name string | code string | asm string | file string |
|---|---|---|---|
testing::internal::PatternMatchesString(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&, char const*, char const*) | static bool PatternMatchesString(const std::string& name_str,
const char* pattern, const char* pattern_end) {
const char* name = name_str.c_str();
const char* const name_begin = name;
const char* const name_end = name + name_str.size();
const char* pattern_next = pattern;
con... | pushq %rbp
movq %rsp, %rbp
subq $0x60, %rsp
movq %rdi, -0x10(%rbp)
movq %rsi, -0x18(%rbp)
movq %rdx, -0x20(%rbp)
movq -0x10(%rbp), %rdi
callq 0x9270
movq %rax, -0x28(%rbp)
movq -0x28(%rbp), %rax
movq %rax, -0x30(%rbp)
movq -0x28(%rbp), %rax
movq %rax, -0x50(%rbp)
movq -0x10(%rbp), %rdi
callq 0x9460
movq %rax, %rcx
movq... | /alk[P]gperftools/vendor/googletest/googletest/src/gtest.cc |
testing::internal::GTestFlagSaver::GTestFlagSaver() | GTestFlagSaver() {
also_run_disabled_tests_ = GTEST_FLAG_GET(also_run_disabled_tests);
break_on_failure_ = GTEST_FLAG_GET(break_on_failure);
catch_exceptions_ = GTEST_FLAG_GET(catch_exceptions);
color_ = GTEST_FLAG_GET(color);
death_test_style_ = GTEST_FLAG_GET(death_test_style);
death_test_use_... | pushq %rbp
movq %rsp, %rbp
subq $0x50, %rsp
movq %rdi, -0x8(%rbp)
movq -0x8(%rbp), %rdi
movq %rdi, -0x28(%rbp)
addq $0x8, %rdi
movq %rdi, -0x20(%rbp)
callq 0x9a00
movq -0x28(%rbp), %rdi
addq $0x28, %rdi
movq %rdi, -0x50(%rbp)
callq 0x9a00
movq -0x28(%rbp), %rdi
addq $0x50, %rdi
movq %rdi, -0x48(%rbp)
callq 0x9a00
movq ... | /alk[P]gperftools/vendor/googletest/googletest/src/gtest-internal-inl.h |
testing::internal::ScopedPrematureExitFile::~ScopedPrematureExitFile() | ~ScopedPrematureExitFile() {
#ifndef GTEST_OS_ESP8266
if (!premature_exit_filepath_.empty()) {
int retval = remove(premature_exit_filepath_.c_str());
if (retval) {
GTEST_LOG_(ERROR) << "Failed to remove premature exit filepath \""
<< premature_exit_filepath_ << "\" with... | pushq %rbp
movq %rsp, %rbp
subq $0x30, %rsp
movq %rdi, -0x8(%rbp)
movq -0x8(%rbp), %rdi
movq %rdi, -0x18(%rbp)
callq 0x9bf0
testb $0x1, %al
jne 0x37887
movq -0x18(%rbp), %rdi
callq 0x9270
movq %rax, %rdi
callq 0x9cd0
movl %eax, -0xc(%rbp)
cmpl $0x0, -0xc(%rbp)
je 0x37885
leaq 0x2a029(%rip), %rdx # 0x6183b
leaq -0x1... | /alk[P]gperftools/vendor/googletest/googletest/src/gtest.cc |
void testing::internal::ParseGoogleTestFlagsOnlyImpl<char>(int*, char**) | void ParseGoogleTestFlagsOnlyImpl(int* argc, CharType** argv) {
std::string flagfile_value;
for (int i = 1; i < *argc; i++) {
const std::string arg_string = StreamableToString(argv[i]);
const char* const arg = arg_string.c_str();
using internal::ParseFlag;
bool remove_flag = false;
if (ParseGo... | pushq %rbp
movq %rsp, %rbp
subq $0x80, %rsp
movq %rdi, -0x8(%rbp)
