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
|
|---|---|---|---|---|---|---|
2,113,300 | ncnn::UnaryOp_x86_avx512_functor::unary_op_tan::func(float const&) const | float func(const float& x) const
{
return (float)tan(x);
} | subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%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/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,301 | ncnn::UnaryOp_x86_avx512_functor::unary_op_asin::func_pack16(float vector[16] const&) const | __m512 func_pack16(const __m512& x) const
{
//TODO avx512 optimize
float tmp[16];
_mm512_storeu_ps(tmp, x);
for (int i = 0; i < 16; i++)
tmp[i] = asin(tmp[i]);
return _mm512_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x70(%rsp)
movq %rsi, 0x68(%rsp)
leaq 0x20(%rsp), %rax
movq 0x68(%rsp), %rcx
vmovaps (%rcx), %zmm0
movq %rax, 0xe8(%rsp)
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm0
movq 0xe8(%rsp), %rax
vmovups %zmm0, (%rax)
movl $0x0, 0x1c(%rsp)
cmpl $0x10, 0x1c(%rsp)
jge 0x1657f69
movslq 0x1c(%rsp), %rax
vmovss 0x20(%rsp,%rax,4), %xmm0
vzeroupper
callq 0x163a350
movslq 0x1c(%rsp), %rax
vmovss %xmm0, 0x20(%rsp,%rax,4)
movl 0x1c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c(%rsp)
jmp 0x1657f37
leaq 0x20(%rsp), %rax
movq %rax, 0x78(%rsp)
movq 0x78(%rsp), %rax
vmovups (%rax), %zmm0
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/unaryop_x86_avx512.cpp |
2,113,302 | ncnn::UnaryOp_x86_avx512_functor::unary_op_asin::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
//TODO avx optimize
float tmp[8];
_mm256_storeu_ps(tmp, x);
tmp[0] = asin(tmp[0]);
tmp[1] = asin(tmp[1]);
tmp[2] = asin(tmp[2]);
tmp[3] = asin(tmp[3]);
tmp[4] = asin(tmp[4]);
tmp[5] = asin(tmp[5]);
tmp[6] = asin(tmp[6]);
tmp[7] = asin(tmp[7]);
return _mm256_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x30(%rsp)
movq %rsi, 0x28(%rsp)
movq %rsp, %rax
movq 0x28(%rsp), %rcx
vmovaps (%rcx), %ymm0
movq %rax, 0x68(%rsp)
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm0
movq 0x68(%rsp), %rax
vmovups %ymm0, (%rax)
vmovss (%rsp), %xmm0
vzeroupper
callq 0x163a350
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0xc(%rsp)
vmovss 0x10(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x10(%rsp)
vmovss 0x14(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x14(%rsp)
vmovss 0x18(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x18(%rsp)
vmovss 0x1c(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x1c(%rsp)
movq %rsp, %rax
movq %rax, 0x38(%rsp)
movq 0x38(%rsp), %rax
vmovups (%rax), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nop
| ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json | O0 | ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,303 | ncnn::UnaryOp_x86_avx512_functor::unary_op_asin::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = asin(tmp[0]);
tmp[1] = asin(tmp[1]);
tmp[2] = asin(tmp[2]);
tmp[3] = asin(tmp[3]);
return _mm_loadu_ps(tmp);
} | subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
vmovaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
vmovups %xmm0, (%rax)
vmovss (%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a350
vmovss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
vmovups (%rax), %xmm0
addq $0x48, %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/unaryop_x86_avx512.cpp |
2,113,304 | ncnn::UnaryOp_x86_avx512_functor::unary_op_asin::func(float const&) const | float func(const float& x) const
{
return (float)asin(x);
} | subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%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/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,305 | ncnn::UnaryOp_x86_avx512_functor::unary_op_acos::func_pack16(float vector[16] const&) const | __m512 func_pack16(const __m512& x) const
{
//TODO avx512 optimize
float tmp[16];
_mm512_storeu_ps(tmp, x);
for (int i = 0; i < 16; i++)
tmp[i] = acos(tmp[i]);
return _mm512_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x70(%rsp)
movq %rsi, 0x68(%rsp)
leaq 0x20(%rsp), %rax
movq 0x68(%rsp), %rcx
vmovaps (%rcx), %zmm0
movq %rax, 0xe8(%rsp)
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm0
movq 0xe8(%rsp), %rax
vmovups %zmm0, (%rax)
movl $0x0, 0x1c(%rsp)
cmpl $0x10, 0x1c(%rsp)
jge 0x16581b9
movslq 0x1c(%rsp), %rax
vmovss 0x20(%rsp,%rax,4), %xmm0
vzeroupper
callq 0x163a3a0
movslq 0x1c(%rsp), %rax
vmovss %xmm0, 0x20(%rsp,%rax,4)
movl 0x1c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c(%rsp)
jmp 0x1658187
leaq 0x20(%rsp), %rax
movq %rax, 0x78(%rsp)
movq 0x78(%rsp), %rax
vmovups (%rax), %zmm0
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/unaryop_x86_avx512.cpp |
2,113,306 | ncnn::UnaryOp_x86_avx512_functor::unary_op_acos::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
//TODO avx optimize
float tmp[8];
_mm256_storeu_ps(tmp, x);
tmp[0] = acos(tmp[0]);
tmp[1] = acos(tmp[1]);
tmp[2] = acos(tmp[2]);
tmp[3] = acos(tmp[3]);
tmp[4] = acos(tmp[4]);
tmp[5] = acos(tmp[5]);
tmp[6] = acos(tmp[6]);
tmp[7] = acos(tmp[7]);
return _mm256_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x30(%rsp)
movq %rsi, 0x28(%rsp)
movq %rsp, %rax
movq 0x28(%rsp), %rcx
vmovaps (%rcx), %ymm0
movq %rax, 0x68(%rsp)
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm0
movq 0x68(%rsp), %rax
vmovups %ymm0, (%rax)
vmovss (%rsp), %xmm0
vzeroupper
callq 0x163a3a0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0xc(%rsp)
vmovss 0x10(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x10(%rsp)
vmovss 0x14(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x14(%rsp)
vmovss 0x18(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x18(%rsp)
vmovss 0x1c(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x1c(%rsp)
movq %rsp, %rax
movq %rax, 0x38(%rsp)
movq 0x38(%rsp), %rax
vmovups (%rax), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nop
| ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json | O0 | ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,307 | ncnn::UnaryOp_x86_avx512_functor::unary_op_acos::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = acos(tmp[0]);
tmp[1] = acos(tmp[1]);
tmp[2] = acos(tmp[2]);
tmp[3] = acos(tmp[3]);
return _mm_loadu_ps(tmp);
} | subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
vmovaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
vmovups %xmm0, (%rax)
vmovss (%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a3a0
vmovss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
vmovups (%rax), %xmm0
addq $0x48, %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/unaryop_x86_avx512.cpp |
2,113,308 | ncnn::UnaryOp_x86_avx512_functor::unary_op_acos::func(float const&) const | float func(const float& x) const
{
return (float)acos(x);
} | subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%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/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,309 | ncnn::UnaryOp_x86_avx512_functor::unary_op_atan::func_pack16(float vector[16] const&) const | __m512 func_pack16(const __m512& x) const
{
//TODO avx512 optimize
float tmp[16];
_mm512_storeu_ps(tmp, x);
for (int i = 0; i < 16; i++)
tmp[i] = atan(tmp[i]);
return _mm512_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x70(%rsp)
movq %rsi, 0x68(%rsp)
leaq 0x20(%rsp), %rax
movq 0x68(%rsp), %rcx
vmovaps (%rcx), %zmm0
movq %rax, 0xe8(%rsp)
vmovaps %zmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %zmm0
movq 0xe8(%rsp), %rax
vmovups %zmm0, (%rax)
movl $0x0, 0x1c(%rsp)
cmpl $0x10, 0x1c(%rsp)
jge 0x1658409
movslq 0x1c(%rsp), %rax
vmovss 0x20(%rsp,%rax,4), %xmm0
vzeroupper
callq 0x163a3f0
movslq 0x1c(%rsp), %rax
vmovss %xmm0, 0x20(%rsp,%rax,4)
movl 0x1c(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x1c(%rsp)
jmp 0x16583d7
leaq 0x20(%rsp), %rax
