name string | code string | asm string | file string |
|---|---|---|---|
ncnn::convolution_pack8_avx(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static void convolution_pack8_avx(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data_packed, const Mat& bias_data, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int activation_type, const Mat& activation_params, const Option& opt)
{
int w = bottom_blob.w;
int... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x5580, %rsp # imm = 0x5580
movq 0x40(%rbp), %rax
movq 0x38(%rbp), %rax
movl 0x30(%rbp), %eax
movl 0x28(%rbp), %eax
movl 0x20(%rbp), %eax
movl 0x18(%rbp), %eax
movl 0x10(%rbp), %eax
movq %rdi, 0x908(%rsp)
movq %rsi, 0x900(%rsp)
movq %rdx, 0x8f8(%rsp)
movq %rc... | /ysh329[P]ncnn/src/layer/x86/convolution_pack8.h |
ncnn::convolution_im2col_sgemm_pack4to8_avx(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, ncnn::Option const&) | static void convolution_im2col_sgemm_pack4to8_avx(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Mat& _bias, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, const Option& opt)
{
int w = bottom_blob.w;
int inch = bottom_blob.c;
int outw = top_blob.w;
... | subq $0x408, %rsp # imm = 0x408
movq 0x430(%rsp), %rax
movl 0x428(%rsp), %eax
movl 0x420(%rsp), %eax
movl 0x418(%rsp), %eax
movl 0x410(%rsp), %eax
movq %rdi, 0x1d0(%rsp)
movq %rsi, 0x1c8(%rsp)
movq %rdx, 0x1c0(%rsp)
movq %rcx, 0x1b8(%rsp)
movl %r8d, 0x1b4(%rsp)
movl %r9d, 0x1b0(%rsp)
movq 0x1d0(%rsp), %rax
m... | /ysh329[P]ncnn/src/layer/x86/convolution_sgemm_pack4to8.h |
ncnn::conv1x1s2_sgemm_pack8to4_avx(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | static void conv1x1s2_sgemm_pack8to4_avx(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Mat& _bias, const Option& opt)
{
int w = bottom_blob.w;
int channels = bottom_blob.c;
size_t elemsize = bottom_blob.elemsize;
int elempack = bottom_blob.elempack;
int outw = top_blob.w;
int ... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x420, %rsp # imm = 0x420
movq %rdi, 0x1e8(%rsp)
movq %rsi, 0x1e0(%rsp)
movq %rdx, 0x1d8(%rsp)
movq %rcx, 0x1d0(%rsp)
movq %r8, 0x1c8(%rsp)
movq 0x1e8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1c4(%rsp)
movq 0x1e8(%rsp), %rax
movl 0x38(%rax), %eax
movl ... | /ysh329[P]ncnn/src/layer/x86/convolution_1x1_pack8to4.h |
ncnn::convolution_pack8to4_avx(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static void convolution_pack8to4_avx(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data_packed, const Mat& bias_data, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int activation_type, const Mat& activation_params, const Option& opt)
{
int w = bottom_blob.w;
... | subq $0x29b8, %rsp # imm = 0x29B8
movq 0x29f0(%rsp), %rax
movq 0x29e8(%rsp), %rax
movl 0x29e0(%rsp), %eax
movl 0x29d8(%rsp), %eax
movl 0x29d0(%rsp), %eax
movl 0x29c8(%rsp), %eax
movl 0x29c0(%rsp), %eax
movq %rdi, 0x560(%rsp)
movq %rsi, 0x558(%rsp)
movq %rdx, 0x550(%rsp)
movq %rcx, 0x548(%rsp)
movl %r8d, 0x544... | /ysh329[P]ncnn/src/layer/x86/convolution_pack8to4.h |
ncnn::convolution_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static void convolution_pack4_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data_packed, const Mat& bias_data, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int activation_type, const Mat& activation_params, const Option& opt)
{
int w = bottom_blob.w;
int... | subq $0x2648, %rsp # imm = 0x2648
movq 0x2680(%rsp), %rax
movq 0x2678(%rsp), %rax
movl 0x2670(%rsp), %eax
movl 0x2668(%rsp), %eax
movl 0x2660(%rsp), %eax
movl 0x2658(%rsp), %eax
movl 0x2650(%rsp), %eax
movq %rdi, 0x410(%rsp)
movq %rsi, 0x408(%rsp)
movq %rdx, 0x400(%rsp)
movq %rcx, 0x3f8(%rsp)
movl %r8d, 0x3f4... | /ysh329[P]ncnn/src/layer/x86/convolution_pack4.h |
ncnn::conv1x1s2_sgemm_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | static void conv1x1s2_sgemm_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Mat& _bias, const Option& opt)
{
int w = bottom_blob.w;
int channels = bottom_blob.c;
size_t elemsize = bottom_blob.elemsize;
int elempack = bottom_blob.elempack;
int outw = top_blob.w;
int outh = to... | subq $0x398, %rsp # imm = 0x398
movq %rdi, 0x1b8(%rsp)
movq %rsi, 0x1b0(%rsp)
movq %rdx, 0x1a8(%rsp)
movq %rcx, 0x1a0(%rsp)
movq %r8, 0x198(%rsp)
movq 0x1b8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x194(%rsp)
movq 0x1b8(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x190(%rsp)
movq 0x1b8(%rsp), %rax
movq... | /ysh329[P]ncnn/src/layer/x86/convolution_1x1.h |
ncnn::Convolution_x86_fma::forward(std::vector<ncnn::Mat, std::allocator<ncnn::Mat>> const&, std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int Convolution_x86_fma::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& top_blobs, const Option& opt) const
{
const Mat& bottom_blob = bottom_blobs[0];
const Mat& _weight_data = bottom_blobs[1];
Mat& top_blob = top_blobs[0];
const int _kernel_w = _weight_data.w;
const int _kernel_h... | subq $0x448, %rsp # imm = 0x448
movq %rdi, 0x2b0(%rsp)
movq %rsi, 0x2a8(%rsp)
movq %rdx, 0x2a0(%rsp)
movq %rcx, 0x298(%rsp)
movq 0x2b0(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0x2a8(%rsp), %rdi
xorl %eax, %eax
movl %eax, %esi
movq %rsi, 0x100(%rsp)
callq 0xb5820
movq %rax, 0x290(%rsp)
movq 0x2a8(%rsp), %rdi
mo... | /ysh329[P]ncnn/build_O0/src/layer/x86/convolution_x86_fma.cpp |
ncnn::conv3x3s1_winograd63_transform_output_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Option const&) | static void conv3x3s1_winograd63_transform_output_sse(const Mat& top_blob_tm, Mat& top_blob, const Mat& bias, const Option& opt)
{
const int outw = top_blob.w;
const int outh = top_blob.h;
const int outch = top_blob.c;
const int w_tiles = outw / 6;
const int h_tiles = outh / 6;
const int tiles ... | subq $0x438, %rsp # imm = 0x438
movq %rdi, 0x290(%rsp)
movq %rsi, 0x288(%rsp)
movq %rdx, 0x280(%rsp)
movq %rcx, 0x278(%rsp)
movq 0x288(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x274(%rsp)
movq 0x288(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x270(%rsp)
movq 0x288(%rsp), %rax
movl 0x38(%rax), %eax
movl... | /ysh329[P]ncnn/src/layer/x86/convolution_winograd_transform.h |
ncnn::conv3x3s1_winograd63_transform_input_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void conv3x3s1_winograd63_transform_input_pack4_sse(const Mat& bottom_blob, Mat& bottom_blob_tm, const Option& opt)