movq %rsi, -0x10(%rbp)
leaq -0x30(%rbp), %rdi
callq 0x9a00
movl $0x1, -0x34(%rbp)
movl -0x34(%rbp), %eax
movq -0x8(%rbp), %rcx
cmpl (%rcx), %eax
jge 0x394b0
movq -0x10(%rbp), %rax
movslq -0x34(%rbp), %rcx
leaq (%rax,%rcx,8), %rsi
leaq -0x58(%rbp), %rdi
c... | /alk[P]gperftools/vendor/googletest/googletest/src/gtest.cc |
environ_telnet | static int _environ_telnet(telnet_t *telnet, unsigned char type,
char* buffer, size_t size) {
telnet_event_t ev;
struct telnet_environ_t *values = 0;
char *c, *last, *out;
size_t index, count;
/* if we have no data, just pass it through */
if (size == 0) {
return 0;
}
/* first byte must be a valid command... | pushq %rbp
movq %rsp, %rbp
subq $0x90, %rsp
movb %sil, %al
movq %rdi, -0x10(%rbp)
movb %al, -0x11(%rbp)
movq %rdx, -0x20(%rbp)
movq %rcx, -0x28(%rbp)
movq $0x0, -0x58(%rbp)
cmpq $0x0, -0x28(%rbp)
jne 0x43a8
movl $0x0, -0x4(%rbp)
jmp 0x47e9
movq -0x20(%rbp), %rax
movsbl (%rax), %eax
cmpl $0x1, %eax
je 0x4405
movq -0x20(... | /seanmiddleditch[P]libtelnet/libtelnet.c |
mssp_telnet | static int _mssp_telnet(telnet_t *telnet, char* buffer, size_t size) {
telnet_event_t ev;
struct telnet_environ_t *values;
char *var = 0;
char *c, *last, *out;
size_t i, count;
unsigned char next_type;
/* if we have no data, just pass it through */
if (size == 0) {
return 0;
}
/* first byte must be a VAR ... | pushq %rbp
movq %rsp, %rbp
subq $0xa0, %rsp
movq %rdi, -0x10(%rbp)
movq %rsi, -0x18(%rbp)
movq %rdx, -0x20(%rbp)
movq $0x0, -0x58(%rbp)
cmpq $0x0, -0x20(%rbp)
jne 0x4832
movl $0x0, -0x4(%rbp)
jmp 0x4ae4
movq -0x18(%rbp), %rax
movsbl (%rax), %eax
cmpl $0x1, %eax
je 0x4870
movq -0x10(%rbp), %rdi
movl $0x2a6, %esi ... | /seanmiddleditch[P]libtelnet/libtelnet.c |
apngasm::APNGAsm::savePNGs(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&) const | bool APNGAsm::savePNGs(const std::string& outputDir) const
{
const int count = _frames.size();
for(int i = 0; i < count; ++i)
{
const std::string outputPath = _listener->onCreatePngPath(outputDir, i);
if( !_listener->onPreSave(outputPath) )
return false;
if( !_frames[i].save(... | pushq %rbp
movq %rsp, %rbp
subq $0x70, %rsp
movq %rdi, -0x10(%rbp)
movq %rsi, -0x18(%rbp)
movq -0x10(%rbp), %rdi
movq %rdi, -0x58(%rbp)
callq 0xb8e40
movl %eax, -0x1c(%rbp)
movl $0x0, -0x20(%rbp)
movl -0x20(%rbp), %eax
cmpl -0x1c(%rbp), %eax
jge 0xcef33
movq -0x58(%rbp), %rax
movq 0x20(%rax), %rsi
movq -0x18(%rbp), %rd... | /apngasm[P]apngasm/lib/src/apngasm.cpp |
apngasm::listener::APNGAsmListener::~APNGAsmListener() | class APNGASM_DECLSPEC APNGAsmListener : public IAPNGAsmListener {
public:
// Called before add frame.
// Return true if can add.
bool onPreAddFrame(const std::string &filePath, unsigned int delayNum,
unsigned int delayDen) const;
// Called before add frame.
// Return true if can add.