movq %rax, 0x78(%rsp)
movq 0x78(%rsp), %rax
vmovups (%rax), %zmm0
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/unaryop_x86_avx512.cpp |
2,113,310 | ncnn::UnaryOp_x86_avx512_functor::unary_op_atan::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
//TODO avx optimize
float tmp[8];
_mm256_storeu_ps(tmp, x);
tmp[0] = atan(tmp[0]);
tmp[1] = atan(tmp[1]);
tmp[2] = atan(tmp[2]);
tmp[3] = atan(tmp[3]);
tmp[4] = atan(tmp[4]);
tmp[5] = atan(tmp[5]);
tmp[6] = atan(tmp[6]);
tmp[7] = atan(tmp[7]);
return _mm256_loadu_ps(tmp);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x30(%rsp)
movq %rsi, 0x28(%rsp)
movq %rsp, %rax
movq 0x28(%rsp), %rcx
vmovaps (%rcx), %ymm0
movq %rax, 0x68(%rsp)
vmovaps %ymm0, 0x40(%rsp)
vmovaps 0x40(%rsp), %ymm0
movq 0x68(%rsp), %rax
vmovups %ymm0, (%rax)
vmovss (%rsp), %xmm0
vzeroupper
callq 0x163a3f0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0xc(%rsp)
vmovss 0x10(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x10(%rsp)
vmovss 0x14(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x14(%rsp)
vmovss 0x18(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x18(%rsp)
vmovss 0x1c(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x1c(%rsp)
movq %rsp, %rax
movq %rax, 0x38(%rsp)
movq 0x38(%rsp), %rax
vmovups (%rax), %ymm0
movq %rbp, %rsp
popq %rbp
retq
nop
| ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json | O0 | ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,311 | ncnn::UnaryOp_x86_avx512_functor::unary_op_atan::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
//TODO sse optimize
float tmp[4];
_mm_storeu_ps(tmp, x);
tmp[0] = atan(tmp[0]);
tmp[1] = atan(tmp[1]);
tmp[2] = atan(tmp[2]);
tmp[3] = atan(tmp[3]);
return _mm_loadu_ps(tmp);
} | subq $0x48, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq %rsp, %rax
movq 0x10(%rsp), %rcx
vmovaps (%rcx), %xmm0
movq %rax, 0x38(%rsp)
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x20(%rsp), %xmm0
movq 0x38(%rsp), %rax
vmovups %xmm0, (%rax)
vmovss (%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x4(%rsp)
vmovss 0x8(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0x8(%rsp)
vmovss 0xc(%rsp), %xmm0
callq 0x163a3f0
vmovss %xmm0, 0xc(%rsp)
movq %rsp, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
vmovups (%rax), %xmm0
addq $0x48, %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/unaryop_x86_avx512.cpp |
2,113,312 | ncnn::UnaryOp_x86_avx512_functor::unary_op_atan::func(float const&) const | float func(const float& x) const
{
return (float)atan(x);
} | subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%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/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,313 | ncnn::UnaryOp_x86_avx512_functor::unary_op_reciprocal::func_pack16(float vector[16] const&) const | __m512 func_pack16(const __m512& x) const
{
return _mm512_div_ps(*(__m512*)_ps512_1, x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x100, %rsp # imm = 0x100
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
vmovaps 0x7c484d(%rip), %zmm1 # 0x1e1ce40
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm1, 0x80(%rsp)
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0x80(%rsp), %zmm0
vdivps 0x40(%rsp), %zmm0, %zmm0
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/unaryop_x86_avx512.cpp |
2,113,314 | ncnn::UnaryOp_x86_avx512_functor::unary_op_reciprocal::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return _mm256_div_ps(*(__m256*)_ps256_1, x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x80, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
vmovaps 0x7c4b6f(%rip), %ymm1 # 0x1e1d1c0
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
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/unaryop_x86_avx512.cpp |
2,113,315 | ncnn::UnaryOp_x86_avx512_functor::unary_op_reciprocal::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return _mm_div_ps(*(__m128*)_ps_1, x);
} | movq %rdi, -0x30(%rsp)
movq %rsi, -0x38(%rsp)
vmovaps 0x7c4cee(%rip), %xmm1 # 0x1e1d380
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 %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/unaryop_x86_avx512.cpp |
2,113,316 | ncnn::UnaryOp_x86_avx512_functor::unary_op_reciprocal::func(float const&) const | float func(const float& x) const
{
return 1.f / x;
} | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x10(%rsp), %rax
vmovss 0x7b092d(%rip), %xmm0 # 0x1e09004
vdivss (%rax), %xmm0, %xmm0
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/unaryop_x86_avx512.cpp |
2,113,317 | ncnn::UnaryOp_x86_avx512_functor::unary_op_tanh::func_pack16(float vector[16] const&) const | __m512 func_pack16(const __m512& x) const
{
return tanh_avx512(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x1500, %rsp # imm = 0x1500
movq %rdi, 0x38(%rsp)
movq %rsi, 0x30(%rsp)
movq 0x30(%rsp), %rax
vmovaps (%rax), %zmm0
vmovaps %zmm0, 0xc0(%rsp)
movl $0x3f800000, 0x27c(%rsp) # imm = 0x3F800000
vbroadcastss 0x27c(%rsp), %zmm0
vmovaps %zmm0, 0x200(%rsp)
vmovaps 0x200(%rsp), %zmm0
vmovaps %zmm0, 0x80(%rsp)
movl $0x40000000, 0x1fc(%rsp) # imm = 0x40000000
vbroadcastss 0x1fc(%rsp), %zmm0
vmovaps %zmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %zmm0
vmovaps %zmm0, 0x40(%rsp)
vmovaps 0xc0(%rsp), %zmm1
vmovaps 0x40(%rsp), %zmm0
vmovaps %zmm1, 0x140(%rsp)
vmovaps %zmm0, 0x100(%rsp)
vmovaps 0x140(%rsp), %zmm0
vmovaps 0x100(%rsp), %zmm1
vmulps %zmm1, %zmm0, %zmm0
vmovaps %zmm0, 0x380(%rsp)
movl $0x3f800000, 0x143c(%rsp) # imm = 0x3F800000
vbroadcastss 0x143c(%rsp), %zmm0
vmovaps %zmm0, 0x13c0(%rsp)
vmovaps 0x13c0(%rsp), %zmm0
vmovaps %zmm0, 0x340(%rsp)
vmovaps 0x340(%rsp), %zmm1
vpxor %xmm0, %xmm0, %xmm0
vmovaps %zmm0, 0x440(%rsp)
vmovaps 0x440(%rsp), %zmm3
vmovaps 0x380(%rsp), %zmm2
vmovaps %zmm3, 0x400(%rsp)
vmovaps %zmm2, 0x3c0(%rsp)
vmovaps 0x400(%rsp), %zmm2
vmovaps 0x3c0(%rsp), %zmm3
vsubps %zmm3, %zmm2, %zmm2
vmovaps %zmm2, 0x640(%rsp)
vmovaps %zmm0, 0x6c0(%rsp)
vmovaps 0x6c0(%rsp), %zmm0
vmovaps %zmm0, 0x600(%rsp)
vmovaps 0x7c45fb(%rip), %zmm0 # 0x1e1ce40
vmovaps %zmm0, 0x540(%rsp)
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4621(%rip), %zmm0 # 0x1e1ce80
vmovaps %zmm2, 0x840(%rsp)
vmovaps %zmm0, 0x800(%rsp)
vmovaps 0x840(%rsp), %zmm0
vmovaps 0x800(%rsp), %zmm2
vminps %zmm2, %zmm0, %zmm0
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4621(%rip), %zmm0 # 0x1e1cec0
vmovaps %zmm2, 0x8c0(%rsp)
vmovaps %zmm0, 0x880(%rsp)
vmovaps 0x8c0(%rsp), %zmm0
vmovaps 0x880(%rsp), %zmm2
vmaxps %zmm2, %zmm0, %zmm0
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x640(%rsp), %zmm3
vmovaps 0x7c4621(%rip), %zmm2 # 0x1e1cf00
vmovaps 0x7c4657(%rip), %zmm0 # 0x1e1cf40
vmovaps %zmm3, 0xe00(%rsp)
vmovaps %zmm2, 0xdc0(%rsp)
vmovaps %zmm0, 0xd80(%rsp)
vmovaps 0xe00(%rsp), %zmm2
vmovaps 0xdc0(%rsp), %zmm0
vmovaps 0xd80(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x5c0(%rsp)
vmovaps 0x5c0(%rsp), %zmm0
vrndscaleps $0x1, %zmm0, %zmm0
vmovaps %zmm0, 0x600(%rsp)
vmovaps 0x600(%rsp), %zmm2
vmovaps 0x5c0(%rsp), %zmm0
vcmpltps %zmm2, %zmm0, %k0
kmovw %k0, 0x53e(%rsp)
vmovaps 0x600(%rsp), %zmm2
movw 0x53e(%rsp), %ax
vmovaps 0x540(%rsp), %zmm0
vmovaps %zmm2, 0xf00(%rsp)
movw %ax, 0xefe(%rsp)
vmovaps %zmm2, 0xe80(%rsp)
vmovaps %zmm0, 0xe40(%rsp)
vmovaps 0xe80(%rsp), %zmm2
vmovaps 0xe40(%rsp), %zmm0
kmovw 0xefe(%rsp), %k1
vmovaps %zmm2, 0xf80(%rsp)
vmovaps %zmm0, 0xf40(%rsp)
vmovaps 0xf80(%rsp), %zmm2
vmovaps 0xf40(%rsp), %zmm3
vmovaps 0xf00(%rsp), %zmm0
vsubps %zmm3, %zmm2, %zmm0 {%k1}
vmovaps %zmm0, 0x5c0(%rsp)
vmovaps 0x5c0(%rsp), %zmm3
vmovaps 0x7c4589(%rip), %zmm2 # 0x1e1cf80
vmovaps 0x640(%rsp), %zmm0
vmovaps %zmm3, 0x1100(%rsp)
vmovaps %zmm2, 0x10c0(%rsp)
vmovaps %zmm0, 0x1080(%rsp)
vmovaps 0x1100(%rsp), %zmm2