{
const int w = bottom_blob.w;
const int h = bottom_blob.h;
const int inch = bottom_blob.c;
const int w_tiles = (w - 2) / 6;
const int h_tiles = (h - 2) / 6;
const int tile... | subq $0x1ce8, %rsp # imm = 0x1CE8
movq %rdi, 0xb40(%rsp)
movq %rsi, 0xb38(%rsp)
movq %rdx, 0xb30(%rsp)
movq 0xb40(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb2c(%rsp)
movq 0xb40(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xb28(%rsp)
movq 0xb40(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0xb24(%rsp)
mov... | /ysh329[P]ncnn/src/layer/x86/convolution_winograd_transform_pack4.h |
ncnn::convolution_winograd_dot_pack4_sse(ncnn::Mat&, int, ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void convolution_winograd_dot_pack4_sse(Mat& bottom_blob_tm, int outch, const Mat& kernel_tm, Mat& top_blob_tm, const Option& opt)
{
// Mat bottom_blob_tm(tiles, 16/36/64, inch, 16u, 4, opt.workspace_allocator);
const int tiles = bottom_blob_tm.w;
const int batch = bottom_blob_tm.h;
const int in... | subq $0x3058, %rsp # imm = 0x3058
movq %rdi, 0xe30(%rsp)
movl %esi, 0xe2c(%rsp)
movq %rdx, 0xe20(%rsp)
movq %rcx, 0xe18(%rsp)
movq %r8, 0xe10(%rsp)
movq 0xe30(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xe0c(%rsp)
movq 0xe30(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xe08(%rsp)
movq 0xe30(%rsp), %rax
mov... | /ysh329[P]ncnn/src/layer/x86/convolution_winograd_dot_pack4.h |
ncnn::conv3x3s1_winograd63_transform_output_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Option const&) | static void conv3x3s1_winograd63_transform_output_pack4_sse(const Mat& top_blob_tm, Mat& top_blob, const Mat& bias, const Option& opt)
{
const int outw = top_blob.w;
const int outh = top_blob.h;
const int outch = top_blob.c;
const int w_tiles = outw / 6;
const int h_tiles = outh / 6;
const int ... | subq $0x1998, %rsp # imm = 0x1998
movq %rdi, 0xa70(%rsp)
movq %rsi, 0xa68(%rsp)
movq %rdx, 0xa60(%rsp)
movq %rcx, 0xa58(%rsp)
movq 0xa68(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xa54(%rsp)
movq 0xa68(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xa50(%rsp)
movq 0xa68(%rsp), %rax
movl 0x38(%rax), %eax
mov... | /ysh329[P]ncnn/src/layer/x86/convolution_winograd_transform_pack4.h |
ncnn::Deconvolution::load_model(ncnn::ModelBin const&) | int Deconvolution::load_model(const ModelBin& mb)
{
weight_data = mb.load(weight_data_size, 0);
if (weight_data.empty())
return -100;
if (bias_term)
{
bias_data = mb.load(num_output, 1);
if (bias_data.empty())
return -100;
}
return 0;
} | subq $0x238, %rsp # imm = 0x238
movq %rdi, 0x118(%rsp)
movq %rsi, 0x110(%rsp)
movq 0x118(%rsp), %rax
movq %rax, 0x60(%rsp)
movq 0x110(%rsp), %rsi
movl 0x110(%rax), %edx
movq (%rsi), %rax
leaq 0xc8(%rsp), %rdi
xorl %ecx, %ecx
callq *0x10(%rax)
movq 0x60(%rsp), %rax
addq $0x160, %rax # imm = 0x160
m... | /ysh329[P]ncnn/src/layer/deconvolution.cpp |
ncnn::Deconvolution::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Deconvolution::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
int w = bottom_blob.w;
int h = bottom_blob.h;
size_t elemsize = bottom_blob.elemsize;
const int kernel_extent_w = dilation_w * (kernel_w - 1) + 1;
const int kernel_extent_h = dilation_h * (kernel_h - 1) + 1... | pushq %rbp
pushq %r15
pushq %r14
pushq %rbx
subq $0x1e8, %rsp # imm = 0x1E8
movq %rdi, 0x138(%rsp)
movq %rsi, 0x130(%rsp)
movq %rdx, 0x128(%rsp)
movq %rcx, 0x120(%rsp)
movq 0x138(%rsp), %rax
movq %rax, 0x90(%rsp)
movq 0x130(%rsp), %rcx
movl 0x2c(%rcx), %ecx
movl %ecx, 0x11c(%rsp)
movq 0x130(%rsp), %rcx
movl ... | /ysh329[P]ncnn/src/layer/deconvolution.cpp |
ncnn::deconvolution(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static int deconvolution(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data, const Mat& bias_data, int kernel_w, int kernel_h, int stride_w, int stride_h, int dilation_w, int dilation_h, int activation_type, const Mat& activation_params, const Option& opt)
{
const int outw = top_blob.w;
const int out... | subq $0x498, %rsp # imm = 0x498
movq 0x4d0(%rsp), %rax
movq 0x4c8(%rsp), %rax
movl 0x4c0(%rsp), %eax
movl 0x4b8(%rsp), %eax
movl 0x4b0(%rsp), %eax
movl 0x4a8(%rsp), %eax
movl 0x4a0(%rsp), %eax
movq %rdi, 0x210(%rsp)
movq %rsi, 0x208(%rsp)
movq %rdx, 0x200(%rsp)
movq %rcx, 0x1f8(%rsp)
movl %r8d, 0x1f4(%rsp)
m... | /ysh329[P]ncnn/src/layer/deconvolution.cpp |
ncnn::Deconvolution::cut_padding(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | void Deconvolution::cut_padding(const Mat& top_blob_bordered, Mat& top_blob, const Option& opt) const
{
if (pad_left > 0 || pad_right > 0 || pad_top > 0 || pad_bottom > 0)
{
copy_cut_border(top_blob_bordered, top_blob, pad_top, pad_bottom, pad_left, pad_right, opt);
}
else if (output_w > 0 && ou... | subq $0x98, %rsp
movq %rdi, 0x58(%rsp)
movq %rsi, 0x50(%rsp)
movq %rdx, 0x48(%rsp)
movq %rcx, 0x40(%rsp)
movq 0x58(%rsp), %rax
movq %rax, 0x30(%rsp)
cmpl $0x0, 0xec(%rax)
jg 0x943288
movq 0x30(%rsp), %rax
cmpl $0x0, 0xf0(%rax)
jg 0x943288
movq 0x30(%rsp), %rax
cmpl $0x0, 0xf4(%rax)
jg 0x943288
movq 0x30(%rsp), %rax
cmp... | /ysh329[P]ncnn/src/layer/deconvolution.cpp |
ncnn::deconvolution_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static void deconvolution_pack4_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data_packed, const Mat& bias_data, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int activation_type, const Mat& activation_params, const Option& opt)
{
int outch = top_blob.c;
... | subq $0x2848, %rsp # imm = 0x2848
movq 0x2880(%rsp), %rax
movq 0x2878(%rsp), %rax
movl 0x2870(%rsp), %eax
movl 0x2868(%rsp), %eax
movl 0x2860(%rsp), %eax
movl 0x2858(%rsp), %eax
movl 0x2850(%rsp), %eax
movq %rdi, 0x390(%rsp)
movq %rsi, 0x388(%rsp)
movq %rdx, 0x380(%rsp)
movq %rcx, 0x378(%rsp)
movl %r8d, 0x374... | /ysh329[P]ncnn/src/layer/x86/deconvolution_pack4.h |
ncnn::deconvolution_pack4to1_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Mat const&, ncnn::Option const&) | static void deconvolution_pack4to1_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& weight_data_packed, const Mat& bias_data, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int activation_type, const Mat& activation_params, const Option& opt)
{
int outch = top_blob.c;
... | subq $0x698, %rsp # imm = 0x698