... | pushq %rbp
movq %rsp, %rbp
subq $0x10, %rsp
movq %rdi, -0x8(%rbp)
movq -0x8(%rbp), %rdi
callq 0xbcce0
addq $0x10, %rsp
popq %rbp
retq
nopl (%rax,%rax)
| /apngasm[P]apngasm/lib/src/listener/apngasmlistener.h |
apngasm::APNGFrame::APNGFrame(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&, unsigned int, unsigned int) | APNGFrame::APNGFrame(const std::string &filePath, unsigned delayNum,
unsigned delayDen)
: _pixels(NULL), _width(0), _height(0), _colorType(0), _paletteSize(0),
_transparencySize(0), _delayNum(delayNum), _delayDen(delayDen),
_rows(NULL) {
// TODO save extracted info to self
FILE ... | pushq %rbp
movq %rsp, %rbp
subq $0x70, %rsp
movq %rdi, -0x8(%rbp)
movq %rsi, -0x10(%rbp)
movl %edx, -0x14(%rbp)
movl %ecx, -0x18(%rbp)
movq -0x8(%rbp), %rax
movq %rax, -0x68(%rbp)
movq $0x0, (%rax)
movl $0x0, 0x8(%rax)
movl $0x0, 0xc(%rax)
movb $0x0, 0x10(%rax)
movl $0x0, 0x414(%rax)
movl $0x0, 0x418(%rax)
movl -0x14(%... | /apngasm[P]apngasm/lib/src/apngframe.cpp |
apngasm::spec::(anonymous namespace)::createRelativeDir(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>> const&) | const std::string createRelativeDir(const std::string& from, const std::string& to)
{
std::filesystem::path fromPath = createAbsolutePath(from);
std::filesystem::path toPath = createAbsolutePath(to);
const std::string separatorStr = std::string(1, separator);
// Convert path to na... | pushq %rbp
movq %rsp, %rbp
subq $0x2a0, %rsp # imm = 0x2A0
movq %rdi, -0x238(%rbp)
movq %rdi, %rax
movq %rax, -0x230(%rbp)
movq %rdi, -0x8(%rbp)
movq %rsi, -0x10(%rbp)
movq %rdx, -0x18(%rbp)
movq -0x10(%rbp), %rsi
leaq -0x40(%rbp), %rdi
callq 0x112ff0
movq -0x18(%rbp), %rsi
leaq -0x68(%rbp), %rdi
callq 0x112... | /apngasm[P]apngasm/lib/src/spec/specwriter.cpp |
void readEntry<std::_Deque_iterator<std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>>, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>>&, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char>>*>>(std::_Deque_iterator<std::__cxx11::basic_... | void readEntry(StringListIterator &iter, double &val){
val = std::stod(*iter++);
} | pushq %rbp
pushq %r15
pushq %r14
pushq %r12
pushq %rbx
subq $0x10, %rsp
movq %rsi, %r14
movq (%rdi), %rax
movq 0x18(%rdi), %rcx
leaq 0x20(%rax), %rdx
movq %rdx, (%rdi)
cmpq 0x10(%rdi), %rdx
jne 0xe9e0
leaq 0x8(%rcx), %rdx
movq %rdx, 0x18(%rdi)
movq 0x8(%rcx), %rcx
movq %rcx, 0x8(%rdi)
leaq 0x200(%rcx), %rdx
movq %rdx, ... | /mkstoyanov[P]TASMANIAN/SparseGrids/Benchmarks/benchCommon.hpp |
despot::Tiger::PrintAction(int, std::ostream&) const | void Tiger::PrintAction(ACT_TYPE action, ostream& out) const {
if (action == LEFT) {
out << "Open left" << endl;
} else if (action == RIGHT) {
out << "Open right" << endl;
} else {
out << "Listen" << endl;
}
} | pushq %rax
movq %rdx, %rdi
cmpl $0x1, %esi
leaq 0x175c(%rip), %rax # 0x70dc
leaq 0x1760(%rip), %rcx # 0x70e7
cmoveq %rax, %rcx
testl %esi, %esi
leaq 0x173e(%rip), %rsi # 0x70d2
cmovneq %rcx, %rsi
callq 0x5160
movq %rax, %rdi
popq %rax
jmp 0x50c0
| /AdaCompNUS[P]despot/examples/cpp_models/tiger/src/tiger.cpp |
main | int main(int argc, char** argv)