vmovaps 0x10c0(%rsp), %zmm0
vmovaps 0x1080(%rsp), %zmm3
vfnmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = -(zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x5c0(%rsp), %zmm3
vmovaps 0x7c4571(%rip), %zmm2 # 0x1e1cfc0
vmovaps 0x640(%rsp), %zmm0
vmovaps %zmm3, 0x1040(%rsp)
vmovaps %zmm2, 0x1000(%rsp)
vmovaps %zmm0, 0xfc0(%rsp)
vmovaps 0x1040(%rsp), %zmm2
vmovaps 0x1000(%rsp), %zmm0
vmovaps 0xfc0(%rsp), %zmm3
vfnmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = -(zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x640(%rsp)
vmovaps 0x640(%rsp), %zmm0
vmovaps %zmm0, 0x7c0(%rsp)
vmovaps %zmm0, 0x780(%rsp)
vmovaps 0x7c0(%rsp), %zmm0
vmovaps 0x780(%rsp), %zmm2
vmulps %zmm2, %zmm0, %zmm0
vmovaps %zmm0, 0x600(%rsp)
vmovaps 0x7c452b(%rip), %zmm0 # 0x1e1d000
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4549(%rip), %zmm0 # 0x1e1d040
vmovaps %zmm3, 0xd40(%rsp)
vmovaps %zmm2, 0xd00(%rsp)
vmovaps %zmm0, 0xcc0(%rsp)
vmovaps 0xd40(%rsp), %zmm2
vmovaps 0xd00(%rsp), %zmm0
vmovaps 0xcc0(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4531(%rip), %zmm0 # 0x1e1d080
vmovaps %zmm3, 0xc80(%rsp)
vmovaps %zmm2, 0xc40(%rsp)
vmovaps %zmm0, 0xc00(%rsp)
vmovaps 0xc80(%rsp), %zmm2
vmovaps 0xc40(%rsp), %zmm0
vmovaps 0xc00(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4519(%rip), %zmm0 # 0x1e1d0c0
vmovaps %zmm3, 0xbc0(%rsp)
vmovaps %zmm2, 0xb80(%rsp)
vmovaps %zmm0, 0xb40(%rsp)
vmovaps 0xbc0(%rsp), %zmm2
vmovaps 0xb80(%rsp), %zmm0
vmovaps 0xb40(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c4501(%rip), %zmm0 # 0x1e1d100
vmovaps %zmm3, 0xb00(%rsp)
vmovaps %zmm2, 0xac0(%rsp)
vmovaps %zmm0, 0xa80(%rsp)
vmovaps 0xb00(%rsp), %zmm2
vmovaps 0xac0(%rsp), %zmm0
vmovaps 0xa80(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x640(%rsp), %zmm2
vmovaps 0x7c44e9(%rip), %zmm0 # 0x1e1d140
vmovaps %zmm3, 0xa40(%rsp)
vmovaps %zmm2, 0xa00(%rsp)
vmovaps %zmm0, 0x9c0(%rsp)
vmovaps 0xa40(%rsp), %zmm2
vmovaps 0xa00(%rsp), %zmm0
vmovaps 0x9c0(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm3
vmovaps 0x600(%rsp), %zmm2
vmovaps 0x640(%rsp), %zmm0
vmovaps %zmm3, 0x980(%rsp)
vmovaps %zmm2, 0x940(%rsp)
vmovaps %zmm0, 0x900(%rsp)
vmovaps 0x980(%rsp), %zmm2
vmovaps 0x940(%rsp), %zmm0
vmovaps 0x900(%rsp), %zmm3
vfmadd213ps %zmm3, %zmm2, %zmm0 # zmm0 = (zmm2 * zmm0) + zmm3
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm2
vmovaps 0x540(%rsp), %zmm0
vmovaps %zmm2, 0x1180(%rsp)
vmovaps %zmm0, 0x1140(%rsp)
vmovaps 0x1180(%rsp), %zmm0
vmovaps 0x1140(%rsp), %zmm2
vaddps %zmm2, %zmm0, %zmm0
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x5c0(%rsp), %zmm0
vmovaps %zmm0, 0x1240(%rsp)
vpxor %xmm2, %xmm2, %xmm2
vcvttps2dq 0x1240(%rsp), %zmm0
vmovdqa64 %zmm2, 0x1380(%rsp)
vmovdqa64 %zmm0, 0x580(%rsp)
vmovdqa64 0x580(%rsp), %zmm2
vmovdqa64 0x7c4419(%rip), %zmm0 # 0x1e1d180
vmovdqa64 %zmm2, 0x12c0(%rsp)
vmovdqa64 %zmm0, 0x1280(%rsp)
vmovdqa64 0x12c0(%rsp), %zmm0
vmovdqa64 0x1280(%rsp), %zmm2
vpaddd %zmm2, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x580(%rsp)
vmovdqa64 0x580(%rsp), %zmm0
vmovdqa64 %zmm0, 0x1340(%rsp)
movl $0x17, 0x133c(%rsp)
vmovdqa64 0x1340(%rsp), %zmm0
vmovd 0x133c(%rsp), %xmm2
vpslld %xmm2, %zmm0, %zmm0
vmovdqa64 %zmm0, 0x580(%rsp)
vmovdqa64 0x580(%rsp), %zmm0
vmovdqa64 %zmm0, 0x680(%rsp)
vmovdqa64 0x680(%rsp), %zmm0
vmovdqa64 %zmm0, 0x480(%rsp)
vmovaps 0x4c0(%rsp), %zmm2
vmovaps 0x480(%rsp), %zmm0
vmovaps %zmm2, 0x740(%rsp)
vmovaps %zmm0, 0x700(%rsp)
vmovaps 0x740(%rsp), %zmm0
vmovaps 0x700(%rsp), %zmm2
vmulps %zmm2, %zmm0, %zmm0
vmovaps %zmm0, 0x4c0(%rsp)
vmovaps 0x4c0(%rsp), %zmm0
vmovaps %zmm1, 0x1200(%rsp)
vmovaps %zmm0, 0x11c0(%rsp)
vmovaps 0x1200(%rsp), %zmm0
vmovaps 0x11c0(%rsp), %zmm2
vaddps %zmm2, %zmm0, %zmm0
vmovaps %zmm1, 0x1480(%rsp)
vmovaps %zmm0, 0x1440(%rsp)
vmovaps 0x1480(%rsp), %zmm4
vmovaps 0x1440(%rsp), %zmm3
vrcp14ps %zmm3, %zmm0
vmulps %zmm0, %zmm4, %zmm1
vmovaps %zmm1, %zmm2
vfmsub213ps %zmm4, %zmm3, %zmm2 # zmm2 = (zmm3 * zmm2) - zmm4
vfnmadd213ps %zmm1, %zmm0, %zmm2 # zmm2 = -(zmm0 * zmm2) + zmm1
vmovaps 0x40(%rsp), %zmm1
vmovaps 0x80(%rsp), %zmm0
vmovaps %zmm2, 0x300(%rsp)
vmovaps %zmm1, 0x2c0(%rsp)
vmovaps %zmm0, 0x280(%rsp)
vmovaps 0x300(%rsp), %zmm1
vmovaps 0x2c0(%rsp), %zmm0
vmovaps 0x280(%rsp), %zmm2
vpbroadcastd 0x7b4fd5(%rip), %zmm3 # 0x1e0deb8
vpxord %zmm3, %zmm2, %zmm2
vfmadd213ps %zmm2, %zmm1, %zmm0 # zmm0 = (zmm1 * zmm0) + zmm2
movq %rbp, %rsp
popq %rbp
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/unaryop_x86_avx512.cpp |
2,113,318 | ncnn::UnaryOp_x86_avx512_functor::unary_op_tanh::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return tanh_avx(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xe40, %rsp # imm = 0xE40
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x60(%rsp)
movl $0x3f800000, 0xdc(%rsp) # imm = 0x3F800000
vmovss 0xdc(%rsp), %xmm0
vmovss %xmm0, 0x11c(%rsp)
vmovss %xmm0, 0x118(%rsp)
vmovss %xmm0, 0x114(%rsp)
vmovss %xmm0, 0x110(%rsp)
vmovss %xmm0, 0x10c(%rsp)
vmovss %xmm0, 0x108(%rsp)
vmovss %xmm0, 0x104(%rsp)
vmovss %xmm0, 0x100(%rsp)
vmovss 0x104(%rsp), %xmm1
vmovss 0x100(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x108(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x10c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1,2],xmm1[0]
vmovss 0x114(%rsp), %xmm2
vmovss 0x110(%rsp), %xmm1
vinsertps $0x10, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
vmovss 0x118(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1],xmm2[0],xmm1[3]
vmovss 0x11c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1,2],xmm2[0]
vmovaps %xmm1, 0xf0(%rsp)
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %ymm0
vmovaps %ymm0, 0x40(%rsp)
movl $0x40000000, 0xd8(%rsp) # imm = 0x40000000
vmovss 0xd8(%rsp), %xmm0
vmovss %xmm0, 0x15c(%rsp)
vmovss %xmm0, 0x158(%rsp)
vmovss %xmm0, 0x154(%rsp)
vmovss %xmm0, 0x150(%rsp)
vmovss %xmm0, 0x14c(%rsp)
vmovss %xmm0, 0x148(%rsp)
vmovss %xmm0, 0x144(%rsp)
vmovss %xmm0, 0x140(%rsp)
vmovss 0x144(%rsp), %xmm1
vmovss 0x140(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0x148(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0x14c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1,2],xmm1[0]
vmovss 0x154(%rsp), %xmm2
vmovss 0x150(%rsp), %xmm1
vinsertps $0x10, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
vmovss 0x158(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1],xmm2[0],xmm1[3]
vmovss 0x15c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1,2],xmm2[0]
vmovaps %xmm1, 0x130(%rsp)
vmovaps %xmm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
vmovaps %ymm0, 0x20(%rsp)
vmovaps 0x60(%rsp), %ymm1
vmovaps 0x20(%rsp), %ymm0
vmovaps %ymm1, 0xa0(%rsp)
vmovaps %ymm0, 0x80(%rsp)
vmovaps 0xa0(%rsp), %ymm0
vmovaps 0x80(%rsp), %ymm1
vmulps %ymm1, %ymm0, %ymm0
vmovaps %ymm0, 0x1e0(%rsp)
movl $0x3f800000, 0xd9c(%rsp) # imm = 0x3F800000
vmovss 0xd9c(%rsp), %xmm0
vmovss %xmm0, 0xe2c(%rsp)
vmovss %xmm0, 0xe28(%rsp)
vmovss %xmm0, 0xe24(%rsp)
vmovss %xmm0, 0xe20(%rsp)
vmovss %xmm0, 0xe1c(%rsp)
vmovss %xmm0, 0xe18(%rsp)
vmovss %xmm0, 0xe14(%rsp)
vmovss %xmm0, 0xe10(%rsp)
vmovss 0xe14(%rsp), %xmm1
vmovss 0xe10(%rsp), %xmm0
vinsertps $0x10, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0],xmm1[0],xmm0[2,3]
vmovss 0xe18(%rsp), %xmm1
vinsertps $0x20, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1],xmm1[0],xmm0[3]
vmovss 0xe1c(%rsp), %xmm1
vinsertps $0x30, %xmm1, %xmm0, %xmm0 # xmm0 = xmm0[0,1,2],xmm1[0]
vmovss 0xe24(%rsp), %xmm2
vmovss 0xe20(%rsp), %xmm1
vinsertps $0x10, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0],xmm2[0],xmm1[2,3]
vmovss 0xe28(%rsp), %xmm2
vinsertps $0x20, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1],xmm2[0],xmm1[3]
vmovss 0xe2c(%rsp), %xmm2