movq 0x6d0(%rsp), %rax
movq 0x6c8(%rsp), %rax
movl 0x6c0(%rsp), %eax
movl 0x6b8(%rsp), %eax
movl 0x6b0(%rsp), %eax
movl 0x6a8(%rsp), %eax
movl 0x6a0(%rsp), %eax
movq %rdi, 0x268(%rsp)
movq %rsi, 0x260(%rsp)
movq %rdx, 0x258(%rsp)
movq %rcx, 0x250(%rsp)
movl %r8d, 0x24c(%rsp)
m... | /ysh329[P]ncnn/src/layer/x86/deconvolution_pack4to1.h |
virtual thunk to ncnn::Deconvolution_x86::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Deconvolution_x86::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
// deconvolv with NxN kernel
// value = value + bias
int w = bottom_blob.w;
int h = bottom_blob.h;
size_t elemsize = bottom_blob.elemsize;
int elempack = bottom_blob.elempack;
// NCNN_LOGE(... | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq %rcx, -0x20(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x48(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
movq -0x18(%rsp), %rdx
movq -0x20(%rsp), %rcx
jmp 0x944bf0
nopw (%rax,%rax)
| /ysh329[P]ncnn/src/layer/x86/deconvolution_x86.cpp |
virtual thunk to ncnn::Deconvolution_x86_avx512::create_pipeline(ncnn::Option const&) | int Deconvolution_x86_avx512::create_pipeline(const Option& opt)
{
const int maxk = kernel_w * kernel_h;
int num_input = weight_data_size / maxk / num_output;
Mat weight_data_transposed(weight_data.w);
{
float* pt = weight_data_transposed;
const float* p = weight_data;
for (int... | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x30(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
jmp 0x950790
nopw %cs:(%rax,%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/deconvolution_x86_avx512.cpp |
ncnn::Deconvolution_x86_avx512::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Deconvolution_x86_avx512::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
// deconvolv with NxN kernel
// value = value + bias
int w = bottom_blob.w;
int h = bottom_blob.h;
size_t elemsize = bottom_blob.elemsize;
int elempack = bottom_blob.elempack;
// NCN... | pushq %rbp
pushq %r15
pushq %r14
pushq %rbx
subq $0x748, %rsp # imm = 0x748
movq %rdi, 0x3a0(%rsp)
movq %rsi, 0x398(%rsp)
movq %rdx, 0x390(%rsp)
movq %rcx, 0x388(%rsp)
movq 0x3a0(%rsp), %rcx
movq %rcx, 0x1b0(%rsp)
movq 0x398(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x384(%rsp)
movq 0x398(%rsp), %rax
movl... | /ysh329[P]ncnn/build_O0/src/layer/x86/deconvolution_x86_avx512.cpp |
ncnn::Dropout_x86_avx::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int Dropout_x86_avx::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
if (scale == 1.f)
{
return 0;
}
#if __SSE2__
int dims = bottom_top_blob.dims;
int elempack = bottom_top_blob.elempack;
#if __AVX__
#if __AVX512F__
if (elempack == 16)
{
Mat tmp;
co... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x760, %rsp # imm = 0x760
movq %rdi, 0x2b8(%rsp)
movq %rsi, 0x2b0(%rsp)
movq %rdx, 0x2a8(%rsp)
movq 0x2b8(%rsp), %rax
movq %rax, 0x58(%rsp)
movq (%rax), %rcx
movq -0x18(%rcx), %rcx
vmovss 0xd0(%rax,%rcx), %xmm0
vmovss 0x14243e4(%rip), %xmm1 # 0x1e002d0
vuco... | /ysh329[P]ncnn/build_O0/src/layer/x86/dropout_x86_avx.cpp |
virtual thunk to ncnn::Dropout_x86_avx::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int Dropout_x86_avx::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
if (scale == 1.f)
{
return 0;
}
#if __SSE2__
int dims = bottom_top_blob.dims;
int elempack = bottom_top_blob.elempack;
#if __AVX__
#if __AVX512F__
if (elempack == 16)
{
Mat tmp;
co... | 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 0x9dbea0
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/dropout_x86_avx.cpp |
ncnn::ELU_x86_avx::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int ELU_x86_avx::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int channels = bottom_top_blob.c;
int elempack = bottom_top_blob.elempack;
int size = w * h * elempack;
#pragma omp parallel for num_threads(opt.num_threads)
... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x1960, %rsp # imm = 0x1960
movq %rdi, 0x130(%rsp)
movq %rsi, 0x128(%rsp)
movq %rdx, 0x120(%rsp)
movq 0x130(%rsp), %rax
movq %rax, 0x38(%rsp)
movq 0x128(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x11c(%rsp)
movq 0x128(%rsp), %rax
movl 0x30(%rax), %eax
movl... | /ysh329[P]ncnn/build_O0/src/layer/x86/elu_x86_avx.cpp |
int ncnn::binary_op_scalar_inplace<ncnn::BinaryOp_x86_fma_functor::binary_op_rsub>(ncnn::Mat&, float, ncnn::Option const&) | static int binary_op_scalar_inplace(Mat& a, float b, 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... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x2e0, %rsp # imm = 0x2E0
movq %rdi, 0x140(%rsp)
vmovss %xmm0, 0x13c(%rsp)
movq %rsi, 0x130(%rsp)
movq 0x140(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x128(%rsp)
movq 0x140(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x124(%rsp)
movq 0x140(%rsp), %rax
m... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
virtual thunk to ncnn::BinaryOp_x86_fma::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int BinaryOp_x86_fma::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace BinaryOp_x86_fma_functor;
if (op_type == Operation_ADD)
return binary_op_scalar_inplace<binary_op_add>(bottom_top_blob, b, opt);
if (op_type == Operation_SUB)
return binary_op_scalar_inpl... | 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 0x15243d0
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_fma_functor::binary_op_add>(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.... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x560, %rsp # imm = 0x560
movq %rdi, 0x270(%rsp)
movq %rsi, 0x268(%rsp)
movq %rdx, 0x260(%rsp)
movq %rcx, 0x258(%rsp)
movq 0x270(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x250(%rsp)
movq 0x270(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x24c(%rsp)
movq... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_6_11_16_25<ncnn::BinaryOp_x86_fma_functor::binary_op_add>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_fma_functor::binary_op_add>(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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_fma_functor::binary_op_sub>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_fma_functor::binary_op_mul>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_fma_functor::binary_op_div>(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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_fma_functor::binary_op_max>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_fma.cpp |
