
if (op_type == Operation_SQUARE)
return unary_op_inplace<unary_op_square>(bottom_top_blob, opt);
if (op_type == Operation_SQRT)
return unary_op_inplace<unary_op_sqrt>(bottom_top_blob, opt);
if (op_type == Operation_RSQRT)
return unary_op_inplace<unary_op_rsqrt>(bottom_top_blob, opt);
if (op_type == Operation_EXP)
return unary_op_inplace<unary_op_exp>(bottom_top_blob, opt);
if (op_type == Operation_LOG)
return unary_op_inplace<unary_op_log>(bottom_top_blob, opt);
if (op_type == Operation_SIN)
return unary_op_inplace<unary_op_sin>(bottom_top_blob, opt);
if (op_type == Operation_COS)
return unary_op_inplace<unary_op_cos>(bottom_top_blob, opt);
if (op_type == Operation_TAN)
return unary_op_inplace<unary_op_tan>(bottom_top_blob, opt);
if (op_type == Operation_ASIN)
return unary_op_inplace<unary_op_asin>(bottom_top_blob, opt);
if (op_type == Operation_ACOS)
return unary_op_inplace<unary_op_acos>(bottom_top_blob, opt);
if (op_type == Operation_ATAN)
return unary_op_inplace<unary_op_atan>(bottom_top_blob, opt);
if (op_type == Operation_RECIPROCAL)
return unary_op_inplace<unary_op_reciprocal>(bottom_top_blob, opt);
if (op_type == Operation_TANH)
return unary_op_inplace<unary_op_tanh>(bottom_top_blob, opt);
return 0;
}
|
movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x58(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
movq -0x18(%rsp), %rdx
jmp 0x163a500
nopl (%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,198 |
ncnn::UnaryOp_x86_functor::unary_op_abs::func_pack4(float vector[4] const&) const
|
__m128 func_pack4(const __m128& x) const
{
return abs_sse(x);
}
|
movq %rdi, -0x70(%rsp)
movq %rsi, -0x78(%rsp)
movq -0x78(%rsp), %rax
movaps (%rax), %xmm0
movaps %xmm0, -0x58(%rsp)
movl $0x80000000, -0x3c(%rsp) # imm = 0x80000000
movss -0x3c(%rsp), %xmm0
movss %xmm0, -0x4(%rsp)
movss -0x4(%rsp), %xmm0
shufps $0x0, %xmm0, %xmm0 # xmm0 = xmm0[0,0,0,0]
movaps %xmm0, -0x18(%rsp)
movaps -0x18(%rsp), %xmm0
movaps %xmm0, -0x68(%rsp)
movaps -0x68(%rsp), %xmm1
movaps -0x58(%rsp), %xmm0
movaps %xmm1, -0x28(%rsp)
movaps %xmm0, -0x38(%rsp)
movaps -0x28(%rsp), %xmm0
pcmpeqd %xmm1, %xmm1
pxor %xmm1, %xmm0
movaps -0x38(%rsp), %xmm1
pand %xmm1, %xmm0
retq
nop
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
2,113,199 |
ncnn::UnaryOp_x86_functor::unary_op_abs::func(float const&) const
|
float func(const float& x) const
{
return (float)fabs(x);
}
|
subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
movss (%rax), %xmm0
callq 0x102e2f0
addq $0x18, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax,%rax)
|
ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json
|
O0
|
ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.