vinsertps $0x30, %xmm2, %xmm1, %xmm1 # xmm1 = xmm1[0,1,2],xmm2[0]
vmovaps %xmm1, 0xdf0(%rsp)
vmovaps %xmm0, 0xde0(%rsp)
vmovaps 0xde0(%rsp), %ymm0
vmovaps %ymm0, 0x1c0(%rsp)
vmovaps 0x1c0(%rsp), %ymm1
vpxor %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %ymm3
vmovaps 0x1e0(%rsp), %ymm2
vmovaps %ymm3, 0x220(%rsp)
vmovaps %ymm2, 0x200(%rsp)
vmovaps 0x220(%rsp), %ymm2
vmovaps 0x200(%rsp), %ymm3
vsubps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0x340(%rsp)
vmovaps %ymm0, 0x3e0(%rsp)
vmovaps 0x3e0(%rsp), %ymm0
vmovaps %ymm0, 0x320(%rsp)
vbroadcastss 0x7afd64(%rip), %ymm0 # 0x1e09004
vmovaps %ymm0, 0x2c0(%rsp)
vmovaps 0x340(%rsp), %ymm0
vmovaps %ymm0, 0x4a0(%rsp)
vbroadcastss 0x7b2afc(%rip), %ymm0 # 0x1e0bdc0
vmovaps %ymm0, 0x480(%rsp)
vmovaps 0x4a0(%rsp), %ymm0
vmovaps 0x480(%rsp), %ymm2
vminps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x340(%rsp)
vmovaps 0x340(%rsp), %ymm0
vmovaps %ymm0, 0x380(%rsp)
vbroadcastss 0x7b2abd(%rip), %ymm0 # 0x1e0bdc4
vmovaps %ymm0, 0x360(%rsp)
vmovaps 0x380(%rsp), %ymm0
vmovaps 0x360(%rsp), %ymm2
vmaxps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x340(%rsp)
vmovaps 0x340(%rsp), %ymm0
vmovaps %ymm0, 0x740(%rsp)
vbroadcastss 0x7b2a7e(%rip), %ymm0 # 0x1e0bdc8
vmovaps %ymm0, 0x720(%rsp)
vbroadcastss 0x7afcc0(%rip), %ymm0 # 0x1e0901c
vmovaps %ymm0, 0x700(%rsp)
vmovaps 0x740(%rsp), %ymm4
vmovaps 0x720(%rsp), %ymm3
vmovaps 0x700(%rsp), %ymm2
vmovaps %ymm4, 0x9e0(%rsp)
vmovaps %ymm3, 0x9c0(%rsp)
vmovaps %ymm2, 0x9a0(%rsp)
vmovaps 0x9e0(%rsp), %ymm3
vmovaps 0x9c0(%rsp), %ymm2
vmovaps 0x9a0(%rsp), %ymm4
vfmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm2
vroundps $0x1, %ymm2, %ymm2
vmovaps %ymm2, 0x320(%rsp)
vmovaps 0x320(%rsp), %ymm3
vmovaps 0x300(%rsp), %ymm2
vcmpltps %ymm3, %ymm2, %ymm2
vmovdqa %ymm2, 0x2a0(%rsp)
vmovaps 0x2a0(%rsp), %ymm3
vmovaps 0x2c0(%rsp), %ymm2
vmovaps %ymm3, 0x780(%rsp)
vmovaps %ymm2, 0x760(%rsp)
vmovdqa 0x780(%rsp), %ymm2
vmovdqa 0x760(%rsp), %ymm3
vpand %ymm3, %ymm2, %ymm2
vmovdqa %ymm2, 0x2a0(%rsp)
vmovaps 0x320(%rsp), %ymm3
vmovaps 0x2a0(%rsp), %ymm2
vmovaps %ymm3, 0x3c0(%rsp)
vmovaps %ymm2, 0x3a0(%rsp)
vmovaps 0x3c0(%rsp), %ymm2
vmovaps 0x3a0(%rsp), %ymm3
vsubps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0x300(%rsp)
vmovaps 0x300(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x840(%rsp)
vbroadcastss 0x7b2926(%rip), %ymm3 # 0x1e0bdcc
vmovaps %ymm3, 0x820(%rsp)
vmovaps %ymm2, 0x800(%rsp)
vmovaps 0x840(%rsp), %ymm4
vmovaps 0x820(%rsp), %ymm3
vmovaps 0x800(%rsp), %ymm2
vmovaps %ymm4, 0xc80(%rsp)
vmovaps %ymm3, 0xc60(%rsp)
vmovaps %ymm2, 0xc40(%rsp)
vmovaps 0xc80(%rsp), %ymm3
vmovaps 0xc60(%rsp), %ymm2
vmovaps 0xc40(%rsp), %ymm4
vfnmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = -(ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x340(%rsp)
vmovaps 0x300(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x7e0(%rsp)
vbroadcastss 0x7b2895(%rip), %ymm3 # 0x1e0bdd0
vmovaps %ymm3, 0x7c0(%rsp)
vmovaps %ymm2, 0x7a0(%rsp)
vmovaps 0x7e0(%rsp), %ymm4
vmovaps 0x7c0(%rsp), %ymm3
vmovaps 0x7a0(%rsp), %ymm2
vmovaps %ymm4, 0xce0(%rsp)
vmovaps %ymm3, 0xcc0(%rsp)
vmovaps %ymm2, 0xca0(%rsp)
vmovaps 0xce0(%rsp), %ymm3
vmovaps 0xcc0(%rsp), %ymm2
vmovaps 0xca0(%rsp), %ymm4
vfnmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = -(ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x340(%rsp)
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm2, 0x460(%rsp)
vmovaps %ymm2, 0x440(%rsp)
vmovaps 0x460(%rsp), %ymm2
vmovaps 0x440(%rsp), %ymm3
vmulps %ymm3, %ymm2, %ymm2
vmovaps %ymm2, 0x320(%rsp)
vbroadcastss 0x7b27e5(%rip), %ymm2 # 0x1e0bdd4
vmovaps %ymm2, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x6e0(%rsp)
vmovaps %ymm2, 0x6c0(%rsp)
vbroadcastss 0x7b27b3(%rip), %ymm2 # 0x1e0bdd8
vmovaps %ymm2, 0x6a0(%rsp)
vmovaps 0x6e0(%rsp), %ymm4
vmovaps 0x6c0(%rsp), %ymm3
vmovaps 0x6a0(%rsp), %ymm2
vmovaps %ymm4, 0xa40(%rsp)
vmovaps %ymm3, 0xa20(%rsp)
vmovaps %ymm2, 0xa00(%rsp)
vmovaps 0xa40(%rsp), %ymm3
vmovaps 0xa20(%rsp), %ymm2
vmovaps 0xa00(%rsp), %ymm4
vfmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x680(%rsp)
vmovaps %ymm2, 0x660(%rsp)
vbroadcastss 0x7b2722(%rip), %ymm2 # 0x1e0bddc
vmovaps %ymm2, 0x640(%rsp)
vmovaps 0x680(%rsp), %ymm4
vmovaps 0x660(%rsp), %ymm3
vmovaps 0x640(%rsp), %ymm2
vmovaps %ymm4, 0xaa0(%rsp)
vmovaps %ymm3, 0xa80(%rsp)
vmovaps %ymm2, 0xa60(%rsp)
vmovaps 0xaa0(%rsp), %ymm3
vmovaps 0xa80(%rsp), %ymm2
vmovaps 0xa60(%rsp), %ymm4
vfmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x620(%rsp)
vmovaps %ymm2, 0x600(%rsp)
vbroadcastss 0x7b2691(%rip), %ymm2 # 0x1e0bde0
vmovaps %ymm2, 0x5e0(%rsp)
vmovaps 0x620(%rsp), %ymm4
vmovaps 0x600(%rsp), %ymm3
vmovaps 0x5e0(%rsp), %ymm2
vmovaps %ymm4, 0xb00(%rsp)
vmovaps %ymm3, 0xae0(%rsp)
vmovaps %ymm2, 0xac0(%rsp)
vmovaps 0xb00(%rsp), %ymm3
vmovaps 0xae0(%rsp), %ymm2
vmovaps 0xac0(%rsp), %ymm4
vfmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x5c0(%rsp)
vmovaps %ymm2, 0x5a0(%rsp)
vbroadcastss 0x7b2600(%rip), %ymm2 # 0x1e0bde4
vmovaps %ymm2, 0x580(%rsp)
vmovaps 0x5c0(%rsp), %ymm4
vmovaps 0x5a0(%rsp), %ymm3
vmovaps 0x580(%rsp), %ymm2
vmovaps %ymm4, 0xb60(%rsp)
vmovaps %ymm3, 0xb40(%rsp)
vmovaps %ymm2, 0xb20(%rsp)
vmovaps 0xb60(%rsp), %ymm3
vmovaps 0xb40(%rsp), %ymm2
vmovaps 0xb20(%rsp), %ymm4
vfmadd213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) + ymm4
vmovaps %ymm2, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x340(%rsp), %ymm2
vmovaps %ymm3, 0x560(%rsp)
vmovaps %ymm2, 0x540(%rsp)
vmovaps %ymm0, 0x520(%rsp)
vmovaps 0x560(%rsp), %ymm3
vmovaps 0x540(%rsp), %ymm2
vmovaps 0x520(%rsp), %ymm0
vmovaps %ymm3, 0xbc0(%rsp)
vmovaps %ymm2, 0xba0(%rsp)
vmovaps %ymm0, 0xb80(%rsp)
vmovaps 0xbc0(%rsp), %ymm2
vmovaps 0xba0(%rsp), %ymm0
vmovaps 0xb80(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm3
vmovaps 0x320(%rsp), %ymm2
vmovaps 0x340(%rsp), %ymm0
vmovaps %ymm3, 0x500(%rsp)
vmovaps %ymm2, 0x4e0(%rsp)
vmovaps %ymm0, 0x4c0(%rsp)
vmovaps 0x500(%rsp), %ymm3
vmovaps 0x4e0(%rsp), %ymm2
vmovaps 0x4c0(%rsp), %ymm0
vmovaps %ymm3, 0xc20(%rsp)
vmovaps %ymm2, 0xc00(%rsp)
vmovaps %ymm0, 0xbe0(%rsp)
vmovaps 0xc20(%rsp), %ymm2
vmovaps 0xc00(%rsp), %ymm0
vmovaps 0xbe0(%rsp), %ymm3
vfmadd213ps %ymm3, %ymm2, %ymm0 # ymm0 = (ymm2 * ymm0) + ymm3
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm2
vmovaps 0x2c0(%rsp), %ymm0
vmovaps %ymm2, 0x880(%rsp)
vmovaps %ymm0, 0x860(%rsp)
vmovaps 0x880(%rsp), %ymm0
vmovaps 0x860(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x300(%rsp), %ymm0
vmovaps %ymm0, 0x8e0(%rsp)
vcvttps2dq 0x8e0(%rsp), %ymm0
vmovdqa %ymm0, 0x2e0(%rsp)
vmovdqa 0x2e0(%rsp), %ymm0
vmovdqa %ymm0, 0x920(%rsp)
vpbroadcastq 0x7b2429(%rip), %ymm0 # 0x1e0be18
vmovdqa %ymm0, 0x900(%rsp)
vmovdqa 0x920(%rsp), %ymm2
vmovdqa 0x900(%rsp), %ymm0
vmovdqa %ymm2, 0xd20(%rsp)
vmovdqa %ymm0, 0xd00(%rsp)
vmovdqa 0xd20(%rsp), %ymm0
vmovdqa 0xd00(%rsp), %ymm2
vpaddd %ymm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2e0(%rsp)
vmovdqa 0x2e0(%rsp), %ymm0
vmovdqa %ymm0, 0x960(%rsp)
movl $0x17, 0x95c(%rsp)
vmovdqa 0x960(%rsp), %ymm0
movl 0x95c(%rsp), %eax
vmovdqa %ymm0, 0xd60(%rsp)
movl %eax, 0xd5c(%rsp)
vmovdqa 0xd60(%rsp), %ymm0
vmovd 0xd5c(%rsp), %xmm2
vpslld %xmm2, %ymm0, %ymm0
vmovdqa %ymm0, 0x2e0(%rsp)
vmovdqa 0x2e0(%rsp), %ymm0
vmovdqa %ymm0, 0x980(%rsp)
vmovdqa 0x980(%rsp), %ymm0
vmovdqa %ymm0, 0x260(%rsp)
vmovaps 0x280(%rsp), %ymm2
vmovaps 0x260(%rsp), %ymm0
vmovaps %ymm2, 0x420(%rsp)
vmovaps %ymm0, 0x400(%rsp)
vmovaps 0x420(%rsp), %ymm0
vmovaps 0x400(%rsp), %ymm2
vmulps %ymm2, %ymm0, %ymm0
vmovaps %ymm0, 0x280(%rsp)
vmovaps 0x280(%rsp), %ymm0
vmovaps %ymm1, 0x8c0(%rsp)
vmovaps %ymm0, 0x8a0(%rsp)
vmovaps 0x8c0(%rsp), %ymm0
vmovaps 0x8a0(%rsp), %ymm2
vaddps %ymm2, %ymm0, %ymm0
vmovaps %ymm1, 0xdc0(%rsp)
vmovaps %ymm0, 0xda0(%rsp)