{
if (argc != 2)
{
fprintf(stderr, "Usage: %s [imagepath]\n", argv[0]);
return -1;
}
const char* imagepath = argv[1];
cv::Mat m = cv::imread(imagepath, 1);
if (m.empty())
{
fprintf(stderr, "cv::imread %s failed\n", imagepath);
... | subq $0xe8, %rsp
movl $0x0, 0xe4(%rsp)
movl %edi, 0xe0(%rsp)
movq %rsi, 0xd8(%rsp)
cmpl $0x2, 0xe0(%rsp)
je 0x2497e
movq 0x1f0268e(%rip), %rax # 0x1f26fe0
movq (%rax), %rdi
movq 0xd8(%rsp), %rax
movq (%rax), %rdx
leaq 0x1e736d1(%rip), %rsi # 0x1e98038
movb $0x0, %al
callq 0x24200
movl $0xffffffff, 0xe4(%rsp) # imm ... | /Tencent[P]ncnn/examples/retinaface.cpp |
draw_faceobjects(cv::Mat const&, std::vector<FaceObject, std::allocator<FaceObject>> const&) | static void draw_faceobjects(const cv::Mat& bgr, const std::vector<FaceObject>& faceobjects)
{
cv::Mat image = bgr.clone();
for (size_t i = 0; i < faceobjects.size(); i++)
{
const FaceObject& obj = faceobjects[i];
fprintf(stderr, "%.5f at %.2f %.2f %.2f x %.2f\n", obj.prob,
... | subq $0x498, %rsp # imm = 0x498
movq %rdi, 0x490(%rsp)
movq %rsi, 0x488(%rsp)
movq 0x490(%rsp), %rsi
leaq 0x428(%rsp), %rdi
callq 0x24140
movq $0x0, 0x420(%rsp)
movq 0x420(%rsp), %rax
movq %rax, 0x38(%rsp)
movq 0x488(%rsp), %rdi
callq 0x30600
movq %rax, %rcx
movq 0x38(%rsp), %rax
cmpq %rcx, %rax
jae 0x29693
... | /Tencent[P]ncnn/examples/retinaface.cpp |
ncnn::Mat::clone_from(ncnn::Mat const&, ncnn::Allocator*) | void Mat::clone_from(const ncnn::Mat& mat, Allocator* allocator)
{
*this = mat.clone(allocator);
} | subq $0x118, %rsp # imm = 0x118
movq %rdi, 0x98(%rsp)
movq %rsi, 0x90(%rsp)
movq %rdx, 0x88(%rsp)
movq 0x98(%rsp), %rax
movq %rax, 0x20(%rsp)
movq 0x90(%rsp), %rsi
movq 0x88(%rsp), %rdx
leaq 0x40(%rsp), %rdi
callq 0x33e40
movq 0x20(%rsp), %rax
movq %rax, 0xc0(%rsp)
leaq 0x40(%rsp), %rax
movq %rax, 0xb8(%rsp)... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::Mat::substract_mean_normalize(float const*, float const*) | void Mat::substract_mean_normalize(const float* mean_vals, const float* norm_vals)
{
Layer* op;
if (mean_vals && !norm_vals)
{
// substract mean only
op = create_layer(LayerType::Bias);
ParamDict pd;
pd.set(0, c);
op->load_param(pd);
Mat weights[1];
... | subq $0x958, %rsp # imm = 0x958
movq %rdi, 0x578(%rsp)
movq %rsi, 0x570(%rsp)
movq %rdx, 0x568(%rsp)
movq 0x578(%rsp), %rax
movq %rax, 0x240(%rsp)
cmpq $0x0, 0x570(%rsp)
je 0x39cff
cmpq $0x0, 0x568(%rsp)
jne 0x39cff
movl $0x3, %edi
callq 0x89490
movq %rax, 0x560(%rsp)
leaq 0x550(%rsp), %rdi
movq %rdi, 0x238(... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::resize_nearest(ncnn::Mat const&, ncnn::Mat&, int, int, ncnn::Option const&) | void resize_nearest(const Mat& src, Mat& dst, int w, int h, const Option& opt)
{
Layer* interp = create_layer(LayerType::Interp);
ParamDict pd;
pd.set(0, 1);
pd.set(3, h);
pd.set(4, w);
interp->load_param(pd);
interp->create_pipeline(opt);
interp->forward(src, dst, opt);
interp-... | subq $0x58, %rsp
movq %rdi, 0x50(%rsp)
movq %rsi, 0x48(%rsp)
movl %edx, 0x44(%rsp)
movl %ecx, 0x40(%rsp)
movq %r8, 0x38(%rsp)
movl $0x32, %edi
callq 0x89490
movq %rax, 0x30(%rsp)