int ncnn::binary_op_2_3_4_20<ncnn::BinaryOp_x86_avx_functor::binary_op_mul>(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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
int ncnn::binary_op_6_11_16_25<ncnn::BinaryOp_x86_avx_functor::binary_op_div>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
int ncnn::binary_op_7_13_19_29<ncnn::BinaryOp_x86_avx_functor::binary_op_max>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
int ncnn::binary_op_6_11_16_25<ncnn::BinaryOp_x86_avx_functor::binary_op_max>(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.... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
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.empt... | 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... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
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 %x... | /ysh329[P]ncnn/build_O0/src/layer/x86/binaryop_x86_avx.cpp |
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++)
{
... | 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(%ra... | /ysh329[P]ncnn/src/layer/x86/unaryop_x86.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_floor::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_floor_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const _... | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x18(%rsp), %rax
vmovaps (%rax), %xmm0
vroundps $0x1, %xmm0, %xmm0
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_ceil::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_ceil_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const __... | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x18(%rsp), %rax
vmovaps (%rax), %xmm0
vroundps $0x2, %xmm0, %xmm0
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_exp::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return exp_ps(x);
} | subq $0x4e8, %rsp # imm = 0x4E8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, (%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x30(%rsp)
vmovaps 0x30(%rsp), %xmm0
vmovaps %xmm0, -0x10(%rsp)
vmovaps 0x7a8df3(%rip), %xmm0 # 0x1e02ab0
vmovaps %xmm0,... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_log::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return log_ps(x);
} | subq $0x6b8, %rsp # imm = 0x6B8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x10(%rsp)
vmovaps 0x7a7958(%rip), %xmm0 # 0x1e02ab0
vmovaps %xmm0, -0x10(%rsp)
vmovaps 0x10(%rsp), %xmm1
vxorps %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x60(%rsp)
vmovaps 0x... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_sin::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return sin256_ps(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xec0, %rsp # imm = 0xEC0
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x240(%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x260(%rsp)
vmovaps 0x260(%rsp), %ymm0
vmovaps %ymm0, 0x200(%rsp)
vmovaps 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_sin::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return sin_ps(x);
} | subq $0x678, %rsp # imm = 0x678
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x40(%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x70(%rsp)
vmovaps 0x70(%rsp), %xmm0
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_cos::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return cos256_ps(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xda0, %rsp # imm = 0xDA0
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x220(%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x240(%rsp)
vmovaps 0x240(%rsp), %ymm0
vmovaps %ymm0, 0x1e0(%rsp)
vmovaps 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_tan::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return tan256_ps(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x13a0, %rsp # imm = 0x13A0
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
movl $0x322bcc77, 0x1bc(%rsp) # imm = 0x322BCC77
vmovss 0x1bc(%rsp), %xmm0
vmovss %xmm0, 0x138c(%rsp)
vmovss %xmm0, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_functor::unary_op_tan::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return tan_ps(x);
} | subq $0x8d8, %rsp # imm = 0x8D8
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, -0x10(%rsp)
movl $0x322bcc77, 0x1c(%rsp) # imm = 0x322BCC77
vbroadcastss 0x1c(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps (%rsp), %xmm0
vmovaps %xmm0, -0x40(%rsp)
vmovaps -... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_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 0x16314b0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_fma_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 $0xf80, %rsp # imm = 0xF80
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... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_fma.cpp |
ncnn::UnaryOp_x86_avx::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int UnaryOp_x86_avx::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_avx_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>(botto... | 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 0x1662b95
movq 0x10(%rsp), %rdi
movq 0x8(%rsp), %rsi
callq 0x1662e70
movl %eax, 0x24(%rsp)
jmp 0x1662e64
movq (%rsp), %rax
mov... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
int ncnn::unary_op_inplace<ncnn::UnaryOp_x86_avx_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++)
{
... | 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, 0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
virtual thunk to ncnn::UnaryOp_x86_avx::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int UnaryOp_x86_avx::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
using namespace UnaryOp_x86_avx_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>(botto... | 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 0x1662b50
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::UnaryOp_x86_avx_functor::unary_op_floor::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
#if __SSE4_1__
return _mm_floor_ps(x);
#endif // __SSE4_1__
// Use negative zero as the sign bit mask.
const __m128 magic_negative_zero = _mm_set_ps1(-0.0f);
// The smallest float number that have no fractional part. (2^23)