vmovaps 0xdc0(%rsp), %ymm4
vmovaps 0xda0(%rsp), %ymm3
vrcpps %ymm3, %ymm0
vmulps %ymm0, %ymm4, %ymm1
vmovaps %ymm1, %ymm2
vfmsub213ps %ymm4, %ymm3, %ymm2 # ymm2 = (ymm3 * ymm2) - ymm4
vfnmadd213ps %ymm1, %ymm0, %ymm2 # ymm2 = -(ymm0 * ymm2) + ymm1
vmovaps 0x20(%rsp), %ymm1
vmovaps 0x40(%rsp), %ymm0
vmovaps %ymm2, 0x1a0(%rsp)
vmovaps %ymm1, 0x180(%rsp)
vmovaps %ymm0, 0x160(%rsp)
vmovaps 0x1a0(%rsp), %ymm1
vmovaps 0x180(%rsp), %ymm0
vmovaps 0x160(%rsp), %ymm2
vpbroadcastd 0x7b4304(%rip), %ymm3 # 0x1e0deb8
vpxor %ymm3, %ymm2, %ymm2
vfmadd213ps %ymm2, %ymm1, %ymm0 # ymm0 = (ymm1 * ymm0) + ymm2
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/unaryop_x86_avx512.cpp |
2,113,319 | ncnn::UnaryOp_x86_avx512_functor::unary_op_tanh::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return tanh_sse(x);
} | subq $0x668, %rsp # imm = 0x668
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x50(%rsp)
movl $0x3f800000, -0x4(%rsp) # imm = 0x3F800000
vbroadcastss -0x4(%rsp), %xmm0
vmovaps %xmm0, -0x20(%rsp)
vmovaps -0x20(%rsp), %xmm0
vmovaps %xmm0, -0x60(%rsp)
movl $0x40000000, -0x24(%rsp) # imm = 0x40000000
vbroadcastss -0x24(%rsp), %xmm0
vmovaps %xmm0, -0x40(%rsp)
vmovaps -0x40(%rsp), %xmm0
vmovaps %xmm0, -0x70(%rsp)
vmovaps -0x50(%rsp), %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x50(%rsp)
vmovaps %xmm0, 0x40(%rsp)
vmovaps 0x50(%rsp), %xmm0
vmovaps 0x40(%rsp), %xmm1
vmulps %xmm1, %xmm0, %xmm0
vmovaps %xmm0, 0x70(%rsp)
movl $0x3f800000, 0x9c(%rsp) # imm = 0x3F800000
vbroadcastss 0x9c(%rsp), %xmm0
vmovaps %xmm0, 0x80(%rsp)
vmovaps 0x80(%rsp), %xmm0
vmovaps %xmm0, 0x60(%rsp)
vmovaps 0x60(%rsp), %xmm2
vmovaps %xmm2, %xmm1
vpxor %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0xc0(%rsp)
vmovaps 0xc0(%rsp), %xmm4
vmovaps 0x70(%rsp), %xmm3
vmovaps %xmm4, 0xb0(%rsp)
vmovaps %xmm3, 0xa0(%rsp)
vmovaps 0xb0(%rsp), %xmm3
vmovaps 0xa0(%rsp), %xmm4
vsubps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x140(%rsp)
vmovaps %xmm0, 0x170(%rsp)
vmovaps 0x170(%rsp), %xmm0
vmovaps %xmm0, 0x130(%rsp)
vbroadcastss 0x7af2fc(%rip), %xmm0 # 0x1e09004
vmovaps %xmm0, 0x100(%rsp)
vmovaps 0x140(%rsp), %xmm0
vmovaps %xmm0, 0x1f0(%rsp)
vbroadcastss 0x7b2094(%rip), %xmm0 # 0x1e0bdc0
vmovaps %xmm0, 0x1e0(%rsp)
vmovaps 0x1f0(%rsp), %xmm0
vmovaps 0x1e0(%rsp), %xmm3
vminps %xmm3, %xmm0, %xmm0
vmovaps %xmm0, 0x140(%rsp)
vmovaps 0x140(%rsp), %xmm0
vmovaps %xmm0, 0x210(%rsp)
vbroadcastss 0x7b2055(%rip), %xmm0 # 0x1e0bdc4
vmovaps %xmm0, 0x200(%rsp)
vmovaps 0x210(%rsp), %xmm0
vmovaps 0x200(%rsp), %xmm3
vmaxps %xmm3, %xmm0, %xmm0
vmovaps %xmm0, 0x140(%rsp)
vmovaps 0x140(%rsp), %xmm0
vmovaps %xmm0, 0x1d0(%rsp)
vbroadcastss 0x7b2016(%rip), %xmm0 # 0x1e0bdc8
vmovaps %xmm0, 0x1c0(%rsp)
vmovaps 0x1d0(%rsp), %xmm0
vmovaps 0x1c0(%rsp), %xmm3
vmulps %xmm3, %xmm0, %xmm0
vmovaps %xmm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %xmm0
vmovaps %xmm0, 0x250(%rsp)
vbroadcastss 0x7af227(%rip), %xmm0 # 0x1e0901c
vmovaps %xmm0, 0x240(%rsp)
vmovaps 0x250(%rsp), %xmm3
vmovaps 0x240(%rsp), %xmm4
vaddps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x120(%rsp)
vmovaps 0x120(%rsp), %xmm3
vmovaps %xmm3, 0x290(%rsp)
vcvttps2dq 0x290(%rsp), %xmm3
vmovdqa %xmm3, 0x110(%rsp)
vmovdqa 0x110(%rsp), %xmm3
vmovdqa %xmm3, 0x2a0(%rsp)
vcvtdq2ps 0x2a0(%rsp), %xmm3
vmovaps %xmm3, 0x130(%rsp)
vmovaps 0x130(%rsp), %xmm4
vmovaps 0x120(%rsp), %xmm3
vmovaps %xmm4, 0x2c0(%rsp)
vmovaps %xmm3, 0x2b0(%rsp)
vmovaps 0x2b0(%rsp), %xmm3
vmovaps 0x2c0(%rsp), %xmm4
vcmpltps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0xf0(%rsp)
vmovaps 0xf0(%rsp), %xmm4
vmovaps 0x100(%rsp), %xmm3
vmovaps %xmm4, 0x2e0(%rsp)
vmovaps %xmm3, 0x2d0(%rsp)
vmovdqa 0x2e0(%rsp), %xmm3
vmovdqa 0x2d0(%rsp), %xmm4
vpand %xmm4, %xmm3, %xmm3
vmovdqa %xmm3, 0xf0(%rsp)
vmovaps 0x130(%rsp), %xmm4
vmovaps 0xf0(%rsp), %xmm3
vmovaps %xmm4, 0x160(%rsp)
vmovaps %xmm3, 0x150(%rsp)
vmovaps 0x160(%rsp), %xmm3
vmovaps 0x150(%rsp), %xmm4
vsubps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x120(%rsp)
vmovaps 0x120(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x340(%rsp)
vbroadcastss 0x7b1e79(%rip), %xmm4 # 0x1e0bdcc
vmovaps %xmm4, 0x330(%rsp)
vmovaps %xmm3, 0x320(%rsp)
vmovaps 0x340(%rsp), %xmm5
vmovaps 0x330(%rsp), %xmm4
vmovaps 0x320(%rsp), %xmm3
vmovaps %xmm5, 0x4e0(%rsp)
vmovaps %xmm4, 0x4d0(%rsp)
vmovaps %xmm3, 0x4c0(%rsp)
vmovaps 0x4e0(%rsp), %xmm4
vmovaps 0x4d0(%rsp), %xmm3
vmovaps 0x4c0(%rsp), %xmm5
vfnmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = -(xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0x140(%rsp)
vmovaps 0x120(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x310(%rsp)
vbroadcastss 0x7b1de8(%rip), %xmm4 # 0x1e0bdd0
vmovaps %xmm4, 0x300(%rsp)
vmovaps %xmm3, 0x2f0(%rsp)
vmovaps 0x310(%rsp), %xmm5
vmovaps 0x300(%rsp), %xmm4
vmovaps 0x2f0(%rsp), %xmm3
vmovaps %xmm5, 0x510(%rsp)
vmovaps %xmm4, 0x500(%rsp)
vmovaps %xmm3, 0x4f0(%rsp)
vmovaps 0x510(%rsp), %xmm4
vmovaps 0x500(%rsp), %xmm3
vmovaps 0x4f0(%rsp), %xmm5
vfnmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = -(xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0x140(%rsp)
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm3, 0x1b0(%rsp)
vmovaps %xmm3, 0x1a0(%rsp)
vmovaps 0x1b0(%rsp), %xmm3
vmovaps 0x1a0(%rsp), %xmm4
vmulps %xmm4, %xmm3, %xmm3
vmovaps %xmm3, 0x130(%rsp)
vbroadcastss 0x7b1d38(%rip), %xmm3 # 0x1e0bdd4
vmovaps %xmm3, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x460(%rsp)
vmovaps %xmm3, 0x450(%rsp)
vbroadcastss 0x7b1d06(%rip), %xmm3 # 0x1e0bdd8
vmovaps %xmm3, 0x440(%rsp)
vmovaps 0x460(%rsp), %xmm5
vmovaps 0x450(%rsp), %xmm4
vmovaps 0x440(%rsp), %xmm3
vmovaps %xmm5, 0x540(%rsp)
vmovaps %xmm4, 0x530(%rsp)
vmovaps %xmm3, 0x520(%rsp)
vmovaps 0x540(%rsp), %xmm4
vmovaps 0x530(%rsp), %xmm3
vmovaps 0x520(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x430(%rsp)
vmovaps %xmm3, 0x420(%rsp)
vbroadcastss 0x7b1c75(%rip), %xmm3 # 0x1e0bddc
vmovaps %xmm3, 0x410(%rsp)
vmovaps 0x430(%rsp), %xmm5
vmovaps 0x420(%rsp), %xmm4
vmovaps 0x410(%rsp), %xmm3
vmovaps %xmm5, 0x570(%rsp)
vmovaps %xmm4, 0x560(%rsp)
vmovaps %xmm3, 0x550(%rsp)
vmovaps 0x570(%rsp), %xmm4
vmovaps 0x560(%rsp), %xmm3
vmovaps 0x550(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x400(%rsp)
vmovaps %xmm3, 0x3f0(%rsp)
vbroadcastss 0x7b1be4(%rip), %xmm3 # 0x1e0bde0
vmovaps %xmm3, 0x3e0(%rsp)
vmovaps 0x400(%rsp), %xmm5
vmovaps 0x3f0(%rsp), %xmm4
vmovaps 0x3e0(%rsp), %xmm3
vmovaps %xmm5, 0x5a0(%rsp)
vmovaps %xmm4, 0x590(%rsp)
vmovaps %xmm3, 0x580(%rsp)
vmovaps 0x5a0(%rsp), %xmm4
vmovaps 0x590(%rsp), %xmm3
vmovaps 0x580(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x3d0(%rsp)
vmovaps %xmm3, 0x3c0(%rsp)
vbroadcastss 0x7b1b53(%rip), %xmm3 # 0x1e0bde4
vmovaps %xmm3, 0x3b0(%rsp)
vmovaps 0x3d0(%rsp), %xmm5
vmovaps 0x3c0(%rsp), %xmm4
vmovaps 0x3b0(%rsp), %xmm3
vmovaps %xmm5, 0x5d0(%rsp)
vmovaps %xmm4, 0x5c0(%rsp)
vmovaps %xmm3, 0x5b0(%rsp)
vmovaps 0x5d0(%rsp), %xmm4
vmovaps 0x5c0(%rsp), %xmm3
vmovaps 0x5b0(%rsp), %xmm5
vfmadd213ps %xmm5, %xmm4, %xmm3 # xmm3 = (xmm4 * xmm3) + xmm5
vmovaps %xmm3, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x140(%rsp), %xmm3
vmovaps %xmm4, 0x3a0(%rsp)
vmovaps %xmm3, 0x390(%rsp)
vmovaps %xmm0, 0x380(%rsp)
vmovaps 0x3a0(%rsp), %xmm4
vmovaps 0x390(%rsp), %xmm3
vmovaps 0x380(%rsp), %xmm0
vmovaps %xmm4, 0x600(%rsp)
vmovaps %xmm3, 0x5f0(%rsp)
vmovaps %xmm0, 0x5e0(%rsp)
vmovaps 0x600(%rsp), %xmm3
vmovaps 0x5f0(%rsp), %xmm0
vmovaps 0x5e0(%rsp), %xmm4
vfmadd213ps %xmm4, %xmm3, %xmm0 # xmm0 = (xmm3 * xmm0) + xmm4
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm4
vmovaps 0x130(%rsp), %xmm3
vmovaps 0x140(%rsp), %xmm0
vmovaps %xmm4, 0x370(%rsp)
vmovaps %xmm3, 0x360(%rsp)
vmovaps %xmm0, 0x350(%rsp)
vmovaps 0x370(%rsp), %xmm4