leaq 0x20(%rsp), %rdi
movq %rdi, 0x8(%rsp)
callq 0x78060
movq 0x8(%rsp), %rdi
xorl %esi, %esi
movl $0x1, %edx
callq 0x79040
jmp 0x3cacc
movl ... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::flatten(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | void flatten(const Mat& src, Mat& dst, const Option& opt)
{
Layer* flatten = create_layer(LayerType::Flatten);
ParamDict pd;
flatten->load_param(pd);
flatten->create_pipeline(opt);
flatten->forward(src, dst, opt);
flatten->destroy_pipeline(opt);
delete flatten;
} | subq $0x58, %rsp
movq %rdi, 0x50(%rsp)
movq %rsi, 0x48(%rsp)
movq %rdx, 0x40(%rsp)
movl $0xe, %edi
callq 0x89490
movq %rax, 0x38(%rsp)
leaq 0x28(%rsp), %rdi
movq %rdi, 0x10(%rsp)
callq 0x78060
movq 0x10(%rsp), %rsi
movq 0x38(%rsp), %rdi
movq (%rdi), %rax
movq 0x10(%rax), %rax
callq *%rax
jmp 0x3cf16
movq 0x38(%rsp), %r... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::cast_int8_to_float32(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | void cast_int8_to_float32(const Mat& src, Mat& dst, const Option& opt)
{
Layer* cast = create_layer(LayerType::Cast);
ParamDict pd;
pd.set(0, 3);
pd.set(1, 1);
cast->load_param(pd);
cast->create_pipeline(opt);
cast->forward(src, dst, opt);
cast->destroy_pipeline(opt);
delete ca... | subq $0x58, %rsp
movq %rdi, 0x50(%rsp)
movq %rsi, 0x48(%rsp)
movq %rdx, 0x40(%rsp)
movl $0x40, %edi
callq 0x89490
movq %rax, 0x38(%rsp)
leaq 0x28(%rsp), %rdi
movq %rdi, 0x10(%rsp)
callq 0x78060
movq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x3, %edx
callq 0x79040
jmp 0x3d104
leaq 0x28(%rsp), %rdi
movl $0x1, %edx
movl %ed... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::cast_bfloat16_to_float32(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | void cast_bfloat16_to_float32(const Mat& src, Mat& dst, const Option& opt)
{
Layer* cast = create_layer(LayerType::Cast);
ParamDict pd;
pd.set(0, 4);
pd.set(1, 1);
cast->load_param(pd);
cast->create_pipeline(opt);
cast->forward(src, dst, opt);
cast->destroy_pipeline(opt);
delet... | subq $0x58, %rsp
movq %rdi, 0x50(%rsp)
movq %rsi, 0x48(%rsp)
movq %rdx, 0x40(%rsp)
movl $0x40, %edi
callq 0x89490
movq %rax, 0x38(%rsp)
leaq 0x28(%rsp), %rdi
movq %rdi, 0x10(%rsp)
callq 0x78060
movq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x4, %edx
callq 0x79040
jmp 0x3d324
leaq 0x28(%rsp), %rdi
movl $0x1, %edx
movl %ed... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::dequantize_from_int32(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, ncnn::Option const&) | void dequantize_from_int32(const Mat& src, Mat& dst, const Mat& scale_data, const Mat& bias_data, const Option& opt)
{
Layer* dequantize = create_layer(LayerType::Dequantize);
ParamDict pd;
pd.set(0, scale_data.w);
pd.set(1, bias_data.w);
dequantize->load_param(pd);
Mat weights[2];
weight... | subq $0x268, %rsp # imm = 0x268
movq %rdi, 0x188(%rsp)
movq %rsi, 0x180(%rsp)
movq %rdx, 0x178(%rsp)
movq %rcx, 0x170(%rsp)
movq %r8, 0x168(%rsp)
movl $0x3a, %edi
callq 0x89490
movq %rax, 0x160(%rsp)
leaq 0x150(%rsp), %rdi
movq %rdi, 0x98(%rsp)
callq 0x78060
movq 0x98(%rsp), %rdi
movq 0x178(%rsp), %rax