const _... | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x18(%rsp), %rax
vmovaps (%rax), %xmm0
vroundps $0x1, %xmm0, %xmm0
addq $0x28, %rsp
retq
nopw %cs:(%rax,%rax)
nopl (%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::UnaryOp_x86_avx_functor::unary_op_exp::func_pack8(float vector[8] const&) const | __m256 func_pack8(const __m256& x) const
{
return exp256_ps(x);
} | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xd80, %rsp # imm = 0xD80
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x10(%rsp), %rax
vmovaps (%rax), %ymm0
vmovaps %ymm0, 0x100(%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %ymm0, 0x1a0(%rsp)
vmovaps 0x1a0(%rsp), %ymm0
vmovaps %ymm0, 0xe0(%rsp)
vmovaps 0x... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::UnaryOp_x86_avx_functor::unary_op_sin::func_pack4(float vector[4] const&) const | __m128 func_pack4(const __m128& x) const
{
return sin_ps(x);
} | subq $0x708, %rsp # imm = 0x708
movq %rdi, -0x78(%rsp)
movq %rsi, -0x80(%rsp)
movq -0x80(%rsp), %rax
vmovaps (%rax), %xmm0
vmovaps %xmm0, 0x40(%rsp)
vxorps %xmm0, %xmm0, %xmm0
vmovaps %xmm0, 0x70(%rsp)
vmovaps 0x70(%rsp), %xmm0
vmovaps %xmm0, 0x20(%rsp)
vmovaps 0x40(%rsp), %xmm0
vmovaps %xmm0, (%rsp)
vmovaps... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::UnaryOp_x86_avx_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 0x1631460
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::UnaryOp_x86_avx_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 0x16314b0
vmovss %xmm0, (%rsp)
vmovss 0x4(%rsp), %xmm0... | /ysh329[P]ncnn/build_O0/src/layer/x86/unaryop_x86_avx.cpp |
ncnn::ConvolutionDepthWise_x86::create_group_ops(ncnn::Option const&) | int ConvolutionDepthWise_x86::create_group_ops(const Option& opt)
{
// create Convolution op for each group
const int maxk = kernel_w * kernel_h;
int channels = (weight_data_size / group) / maxk / (num_output / group) * group;
for (int i = 0; i < (int)group_ops.size(); i++)
delete group_ops[i];... | subq $0x1058, %rsp # imm = 0x1058
movq %rdi, 0x868(%rsp)
movq %rsi, 0x860(%rsp)
movq 0x868(%rsp), %rcx
movq %rcx, 0x288(%rsp)
movq (%rcx), %rax
movq -0x18(%rax), %rax
movl 0xd4(%rcx,%rax), %eax
movq (%rcx), %rdx
movq -0x18(%rdx), %rdx
imull 0xd8(%rcx,%rdx), %eax
movl %eax, 0x85c(%rsp)
movq (%rcx), %rax
movq -... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_x86.cpp |
virtual thunk to ncnn::ConvolutionDepthWise_x86::destroy_pipeline(ncnn::Option const&) | int ConvolutionDepthWise_x86::destroy_pipeline(const Option& opt)
{
if (activation)
{
activation->destroy_pipeline(opt);
delete activation;
activation = 0;
}
for (int i = 0; i < (int)group_ops.size(); i++)
{
group_ops[i]->destroy_pipeline(opt);
delete group_o... | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x38(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
jmp 0x16875a0
nopw %cs:(%rax,%rax)
| /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_x86.cpp |
ncnn::convdw5x5s1_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | static void convdw5x5s1_pack4_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& kernel, const Mat& _bias, const Option& opt)
{
int w = bottom_blob.w;
int outw = top_blob.w;
int outh = top_blob.h;
const int group = bottom_blob.c;
const float* bias = _bias;
#pragma omp parallel for num_thr... | subq $0x3428, %rsp # imm = 0x3428
movq %rdi, 0xd78(%rsp)
movq %rsi, 0xd70(%rsp)
movq %rdx, 0xd68(%rsp)
movq %rcx, 0xd60(%rsp)
movq %r8, 0xd58(%rsp)
movq 0xd78(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xd54(%rsp)
movq 0xd70(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xd50(%rsp)
movq 0xd70(%rsp), %rax
mov... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_5x5_pack4.h |
ncnn::convdw3x3s1_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | static void convdw3x3s1_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& _kernel, const Mat& _bias, const Option& opt)
{
int w = bottom_blob.w;
int outw = top_blob.w;
int outh = top_blob.h;
const int group = bottom_blob.c;
const float* kernel = _kernel;
const float* bias = _bias;
#p... | subq $0x368, %rsp # imm = 0x368
movq %rdi, 0x1c0(%rsp)
movq %rsi, 0x1b8(%rsp)
movq %rdx, 0x1b0(%rsp)
movq %rcx, 0x1a8(%rsp)
movq %r8, 0x1a0(%rsp)
movq 0x1c0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x19c(%rsp)
movq 0x1b8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x198(%rsp)
movq 0x1b8(%rsp), %rax
movl... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_3x3.h |
virtual thunk to ncnn::ConvolutionDepthWise_x86::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int ConvolutionDepthWise_x86::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
#if NCNN_INT8
if (opt.use_int8_inference && int8_scale_term)
{
return forward_int8_x86(bottom_blob, top_blob, opt);
}
#endif
int w = bottom_blob.w;
int h = bottom_blob.h;
int channels... | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq %rdx, -0x18(%rsp)
movq %rcx, -0x20(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x48(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
movq -0x18(%rsp), %rdx
movq -0x20(%rsp), %rcx
jmp 0x16876d0
nopw (%rax,%rax)
| /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_x86.cpp |
ncnn::convdw3x3s1_int8_dequant_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, std::vector<float, std::allocator<float>>, ncnn::Option const&) | static void convdw3x3s1_int8_dequant_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& _kernel, const Mat& _bias, std::vector<float> scales_dequant, const Option& opt)
{
int w = bottom_blob.w;
//int h = bottom_blob.h;
//int inch = bottom_blob.c;
int outw = top_blob.w;
int outh = top_blob.h;
... | subq $0x368, %rsp # imm = 0x368
movq %r8, 0x58(%rsp)
movq %rdi, 0x198(%rsp)
movq %rsi, 0x190(%rsp)
movq %rdx, 0x188(%rsp)
movq %rcx, 0x180(%rsp)
movq %r8, 0x178(%rsp)
movq %r9, 0x170(%rsp)
movq 0x198(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x16c(%rsp)
movq 0x190(%rsp), %rax
movl 0x2c(%rax), %eax
movl %e... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_3x3_int8.h |
ncnn::convdw3x3s2_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | static void convdw3x3s2_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& _kernel, const Mat& _bias, const Option& opt)
{
int w = bottom_blob.w;
int outw = top_blob.w;