vmovaps 0x360(%rsp), %xmm3
vmovaps 0x350(%rsp), %xmm0
vmovaps %xmm4, 0x630(%rsp)
vmovaps %xmm3, 0x620(%rsp)
vmovaps %xmm0, 0x610(%rsp)
vmovaps 0x630(%rsp), %xmm3
vmovaps 0x620(%rsp), %xmm0
vmovaps 0x610(%rsp), %xmm4
vfmadd213ps %xmm4, %xmm3, %xmm0 # xmm0 = (xmm3 * xmm0) + xmm4
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm3
vmovaps 0x100(%rsp), %xmm0
vmovaps %xmm3, 0x230(%rsp)
vmovaps %xmm0, 0x220(%rsp)
vmovaps 0x230(%rsp), %xmm0
vmovaps 0x220(%rsp), %xmm3
vaddps %xmm3, %xmm0, %xmm0
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0x120(%rsp), %xmm0
vmovaps %xmm0, 0x280(%rsp)
vcvttps2dq 0x280(%rsp), %xmm0
vmovdqa %xmm0, 0x110(%rsp)
vmovdqa 0x110(%rsp), %xmm0
vmovdqa %xmm0, 0x480(%rsp)
vpbroadcastq 0x7b197c(%rip), %xmm0 # 0x1e0be18
vmovdqa %xmm0, 0x470(%rsp)
vmovdqa 0x480(%rsp), %xmm0
vmovdqa 0x470(%rsp), %xmm3
vpaddd %xmm3, %xmm0, %xmm0
vmovdqa %xmm0, 0x110(%rsp)
vmovdqa 0x110(%rsp), %xmm0
vmovdqa %xmm0, 0x4a0(%rsp)
movl $0x17, 0x49c(%rsp)
vmovdqa 0x4a0(%rsp), %xmm0
movl 0x49c(%rsp), %eax
vmovd %eax, %xmm3
vpslld %xmm3, %xmm0, %xmm0
vmovdqa %xmm0, 0x110(%rsp)
vmovdqa 0x110(%rsp), %xmm0
vmovdqa %xmm0, 0x4b0(%rsp)
vmovdqa 0x4b0(%rsp), %xmm0
vmovaps %xmm0, 0xd0(%rsp)
vmovaps 0xe0(%rsp), %xmm3
vmovaps 0xd0(%rsp), %xmm0
vmovaps %xmm3, 0x190(%rsp)
vmovaps %xmm0, 0x180(%rsp)
vmovaps 0x190(%rsp), %xmm0
vmulps 0x180(%rsp), %xmm0, %xmm0
vmovaps %xmm0, 0xe0(%rsp)
vmovaps 0xe0(%rsp), %xmm0
vmovaps %xmm2, 0x270(%rsp)
vmovaps %xmm0, 0x260(%rsp)
vmovaps 0x270(%rsp), %xmm0
vaddps 0x260(%rsp), %xmm0, %xmm0
vmovaps %xmm1, 0x650(%rsp)
vmovaps %xmm0, 0x640(%rsp)
vmovaps 0x650(%rsp), %xmm0
vdivps 0x640(%rsp), %xmm0, %xmm1
vmovaps -0x70(%rsp), %xmm0
vmovaps %xmm1, 0x30(%rsp)
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmulps 0x20(%rsp), %xmm0, %xmm1
vmovaps -0x60(%rsp), %xmm0
vmovaps %xmm1, 0x10(%rsp)
vmovaps %xmm0, (%rsp)
vmovaps 0x10(%rsp), %xmm0
vsubps (%rsp), %xmm0, %xmm0
addq $0x668, %rsp # imm = 0x668
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/unaryop_x86_avx512.cpp |
2,113,320 | ncnn::UnaryOp_x86_avx512_functor::unary_op_tanh::func(float const&) const | float func(const float& x) const
{
return (float)tanh(x);
} | subq $0x18, %rsp
movq %rdi, 0x10(%rsp)
movq %rsi, 0x8(%rsp)
movq 0x8(%rsp), %rax
vmovss (%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/build_O0/src/layer/x86/unaryop_x86_avx512.cpp |
2,113,321 | ncnn::UnaryOp_x86_fma::UnaryOp_x86_fma() | UnaryOp_x86_fma::UnaryOp_x86_fma()
{
#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/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
2,113,322 | ncnn::UnaryOp_x86_fma::UnaryOp_x86_fma() | UnaryOp_x86_fma::UnaryOp_x86_fma()
{
#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 0x8775f0(%rip), %rcx # 0x1ed1c88
addq $0x18, %rcx
movq %rcx, (%rax)
leaq 0x8775e2(%rip), %rcx # 0x1ed1c88
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/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
2,113,323 | ncnn::UnaryOp_x86_fma::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int UnaryOp_x86_fma::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_fma_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 0x165a715
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165a9f0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x1, 0xd0(%rax,%rcx)
jne 0x165a742
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165b0a0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x2, 0xd0(%rax,%rcx)
jne 0x165a76f
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165b750
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x3, 0xd0(%rax,%rcx)
jne 0x165a79c
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165be00
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x4, 0xd0(%rax,%rcx)
jne 0x165a7c9
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165c4b0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x5, 0xd0(%rax,%rcx)
jne 0x165a7f6
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165cb60
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x6, 0xd0(%rax,%rcx)
jne 0x165a823
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165d210
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x7, 0xd0(%rax,%rcx)
jne 0x165a850
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165d8c0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x8, 0xd0(%rax,%rcx)
jne 0x165a87d
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165df70
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x9, 0xd0(%rax,%rcx)
jne 0x165a8aa
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165e620
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xa, 0xd0(%rax,%rcx)
jne 0x165a8d7
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165ecd0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xb, 0xd0(%rax,%rcx)
jne 0x165a904
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165f380
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xc, 0xd0(%rax,%rcx)
jne 0x165a931
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x165fa30
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xd, 0xd0(%rax,%rcx)
jne 0x165a95e
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16600e0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xe, 0xd0(%rax,%rcx)
jne 0x165a988
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1660790
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0xf, 0xd0(%rax,%rcx)
jne 0x165a9b2
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1660e40
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
movq (%rsp), %rax
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
cmpl $0x10, 0xd0(%rax,%rcx)
jne 0x165a9dc
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x16614f0
movl %eax, 0x24(%rsp)
jmp 0x165a9e4
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/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
2,113,324 | int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_fma_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;
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x240, %rsp # imm = 0x240
movq %rdi, 0xf8(%rsp)
movq %rsi, 0xf0(%rsp)
movq 0xf8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xe8(%rsp)
movq 0xf8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xe4(%rsp)
movq 0xf8(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xe0(%rsp)
movq 0xf8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xdc(%rsp)
movq 0xf8(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xd8(%rsp)
movl 0xe8(%rsp), %eax
imull 0xe4(%rsp), %eax
imull 0xe0(%rsp), %eax
imull 0xd8(%rsp), %eax
movl %eax, 0xd4(%rsp)
movl $0x0, 0xd0(%rsp)
movl 0xd0(%rsp), %eax
cmpl 0xdc(%rsp), %eax
jge 0x165b096
movq 0xf8(%rsp), %rcx
movl 0xd0(%rsp), %eax
leaq 0x80(%rsp), %rdx
movq %rdx, 0x110(%rsp)
movq %rcx, 0x108(%rsp)
movl %eax, 0x104(%rsp)
movq 0x108(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0x103(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0x104(%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 0x80(%rsp), %r10
movq %r10, 0x1d0(%rsp)
movl %r9d, 0x1cc(%rsp)
movl %r8d, 0x1c8(%rsp)
movl %edi, 0x1c4(%rsp)
movq %rsi, 0x1b8(%rsp)
movq %rdx, 0x1b0(%rsp)
movl %ecx, 0x1ac(%rsp)
movq %rax, 0x1a0(%rsp)
movq 0x1d0(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x1b8(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x1b0(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x1ac(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x1a0(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x1cc(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x1c8(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x1c4(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x1e0(%rsp)
movl $0x10, 0x1dc(%rsp)
movq 0x1e0(%rsp), %rax