int outh = top_blob.h;
const int group = bottom_blob.c;
const int tailstep = w - 2 * outw + w;
const float* kernel = _ke... | subq $0x358, %rsp # imm = 0x358
movq %rdi, 0x1b0(%rsp)
movq %rsi, 0x1a8(%rsp)
movq %rdx, 0x1a0(%rsp)
movq %rcx, 0x198(%rsp)
movq %r8, 0x190(%rsp)
movq 0x1b0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x18c(%rsp)
movq 0x1a8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x188(%rsp)
movq 0x1a8(%rsp), %rax
movl... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_3x3.h |
ncnn::convdw3x3s1_int8_requant_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, std::vector<float, std::allocator<float>>, ncnn::Option const&) | static void convdw3x3s1_int8_requant_sse(const Mat& bottom_blob, Mat& top_blob, const Mat& _kernel, const Mat& _bias, std::vector<float> scales_requant, const Option& opt)
{
int w = bottom_blob.w;
//int h = bottom_blob.h;
//int inch = bottom_blob.c;
int outw = top_blob.w;
int outh = top_blob.h;
... | subq $0x358, %rsp # imm = 0x358
movq %r8, 0x60(%rsp)
movq %rdi, 0x1a0(%rsp)
movq %rsi, 0x198(%rsp)
movq %rdx, 0x190(%rsp)
movq %rcx, 0x188(%rsp)
movq %r8, 0x180(%rsp)
movq %r9, 0x178(%rsp)
movq 0x1a0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x174(%rsp)
movq 0x198(%rsp), %rax
movl 0x2c(%rax), %eax
movl %e... | /ysh329[P]ncnn/src/layer/x86/convolutiondepthwise_3x3_int8.h |
ncnn::ConvolutionDepthWise_x86_avx::forward(std::vector<ncnn::Mat, std::allocator<ncnn::Mat>> const&, std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int ConvolutionDepthWise_x86_avx::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& top_blobs, const Option& opt) const
{
const Mat& bottom_blob = bottom_blobs[0];
const Mat& _weight_data = bottom_blobs[1];
Mat& top_blob = top_blobs[0];
const int _kernel_w = _weight_data.w;
const int ... | subq $0x448, %rsp # imm = 0x448
movq %rdi, 0x2b0(%rsp)
movq %rsi, 0x2a8(%rsp)
movq %rdx, 0x2a0(%rsp)
movq %rcx, 0x298(%rsp)
movq 0x2b0(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0x2a8(%rsp), %rdi
xorl %eax, %eax
movl %eax, %esi
movq %rsi, 0x100(%rsp)
callq 0xb5820
movq %rax, 0x290(%rsp)
movq 0x2a8(%rsp), %rdi
mo... | /ysh329[P]ncnn/build_O0/src/layer/x86/convolutiondepthwise_x86_avx.cpp |
ncnn::resize_bicubic_image_pack4(ncnn::Mat const&, ncnn::Mat&, float*, int*, float*, int*) | static void resize_bicubic_image_pack4(const Mat& src, Mat& dst, float* alpha, int* xofs, float* beta, int* yofs)
{
int w = dst.w;
int h = dst.h;
// loop body
Mat rowsbuf0(w, (size_t)4 * 4u, 4);
Mat rowsbuf1(w, (size_t)4 * 4u, 4);
Mat rowsbuf2(w, (size_t)4 * 4u, 4);
Mat rowsbuf3(w, (size_t)... | subq $0x1fc8, %rsp # imm = 0x1FC8
movq %rdi, 0xa48(%rsp)
movq %rsi, 0xa40(%rsp)
movq %rdx, 0xa38(%rsp)
movq %rcx, 0xa30(%rsp)
movq %r8, 0xa28(%rsp)
movq %r9, 0xa20(%rsp)
movq 0xa40(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xa1c(%rsp)
movq 0xa40(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xa18(%rsp)
movl... | /ysh329[P]ncnn/src/layer/x86/interp_bicubic_pack4.h |
ncnn::resize_bilinear_image(ncnn::Mat const&, ncnn::Mat&, float*, int*, float*, int*) | static void resize_bilinear_image(const Mat& src, Mat& dst, float* alpha, int* xofs, float* beta, int* yofs)
{
int w = dst.w;
int h = dst.h;
// loop body
Mat rowsbuf0(w);
Mat rowsbuf1(w);
float* rows0 = rowsbuf0;
float* rows1 = rowsbuf1;
int prev_sy1 = -2;
for (int dy = 0; dy < h;... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x800, %rsp # imm = 0x800
movq %rdi, 0x3f0(%rsp)
movq %rsi, 0x3e8(%rsp)
movq %rdx, 0x3e0(%rsp)
movq %rcx, 0x3d8(%rsp)
movq %r8, 0x3d0(%rsp)
movq %r9, 0x3c8(%rsp)
movq 0x3e8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x3c4(%rsp)
movq 0x3e8(%rsp), %rax
movl ... | /ysh329[P]ncnn/src/layer/x86/interp_bilinear.h |
ncnn::resize_bicubic_image(ncnn::Mat const&, ncnn::Mat&, float*, int*, float*, int*) | static void resize_bicubic_image(const Mat& src, Mat& dst, float* alpha, int* xofs, float* beta, int* yofs)
{
int w = dst.w;
int h = dst.h;
// loop body
Mat rowsbuf0(w);
Mat rowsbuf1(w);
Mat rowsbuf2(w);
Mat rowsbuf3(w);
float* rows0 = rowsbuf0;
float* rows1 = rowsbuf1;
float* r... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xfa0, %rsp # imm = 0xFA0
movq %rdi, 0x738(%rsp)
movq %rsi, 0x730(%rsp)
movq %rdx, 0x728(%rsp)
movq %rcx, 0x720(%rsp)
movq %r8, 0x718(%rsp)
movq %r9, 0x710(%rsp)
movq 0x730(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x70c(%rsp)
movq 0x730(%rsp), %rax
movl ... | /ysh329[P]ncnn/src/layer/x86/interp_bicubic.h |
ncnn::interpolate_cubic(float, float*) | static inline void interpolate_cubic(float fx, float* coeffs)
{
const float A = -0.75f;
float fx0 = fx + 1;
float fx1 = fx;
float fx2 = 1 - fx;
// float fx3 = 2 - fx;
coeffs[0] = A * fx0 * fx0 * fx0 - 5 * A * fx0 * fx0 + 8 * A * fx0 - 4 * A;
coeffs[1] = (A + 2) * fx1 * fx1 * fx1 - (A + 3) ... | vmovss %xmm0, -0x4(%rsp)
movq %rdi, -0x10(%rsp)
vmovss 0x5c0fcd(%rip), %xmm0 # 0x1e16fb0
vmovss %xmm0, -0x14(%rsp)
vmovss 0x5aa2df(%rip), %xmm0 # 0x1e002d0
vaddss -0x4(%rsp), %xmm0, %xmm0
vmovss %xmm0, -0x18(%rsp)
vmovss -0x4(%rsp), %xmm0
vmovss %xmm0, -0x1c(%rsp)
vmovss 0x5aa2bf(%rip), %xmm0 # 0x1e002d0
vsubss -... | /ysh329[P]ncnn/src/layer/x86/interp_bicubic.h |
ncnn::linear_coeffs(int, int, int*, float*, int) | static void linear_coeffs(int w, int outw, int* xofs, float* alpha, int align_corner)
{
double scale = (double)w / outw;
if (align_corner)
{
scale = (double)(w - 1) / (outw - 1);
}
for (int dx = 0; dx < outw; dx++)
{
float fx = (float)((dx + 0.5) * scale - 0.5);
if (alig... | subq $0x38, %rsp
movl %edi, 0x34(%rsp)
movl %esi, 0x30(%rsp)
movq %rdx, 0x28(%rsp)
movq %rcx, 0x20(%rsp)
movl %r8d, 0x1c(%rsp)
vcvtsi2sdl 0x34(%rsp), %xmm0, %xmm0
vcvtsi2sdl 0x30(%rsp), %xmm1, %xmm1
vdivsd %xmm1, %xmm0, %xmm0
vmovsd %xmm0, 0x10(%rsp)
cmpl $0x0, 0x1c(%rsp)
je 0x1861888
movl 0x34(%rsp), %eax
subl $0x1, %... | /ysh329[P]ncnn/src/layer/x86/interp_bilinear.h |
ncnn::YoloDetectionOutput::destroy_pipeline(ncnn::Option const&) | int YoloDetectionOutput::destroy_pipeline(const Option& opt)
{
if (softmax)
{
softmax->destroy_pipeline(opt);
delete softmax;
softmax = 0;
}
return 0;
} | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x20(%rsp), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x128(%rax)
je 0x18b774b
movq 0x10(%rsp), %rax
movq 0x128(%rax), %rdi
movq 0x18(%rsp), %rsi
movq (%rdi), %rax
callq *0x28(%rax)
movq 0x10(%rsp), %rax
movq 0x128(%rax), %rax
movq %rax, 0x8(%rsp)
cmpq $0x0,... | /ysh329[P]ncnn/src/layer/yolodetectionoutput.cpp |
ncnn::YoloDetectionOutput::forward_inplace(std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int YoloDetectionOutput::forward_inplace(std::vector<Mat>& bottom_top_blobs, const Option& opt) const
{
// gather all box
std::vector<BBoxRect> all_bbox_rects;
std::vector<float> all_bbox_scores;
for (size_t b = 0; b < bottom_top_blobs.size(); b++)
{
Mat& bottom_top_blob = bottom_top_blobs[... | subq $0xba8, %rsp # imm = 0xBA8
movq %rdi, 0x5d8(%rsp)
movq %rsi, 0x5d0(%rsp)
movq %rdx, 0x5c8(%rsp)
movq 0x5d8(%rsp), %rax
movq %rax, 0x1b0(%rsp)
leaq 0x5b0(%rsp), %rdi
callq 0x17fd480
leaq 0x598(%rsp), %rdi
callq 0xf86d00
movq $0x0, 0x590(%rsp)
movq 0x590(%rsp), %rax
movq %rax, 0x1a8(%rsp)
movq 0x5d0(%rsp)... | /ysh329[P]ncnn/src/layer/yolodetectionoutput.cpp |
ncnn::nms_sorted_bboxes(std::vector<ncnn::BBoxRect, std::allocator<ncnn::BBoxRect>> const&, std::vector<unsigned long, std::allocator<unsigned long>>&, float) | static void nms_sorted_bboxes(const std::vector<BBoxRect>& bboxes, std::vector<size_t>& picked, float nms_threshold)
{
picked.clear();
const size_t n = bboxes.size();
std::vector<float> areas(n);
for (size_t i = 0; i < n; i++)
{
const BBoxRect& r = bboxes[i];
float width = r.xmax ... | subq $0xc8, %rsp
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movss %xmm0, 0xb4(%rsp)
movq 0xb8(%rsp), %rdi
callq 0xf872d0
movq 0xc0(%rsp), %rdi
callq 0x17fd500
movq %rax, 0xa8(%rsp)
movq 0xa8(%rsp), %rax
movq %rax, 0x28(%rsp)
leaq 0x8f(%rsp), %rdi
movq %rdi, 0x30(%rsp)
callq 0xf872f0
movq 0x28(%rsp), %rsi
movq 0x30(%rs... | /ysh329[P]ncnn/src/layer/yolodetectionoutput.cpp |
ncnn::intersection_area(ncnn::BBoxRect const&, ncnn::BBoxRect const&) | static inline float intersection_area(const BBoxRect& a, const BBoxRect& b)
{
if (a.xmin > b.xmax || a.xmax < b.xmin || a.ymin > b.ymax || a.ymax < b.ymin)
{
// no intersection
return 0.f;
}
float inter_width = std::min(a.xmax, b.xmax) - std::max(a.xmin, b.xmin);
float inter_height ... | subq $0x28, %rsp
movq %rdi, 0x18(%rsp)
movq %rsi, 0x10(%rsp)
movq 0x18(%rsp), %rax
movss (%rax), %xmm0
movq 0x10(%rsp), %rax
ucomiss 0x8(%rax), %xmm0
ja 0x18ba898
movq 0x18(%rsp), %rax
movss 0x8(%rax), %xmm1
movq 0x10(%rsp), %rax
movss (%rax), %xmm0
ucomiss %xmm1, %xmm0
ja 0x18ba898
movq 0x18(%rsp), %rax
movss 0x4(%rax... | /ysh329[P]ncnn/src/layer/yolodetectionoutput.cpp |
void ncnn::qsort_descent_inplace<ncnn::BBoxRect>(std::vector<ncnn::BBoxRect, std::allocator<ncnn::BBoxRect>>&, std::vector<float, std::allocator<float>>&, int, int) | static void qsort_descent_inplace(std::vector<T>& datas, std::vector<float>& scores, int left, int right)
{
int i = left;
int j = right;
float p = scores[(left + right) / 2];
while (i <= j)
{
while (scores[i] > p)
i++;
while (scores[j] < p)
j--;
if ... | subq $0x38, %rsp
movq %rdi, 0x30(%rsp)
movq %rsi, 0x28(%rsp)
movl %edx, 0x24(%rsp)
movl %ecx, 0x20(%rsp)
movl 0x24(%rsp), %eax
movl %eax, 0x1c(%rsp)
movl 0x20(%rsp), %eax
movl %eax, 0x18(%rsp)
movq 0x28(%rsp), %rdi
movl 0x24(%rsp), %eax
addl 0x20(%rsp), %eax
movl $0x2, %ecx
cltd
idivl %ecx
movslq %eax, %rsi
callq 0xf86... | /ysh329[P]ncnn/src/layer/yolodetectionoutput.cpp |
ncnn::Quantize::load_model(ncnn::ModelBin const&) | int Quantize::load_model(const ModelBin& mb)
{
scale_data = mb.load(scale_data_size, 1);
if (scale_data.empty())
return -100;
return 0;
} | subq $0x138, %rsp # imm = 0x138
movq %rdi, 0xa0(%rsp)
movq %rsi, 0x98(%rsp)
movq 0xa0(%rsp), %rax
movq %rax, 0x30(%rsp)
movq 0x98(%rsp), %rsi
movl 0xd0(%rax), %edx
movq (%rsi), %rax
leaq 0x50(%rsp), %rdi
movl $0x1, %ecx
callq *0x10(%rax)
movq 0x30(%rsp), %rax
addq $0xd8, %rax
movq %rax, 0xc0(%rsp)
leaq 0x50(... | /ysh329[P]ncnn/src/layer/quantize.cpp |
ncnn::Quantize::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Quantize::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
int dims = bottom_blob.dims;
if (dims == 1)
{
int w = bottom_blob.w;
top_blob.create(w, (size_t)1u, opt.blob_allocator);
if (top_blob.empty())
return -100;
const float* ptr ... | subq $0x428, %rsp # imm = 0x428
movq %rdi, 0x1c8(%rsp)
movq %rsi, 0x1c0(%rsp)
movq %rdx, 0x1b8(%rsp)
movq %rcx, 0x1b0(%rsp)
movq 0x1c8(%rsp), %rax
movq %rax, 0x90(%rsp)
movq 0x1c0(%rsp), %rax
movl 0x28(%rax), %eax
movl %eax, 0x1ac(%rsp)
cmpl $0x1, 0x1ac(%rsp)
jne 0x18bb3dc
movq 0x1c0(%rsp), %rax
movl 0x2c(%r... | /ysh329[P]ncnn/src/layer/quantize.cpp |
ncnn::Dequantize::Dequantize() | Dequantize::Dequantize()
{
one_blob_only = true;
support_inplace = false;
} | subq $0x68, %rsp
movq %rdi, 0x30(%rsp)
movq 0x30(%rsp), %rdi
movq %rdi, 0x10(%rsp)
callq 0xad350
movq 0x10(%rsp), %rax
leaq 0x5c84c4(%rip), %rcx # 0x1eca048
addq $0x10, %rcx
movq %rcx, (%rax)
addq $0xd8, %rax
movq %rax, 0x38(%rsp)
movq 0x38(%rsp), %rax
movq %rax, 0x18(%rsp)
movq $0x0, (%rax)
movq $0x0, 0x8(%rax)
mov... | /ysh329[P]ncnn/src/layer/dequantize.cpp |
ncnn::Dequantize::load_model(ncnn::ModelBin const&) | int Dequantize::load_model(const ModelBin& mb)
{
scale_data = mb.load(scale_data_size, 1);
if (scale_data.empty())
return -100;
if (bias_data_size)
{
bias_data = mb.load(bias_data_size, 1);
if (bias_data.empty())
return -100;
}
return 0;
} | subq $0x238, %rsp # imm = 0x238
movq %rdi, 0x118(%rsp)
movq %rsi, 0x110(%rsp)
movq 0x118(%rsp), %rax
movq %rax, 0x60(%rsp)
movq 0x110(%rsp), %rsi
movl 0xd0(%rax), %edx
movq (%rsi), %rax
leaq 0xc8(%rsp), %rdi
movl $0x1, %ecx
callq *0x10(%rax)
movq 0x60(%rsp), %rax
addq $0xd8, %rax
movq %rax, 0x178(%rsp)
leaq ... | /ysh329[P]ncnn/src/layer/dequantize.cpp |
ncnn::Requantize::load_model(ncnn::ModelBin const&) | int Requantize::load_model(const ModelBin& mb)
{
scale_in_data = mb.load(scale_in_data_size, 1);
if (scale_in_data.empty())
return -100;
scale_out_data = mb.load(scale_out_data_size, 1);
if (scale_out_data.empty())
return -100;
if (bias_data_size)
{
bias_data = mb.load(... | subq $0x338, %rsp # imm = 0x338
movq %rdi, 0x190(%rsp)
movq %rsi, 0x188(%rsp)
movq 0x190(%rsp), %rax
movq %rax, 0x90(%rsp)
movq 0x188(%rsp), %rsi
movl 0xd0(%rax), %edx
movq (%rsi), %rax
leaq 0x140(%rsp), %rdi
movl $0x1, %ecx
callq *0x10(%rax)
movq 0x90(%rsp), %rax
addq $0x128, %rax # imm = 0x128