movslq 0x1dc(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x1dc(%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, 0xa8(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x165ac5e
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0xc0(%rsp)
movb $0x1, 0x103(%rsp)
testb $0x1, 0x103(%rsp)
jne 0x165ad8d
leaq 0x80(%rsp), %rax
movq %rax, 0x120(%rsp)
movq 0x120(%rsp), %rax
movq %rax, 0x210(%rsp)
movq 0x210(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165ad33
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x20c(%rsp) # imm = 0xFFFFFFFF
movl 0x20c(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x208(%rsp)
cmpl $0x1, 0x208(%rsp)
jne 0x165ad33
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165ad04
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165ad02
jmp 0x165ad31
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x218(%rsp)
cmpq $0x0, 0x218(%rsp)
je 0x165ad2f
movq 0x218(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165ad31
jmp 0x165ad33
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 0x165ad8b
movq %rax, %rdi
callq 0x678a0
jmp 0x165ad8d
leaq 0x80(%rsp), %rax
movq %rax, 0x118(%rsp)
movq 0x118(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x10(%rsp)
leaq 0x80(%rsp), %rax
movq %rax, 0x128(%rsp)
movq 0x128(%rsp), %rax
movq %rax, 0x200(%rsp)
movq 0x200(%rsp), %rax
movq %rax, 0x8(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165ae6c
movq 0x8(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x1fc(%rsp) # imm = 0xFFFFFFFF
movl 0x1fc(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x1f8(%rsp)
cmpl $0x1, 0x1f8(%rsp)
jne 0x165ae6c
movq 0x8(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165ae3d
movq 0x8(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165ae3b
jmp 0x165ae6a
movq 0x8(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x220(%rsp)
cmpq $0x0, 0x220(%rsp)
je 0x165ae68
movq 0x220(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165ae6a
jmp 0x165ae6c
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 0x165aec4
movq %rax, %rdi
callq 0x678a0
movq 0x10(%rsp), %rax
movq %rax, 0xc8(%rsp)
movl $0x0, 0x70(%rsp)
movl 0x70(%rsp), %eax
addl $0x7, %eax
cmpl 0xd4(%rsp), %eax
jge 0x165af7f
movq 0xc8(%rsp), %rax
movq %rax, 0x138(%rsp)
movq 0x138(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x40(%rsp)
leaq 0xef(%rsp), %rdi
leaq 0x40(%rsp), %rsi
callq 0x1661bd0
vmovaps %ymm0, 0x40(%rsp)
movq 0xc8(%rsp), %rax
vmovaps 0x40(%rsp), %ymm0
movq %rax, 0x170(%rsp)
vmovaps %ymm0, 0x140(%rsp)
vmovaps 0x140(%rsp), %ymm0
movq 0x170(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0xc8(%rsp), %rax
addq $0x20, %rax
movq %rax, 0xc8(%rsp)
movl 0x70(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x70(%rsp)
jmp 0x165aed9
jmp 0x165af81
movl 0x70(%rsp), %eax
addl $0x3, %eax
cmpl 0xd4(%rsp), %eax
jge 0x165b02a
movq 0xc8(%rsp), %rax
movq %rax, 0x178(%rsp)
movq 0x178(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x30(%rsp)
leaq 0xef(%rsp), %rdi
leaq 0x30(%rsp), %rsi
vzeroupper
callq 0x1661c60
vmovaps %xmm0, 0x30(%rsp)
movq 0xc8(%rsp), %rax
vmovaps 0x30(%rsp), %xmm0
movq %rax, 0x198(%rsp)
vmovaps %xmm0, 0x180(%rsp)
vmovaps 0x180(%rsp), %xmm0
movq 0x198(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0xc8(%rsp), %rax
addq $0x10, %rax
movq %rax, 0xc8(%rsp)
movl 0x70(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x70(%rsp)
jmp 0x165af81
jmp 0x165b02c
movl 0x70(%rsp), %eax
cmpl 0xd4(%rsp), %eax
jge 0x165b07e
movq 0xc8(%rsp), %rsi
leaq 0xef(%rsp), %rdi
vzeroupper
callq 0x1661ce0
movq 0xc8(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0xc8(%rsp), %rax
addq $0x4, %rax
movq %rax, 0xc8(%rsp)
movl 0x70(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x70(%rsp)
jmp 0x165b02c
jmp 0x165b080
movl 0xd0(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xd0(%rsp)
jmp 0x165aa9a
xorl %eax, %eax
movq %rbp, %rsp
popq %rbp
vzeroupper
retq
| ysh329[P]ncnn[P]build_O0[P]examples[P]scrfd.asm_src.json | O0 | ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
2,113,325 | int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_fma_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;
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x200, %rsp # imm = 0x200
movq %rdi, 0xe0(%rsp)
movq %rsi, 0xd8(%rsp)
movq 0xe0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xd0(%rsp)
movq 0xe0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xcc(%rsp)
movq 0xe0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xc8(%rsp)
movq 0xe0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xc4(%rsp)
movq 0xe0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xc0(%rsp)
movl 0xd0(%rsp), %eax
imull 0xcc(%rsp), %eax
imull 0xc8(%rsp), %eax
imull 0xc0(%rsp), %eax
movl %eax, 0xbc(%rsp)
movl $0x0, 0xb8(%rsp)
movl 0xb8(%rsp), %eax
cmpl 0xc4(%rsp), %eax
jge 0x165b737
movq 0xe0(%rsp), %rcx
movl 0xb8(%rsp), %eax
leaq 0x68(%rsp), %rdx
movq %rdx, 0xf8(%rsp)
movq %rcx, 0xf0(%rsp)
movl %eax, 0xec(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xeb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xec(%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 0x68(%rsp), %r10
movq %r10, 0x1a8(%rsp)
movl %r9d, 0x1a4(%rsp)
movl %r8d, 0x1a0(%rsp)
movl %edi, 0x19c(%rsp)
movq %rsi, 0x190(%rsp)
movq %rdx, 0x188(%rsp)
movl %ecx, 0x184(%rsp)
movq %rax, 0x178(%rsp)
movq 0x1a8(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x190(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x188(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x184(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x178(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x1a4(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x1a0(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x19c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x1b8(%rsp)
movl $0x10, 0x1b4(%rsp)
movq 0x1b8(%rsp), %rax
movslq 0x1b4(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x1b4(%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, 0x90(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x165b308
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0xa8(%rsp)
movb $0x1, 0xeb(%rsp)
testb $0x1, 0xeb(%rsp)
jne 0x165b434
leaq 0x68(%rsp), %rax
movq %rax, 0x108(%rsp)
movq 0x108(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165b3da
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x1d4(%rsp) # imm = 0xFFFFFFFF
movl 0x1d4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x1d0(%rsp)
cmpl $0x1, 0x1d0(%rsp)
jne 0x165b3da
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165b3ab
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165b3a9
jmp 0x165b3d8
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1e0(%rsp)
cmpq $0x0, 0x1e0(%rsp)
je 0x165b3d6
movq 0x1e0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165b3d8
jmp 0x165b3da
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 0x165b432
movq %rax, %rdi
callq 0x678a0
jmp 0x165b434
leaq 0x68(%rsp), %rax
movq %rax, 0x100(%rsp)
movq 0x100(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x68(%rsp), %rax
movq %rax, 0x110(%rsp)
movq 0x110(%rsp), %rax
movq %rax, 0x1c8(%rsp)
movq 0x1c8(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165b50d
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x1c4(%rsp) # imm = 0xFFFFFFFF
movl 0x1c4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x1c0(%rsp)
cmpl $0x1, 0x1c0(%rsp)
jne 0x165b50d
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165b4de
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165b4dc
jmp 0x165b50b
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1e8(%rsp)
cmpq $0x0, 0x1e8(%rsp)
je 0x165b509
movq 0x1e8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165b50b
jmp 0x165b50d
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 0x165b565