m... | /ysh329[P]ncnn/src/layer/requantize.cpp |
ncnn::float32_to_int8(float) | signed char float32_to_int8(float value)
{
float tmp;
if (value >= 0.f)
tmp = value + 0.5f;
else
tmp = value - 0.5f;
if (tmp > 127)
return 127;
if (tmp < -128)
return -128;
return static_cast<signed char>(tmp);
} | movss %xmm0, -0x8(%rsp)
movss -0x8(%rsp), %xmm0
xorps %xmm1, %xmm1
ucomiss %xmm1, %xmm0
jb 0x1ca1fba
movss 0x15e328(%rip), %xmm0 # 0x1e002d4
addss -0x8(%rsp), %xmm0
movss %xmm0, -0xc(%rsp)
jmp 0x1ca1fd2
movss -0x8(%rsp), %xmm0
movss 0x15e30c(%rip), %xmm1 # 0x1e002d4
subss %xmm1, %xmm0
movss %xmm0, -0xc(%rsp)
movss ... | /ysh329[P]ncnn/src/layer/cast.cpp |
ncnn::cast_fp32_to_fp16_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void cast_fp32_to_fp16_sse(const Mat& bottom_blob, Mat& top_blob, const Option& opt)
{
#if NCNN_RUNTIME_CPU && NCNN_AVX512FP16 && __AVX512F__ && !__AVX512FP16__
if (ncnn::cpu_support_x86_avx512_fp16())
{
cast_fp32_to_fp16_sse_avx512fp16(bottom_blob, top_blob, opt);
return;
}
#endif
#... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x380, %rsp # imm = 0x380
movq %rdi, 0x190(%rsp)
movq %rsi, 0x188(%rsp)
movq %rdx, 0x180(%rsp)
movq 0x190(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x17c(%rsp)
movq 0x190(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x178(%rsp)
movq 0x190(%rsp), %rax
movl... | /ysh329[P]ncnn/src/layer/x86/cast_fp16.h |
ncnn::GRU::GRU() | GRU::GRU()
{
one_blob_only = false;
support_inplace = false;
} | subq $0x98, %rsp
movq %rdi, 0x38(%rsp)
movq 0x38(%rsp), %rdi
movq %rdi, 0x18(%rsp)
callq 0xad350
movq 0x18(%rsp), %rax
leaq 0x1a4531(%rip), %rcx # 0x1ecd868
addq $0x10, %rcx
movq %rcx, (%rax)
addq $0xe0, %rax
movq %rax, 0x40(%rsp)
movq 0x40(%rsp), %rax
movq %rax, 0x20(%rsp)
movq $0x0, (%rax)
movq $0x0, 0x8(%rax)
mov... | /ysh329[P]ncnn/src/layer/gru.cpp |
ncnn::Convolution1D::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Convolution1D::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
Mat bottom_blob_bordered;
make_padding(bottom_blob, bottom_blob_bordered, opt);
if (bottom_blob_bordered.empty())
return -100;
const int w = bottom_blob_bordered.w;
const size_t elemsize = bottom_bl... | pushq %rbx
subq $0x180, %rsp # imm = 0x180
movq %rdi, 0x108(%rsp)
movq %rsi, 0x100(%rsp)
movq %rdx, 0xf8(%rsp)
movq %rcx, 0xf0(%rsp)
movq 0x108(%rsp), %rdi
movq %rdi, 0x68(%rsp)
leaq 0xa8(%rsp), %rdx
movq %rdx, 0x118(%rsp)
movq 0x118(%rsp), %rax
movq %rax, 0x70(%rsp)
movq $0x0, (%rax)
movq $0x0, 0x8(%rax)
mo... | /ysh329[P]ncnn/src/layer/convolution1d.cpp |
ncnn::Convolution1D_x86_fma::forward(std::vector<ncnn::Mat, std::allocator<ncnn::Mat>> const&, std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int Convolution1D_x86_fma::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& top_blobs, const Option& opt) const
{
const Mat& bottom_blob = bottom_blobs[0];
const Mat& _weight_data = bottom_blobs[1];
Mat& top_blob = top_blobs[0];
const int _kernel_w = _weight_data.w;
const int _num_ou... | subq $0x448, %rsp # imm = 0x448
movq %rdi, 0x2b0(%rsp)
movq %rsi, 0x2a8(%rsp)
movq %rdx, 0x2a0(%rsp)
movq %rcx, 0x298(%rsp)
movq 0x2b0(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0x2a8(%rsp), %rdi
xorl %eax, %eax
movl %eax, %esi
movq %rsi, 0x100(%rsp)
callq 0xb5820
movq %rax, 0x290(%rsp)
movq 0x2a8(%rsp), %rdi
mo... | /ysh329[P]ncnn/build_O0/src/layer/x86/convolution1d_x86_fma.cpp |
ncnn::Pooling3D::load_param(ncnn::ParamDict const&) | int Pooling3D::load_param(const ParamDict& pd)
{
pooling_type = pd.get(0, 0);
kernel_w = pd.get(1, 0);
kernel_h = pd.get(11, kernel_w);
kernel_d = pd.get(21, kernel_w);
stride_w = pd.get(2, 1);
stride_h = pd.get(12, stride_w);
stride_d = pd.get(22, stride_w);
pad_left = pd.get(3, 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 0xa14b0
movl %eax, %ecx
movq (%rsp), %rax
movl %ecx, 0xd0(%rax)
movq 0x8(%rsp), %rdi
movl $0x1, %esi
xorl %edx, %edx
callq 0xa14b0
movl %eax, %ecx
movq (%rsp), %... | /ysh329[P]ncnn/src/layer/pooling3d.cpp |
ncnn::DeformableConv2D::load_param(ncnn::ParamDict const&) | int DeformableConv2D::load_param(const ParamDict& pd)
{
num_output = pd.get(0, 0);
kernel_w = pd.get(1, 0);
kernel_h = pd.get(11, kernel_w);
dilation_w = pd.get(2, 1);
dilation_h = pd.get(12, dilation_w);
stride_w = pd.get(3, 1);
stride_h = pd.get(13, stride_w);
pad_left = pd.get(4, 0);
... | subq $0x1b8, %rsp # imm = 0x1B8
movq %rdi, 0xf0(%rsp)
movq %rsi, 0xe8(%rsp)
movq 0xf0(%rsp), %rax
movq %rax, 0x38(%rsp)
movq 0xe8(%rsp), %rdi
xorl %edx, %edx
movl %edx, 0x34(%rsp)
movl %edx, %esi
callq 0xa14b0
movl 0x34(%rsp), %edx
movl %eax, %ecx
movq 0x38(%rsp), %rax
movl %ecx, 0xd0(%rax)
movq 0xe8(%rsp), ... | /ysh329[P]ncnn/src/layer/deformableconv2d.cpp |
virtual thunk to ncnn::DeformableConv2D_x86_fma::destroy_pipeline(ncnn::Option const&) | int DeformableConv2D_x86_fma::destroy_pipeline(const Option& opt)
{
if (inner_product)
{
inner_product->destroy_pipeline(opt);
delete inner_product;
inner_product = 0;
}
if (permute)
{
permute->destroy_pipeline(opt);
delete permute;
permute = 0;
}
... | movq %rdi, -0x8(%rsp)
movq %rsi, -0x10(%rsp)
movq -0x8(%rsp), %rdi
movq (%rdi), %rax
movq -0x38(%rax), %rax
addq %rax, %rdi
movq -0x10(%rsp), %rsi
jmp 0x1df8080
nopw %cs:(%rax,%rax)
| /ysh329[P]ncnn/build_O0/src/layer/x86/deformableconv2d_x86_fma.cpp |
ncnn::DeformableConv2D_x86_fma::forward(std::vector<ncnn::Mat, std::allocator<ncnn::Mat>> const&, std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int DeformableConv2D_x86_fma::forward(const std::vector<Mat>& bottom_blobs, std::vector<Mat>& top_blobs, const Option& opt) const
{
const Mat& bottom_blob = bottom_blobs[0];
const Mat& offset = bottom_blobs[1];
const bool has_mask = (bottom_blobs.size() == 3);
const int w = bottom_blob.w;
const in... | subq $0x7b8, %rsp # imm = 0x7B8
movq %rdi, 0x468(%rsp)
movq %rsi, 0x460(%rsp)
movq %rdx, 0x458(%rsp)
movq %rcx, 0x450(%rsp)
movq 0x468(%rsp), %rax
movq %rax, 0x130(%rsp)
movq 0x460(%rsp), %rdi
xorl %eax, %eax
movl %eax, %esi
callq 0xb5820
movq %rax, 0x448(%rsp)
movq 0x460(%rsp), %rdi
movl $0x1, %esi
callq 0x... | /ysh329[P]ncnn/build_O0/src/layer/x86/deformableconv2d_x86_fma.cpp |
Subsets and Splits
SQL Console for LLM4Binary/decompile-bench
Filters out entries with file names ending in .cpp, providing a basic subset of the dataset that excludes C++ files.