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0xb0(%rsp)
movl $0x0, 0x64(%rsp)
movl 0x64(%rsp), %eax
addl $0x7, %eax
cmpl 0xbc(%rsp), %eax
jge 0x165b620
movq 0xb0(%rsp), %rax
movq %rax, 0x118(%rsp)
movq 0x118(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x40(%rsp)
leaq 0xd7(%rsp), %rdi
leaq 0x40(%rsp), %rsi
callq 0x1661d10
vmovaps %ymm0, 0x40(%rsp)
movq 0xb0(%rsp), %rax
vmovaps 0x40(%rsp), %ymm0
movq %rax, 0x150(%rsp)
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
movq 0x150(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x20, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x64(%rsp)
jmp 0x165b57a
jmp 0x165b622
movl 0x64(%rsp), %eax
addl $0x3, %eax
cmpl 0xbc(%rsp), %eax
jge 0x165b6cb
movq 0xb0(%rsp), %rax
movq %rax, 0x158(%rsp)
movq 0x158(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x30(%rsp)
leaq 0xd7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
vzeroupper
callq 0x1661d60
vmovaps %xmm0, 0x30(%rsp)
movq 0xb0(%rsp), %rax
vmovaps 0x30(%rsp), %xmm0
movq %rax, 0x170(%rsp)
vmovaps %xmm0, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm0
movq 0x170(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x10, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x64(%rsp)
jmp 0x165b622
jmp 0x165b6cd
movl 0x64(%rsp), %eax
cmpl 0xbc(%rsp), %eax
jge 0x165b71f
movq 0xb0(%rsp), %rsi
leaq 0xd7(%rsp), %rdi
vzeroupper
callq 0x1661da0
movq 0xb0(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x4, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x64(%rsp)
jmp 0x165b6cd
jmp 0x165b721
movl 0xb8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xb8(%rsp)
jmp 0x165b14a
xorl %eax, %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/unaryop_x86_fma.cpp |
2,113,326 | int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_fma_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;
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x200, %rsp # imm = 0x200
movq %rdi, 0xe0(%rsp)
movq %rsi, 0xd8(%rsp)
movq 0xe0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xd0(%rsp)
movq 0xe0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xcc(%rsp)
movq 0xe0(%rsp), %rax
movl 0x34(%rax), %eax
movl %eax, 0xc8(%rsp)
movq 0xe0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xc4(%rsp)
movq 0xe0(%rsp), %rax
movl 0x18(%rax), %eax
movl %eax, 0xc0(%rsp)
movl 0xd0(%rsp), %eax
imull 0xcc(%rsp), %eax
imull 0xc8(%rsp), %eax
imull 0xc0(%rsp), %eax
movl %eax, 0xbc(%rsp)
movl $0x0, 0xb8(%rsp)
movl 0xb8(%rsp), %eax
cmpl 0xc4(%rsp), %eax
jge 0x165bde7
movq 0xe0(%rsp), %rcx
movl 0xb8(%rsp), %eax
leaq 0x68(%rsp), %rdx
movq %rdx, 0xf8(%rsp)
movq %rcx, 0xf0(%rsp)
movl %eax, 0xec(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x28(%rsp)
movb $0x0, 0xeb(%rsp)
movl 0x2c(%rax), %r9d
movl 0x30(%rax), %r8d
movl 0x34(%rax), %edi
movq (%rax), %rsi
movq 0x40(%rax), %rcx
movslq 0xec(%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 0x68(%rsp), %r10
movq %r10, 0x1a8(%rsp)
movl %r9d, 0x1a4(%rsp)
movl %r8d, 0x1a0(%rsp)
movl %edi, 0x19c(%rsp)
movq %rsi, 0x190(%rsp)
movq %rdx, 0x188(%rsp)
movl %ecx, 0x184(%rsp)
movq %rax, 0x178(%rsp)
movq 0x1a8(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movq 0x190(%rsp), %rax
movq %rax, (%rcx)
movq $0x0, 0x8(%rcx)
movq 0x188(%rsp), %rax
movq %rax, 0x10(%rcx)
movl 0x184(%rsp), %eax
movl %eax, 0x18(%rcx)
movq 0x178(%rsp), %rax
movq %rax, 0x20(%rcx)
movl $0x3, 0x28(%rcx)
movl 0x1a4(%rsp), %eax
movl %eax, 0x2c(%rcx)
movl 0x1a0(%rsp), %eax
movl %eax, 0x30(%rcx)
movl $0x1, 0x34(%rcx)
movl 0x19c(%rsp), %eax
movl %eax, 0x38(%rcx)
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rdx
imulq %rdx, %rax
imulq 0x10(%rcx), %rax
movq %rax, 0x1b8(%rsp)
movl $0x10, 0x1b4(%rsp)
movq 0x1b8(%rsp), %rax
movslq 0x1b4(%rsp), %rdx
addq %rdx, %rax
subq $0x1, %rax
xorl %edx, %edx
subl 0x1b4(%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, 0x90(%rsp)
cmpl $0x4, 0x28(%rax)
jne 0x165b9b8
movq 0x28(%rsp), %rcx
movslq 0x2c(%rcx), %rax
movslq 0x30(%rcx), %rcx
imulq %rcx, %rax
movq %rax, 0xa8(%rsp)
movb $0x1, 0xeb(%rsp)
testb $0x1, 0xeb(%rsp)
jne 0x165bae4
leaq 0x68(%rsp), %rax
movq %rax, 0x108(%rsp)
movq 0x108(%rsp), %rax
movq %rax, 0x1d8(%rsp)
movq 0x1d8(%rsp), %rax
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165ba8a
movq 0x18(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x1d4(%rsp) # imm = 0xFFFFFFFF
movl 0x1d4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x1d0(%rsp)
cmpl $0x1, 0x1d0(%rsp)
jne 0x165ba8a
movq 0x18(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165ba5b
movq 0x18(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165ba59
jmp 0x165ba88
movq 0x18(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1e0(%rsp)
cmpq $0x0, 0x1e0(%rsp)
je 0x165ba86
movq 0x1e0(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165ba88
jmp 0x165ba8a
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 0x165bae2
movq %rax, %rdi
callq 0x678a0
jmp 0x165bae4
leaq 0x68(%rsp), %rax
movq %rax, 0x100(%rsp)
movq 0x100(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x8(%rsp)
leaq 0x68(%rsp), %rax
movq %rax, 0x110(%rsp)
movq 0x110(%rsp), %rax
movq %rax, 0x1c8(%rsp)
movq 0x1c8(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x8(%rax)
je 0x165bbbd
movq 0x10(%rsp), %rax
movq 0x8(%rax), %rcx
movl $0xffffffff, 0x1c4(%rsp) # imm = 0xFFFFFFFF
movl 0x1c4(%rsp), %eax
lock
xaddl %eax, (%rcx)
movl %eax, 0x1c0(%rsp)
cmpl $0x1, 0x1c0(%rsp)
jne 0x165bbbd
movq 0x10(%rsp), %rax
cmpq $0x0, 0x20(%rax)
je 0x165bb8e
movq 0x10(%rsp), %rax
movq (%rax), %rsi
movq 0x20(%rax), %rdi
movq (%rdi), %rax
movq 0x18(%rax), %rax
vzeroupper
callq *%rax
jmp 0x165bb8c
jmp 0x165bbbb
movq 0x10(%rsp), %rax
movq (%rax), %rax
movq %rax, 0x1e8(%rsp)
cmpq $0x0, 0x1e8(%rsp)
je 0x165bbb9
movq 0x1e8(%rsp), %rdi
vzeroupper
callq 0x5f480
jmp 0x165bbbb
jmp 0x165bbbd
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 0x165bc15
movq %rax, %rdi
callq 0x678a0
movq 0x8(%rsp), %rax
movq %rax, 0xb0(%rsp)
movl $0x0, 0x64(%rsp)
movl 0x64(%rsp), %eax
addl $0x7, %eax
cmpl 0xbc(%rsp), %eax
jge 0x165bcd0
movq 0xb0(%rsp), %rax
movq %rax, 0x118(%rsp)
movq 0x118(%rsp), %rax
vmovups (%rax), %ymm0
vmovaps %ymm0, 0x40(%rsp)
leaq 0xd7(%rsp), %rdi
leaq 0x40(%rsp), %rsi
callq 0x1661dd0
vmovaps %ymm0, 0x40(%rsp)
movq 0xb0(%rsp), %rax
vmovaps 0x40(%rsp), %ymm0
movq %rax, 0x150(%rsp)
vmovaps %ymm0, 0x120(%rsp)
vmovaps 0x120(%rsp), %ymm0
movq 0x150(%rsp), %rax
vmovups %ymm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x20, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x8, %eax
movl %eax, 0x64(%rsp)
jmp 0x165bc2a
jmp 0x165bcd2
movl 0x64(%rsp), %eax
addl $0x3, %eax
cmpl 0xbc(%rsp), %eax
jge 0x165bd7b
movq 0xb0(%rsp), %rax
movq %rax, 0x158(%rsp)
movq 0x158(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x30(%rsp)
leaq 0xd7(%rsp), %rdi
leaq 0x30(%rsp), %rsi
vzeroupper
callq 0x1661df0
vmovaps %xmm0, 0x30(%rsp)
movq 0xb0(%rsp), %rax
vmovaps 0x30(%rsp), %xmm0
movq %rax, 0x170(%rsp)
vmovaps %xmm0, 0x160(%rsp)
vmovaps 0x160(%rsp), %xmm0
movq 0x170(%rsp), %rax
vmovaps %xmm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x10, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x4, %eax
movl %eax, 0x64(%rsp)
jmp 0x165bcd2
jmp 0x165bd7d
movl 0x64(%rsp), %eax
cmpl 0xbc(%rsp), %eax
jge 0x165bdcf
movq 0xb0(%rsp), %rsi
leaq 0xd7(%rsp), %rdi
vzeroupper
callq 0x1661e20
movq 0xb0(%rsp), %rax
vmovss %xmm0, (%rax)
movq 0xb0(%rsp), %rax
addq $0x4, %rax
movq %rax, 0xb0(%rsp)
movl 0x64(%rsp), %eax
addl $0x1, %eax
movl %eax, 0x64(%rsp)
jmp 0x165bd7d
jmp 0x165bdd1
movl 0xb8(%rsp), %eax
addl $0x1, %eax
movl %eax, 0xb8(%rsp)
jmp 0x165b7fa
xorl %eax, %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/unaryop_x86_fma.cpp |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.