movl ... | /Tencent[P]ncnn/src/mat.cpp |
ncnn::Mat::from_pixels(unsigned char const*, int, int, int, int, ncnn::Allocator*) | Mat Mat::from_pixels(const unsigned char* pixels, int type, int w, int h, int stride, Allocator* allocator)
{
Mat m;
if (type & PIXEL_CONVERT_MASK)
{
switch (type)
{
case PIXEL_RGB2BGR:
case PIXEL_BGR2RGB:
from_rgb2bgr(pixels, w, h, stride, m, allocator);
... | subq $0xb8, %rsp
movq %rdi, 0x20(%rsp)
movq %rdi, %rax
movq %rax, 0x28(%rsp)
movq 0xc0(%rsp), %rax
movq %rdi, 0x68(%rsp)
movq %rsi, 0x60(%rsp)
movl %edx, 0x5c(%rsp)
movl %ecx, 0x58(%rsp)
movl %r8d, 0x54(%rsp)
movl %r9d, 0x50(%rsp)
movb $0x0, 0x4f(%rsp)
movq %rdi, 0x70(%rsp)
movq 0x70(%rsp), %rax
movq %rax, 0x30(%rsp)
m... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::from_rgb2gray(unsigned char const*, int, int, int, ncnn::Mat&, ncnn::Allocator*) | static int from_rgb2gray(const unsigned char* rgb, int w, int h, int stride, Mat& m, Allocator* allocator)
{
// coeffs for r g b = 0.299f, 0.587f, 0.114f
const unsigned char Y_shift = 8; //14
const unsigned char R2Y = 77;
const unsigned char G2Y = 150;
const unsigned char B2Y = 29;
m.create(w, ... | subq $0x78, %rsp
movq %rdi, 0x50(%rsp)
movl %esi, 0x4c(%rsp)
movl %edx, 0x48(%rsp)
movl %ecx, 0x44(%rsp)
movq %r8, 0x38(%rsp)
movq %r9, 0x30(%rsp)
movb $0x8, 0x2f(%rsp)
movb $0x4d, 0x2e(%rsp)
movb $-0x6a, 0x2d(%rsp)
movb $0x1d, 0x2c(%rsp)
movq 0x38(%rsp), %rdi
movl 0x4c(%rsp), %esi
movl 0x48(%rsp), %edx
movq 0x30(%rsp)... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::from_rgba2bgra(unsigned char const*, int, int, int, ncnn::Mat&, ncnn::Allocator*) | static int from_rgba2bgra(const unsigned char* rgba, int w, int h, int stride, Mat& m, Allocator* allocator)
{
m.create(w, h, 4, 4u, allocator);
if (m.empty())
return -100;
const int wgap = stride - w * 4;
if (wgap == 0)
{
w = w * h;
h = 1;
}
float* ptr0 = m.channel... | subq $0x4e8, %rsp # imm = 0x4E8
movq %rdi, 0x228(%rsp)
movl %esi, 0x224(%rsp)
movl %edx, 0x220(%rsp)
movl %ecx, 0x21c(%rsp)
movq %r8, 0x210(%rsp)
movq %r9, 0x208(%rsp)
movq 0x210(%rsp), %rdi
movl 0x224(%rsp), %esi
movl 0x220(%rsp), %edx
movq 0x208(%rsp), %r9
movl $0x4, %ecx
movl $0x4, %r8d
callq 0x38aa0
movq... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::from_rgb(unsigned char const*, int, int, int, ncnn::Mat&, ncnn::Allocator*) | static int from_rgb(const unsigned char* rgb, int w, int h, int stride, Mat& m, Allocator* allocator)
{
m.create(w, h, 3, 4u, allocator);
if (m.empty())
return -100;
const int wgap = stride - w * 3;
if (wgap == 0)
{
w = w * h;
h = 1;
}
float* ptr0 = m.channel(0);
... | subq $0x3c8, %rsp # imm = 0x3C8
movq %rdi, 0x1b0(%rsp)
movl %esi, 0x1ac(%rsp)
movl %edx, 0x1a8(%rsp)
movl %ecx, 0x1a4(%rsp)
movq %r8, 0x198(%rsp)
movq %r9, 0x190(%rsp)
movq 0x198(%rsp), %rdi
movl 0x1ac(%rsp), %esi
movl 0x1a8(%rsp), %edx
movq 0x190(%rsp), %r9
movl $0x3, %ecx
movl $0x4, %r8d
callq 0x38aa0
movq... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::from_gray(unsigned char const*, int, int, int, ncnn::Mat&, ncnn::Allocator*) | static int from_gray(const unsigned char* gray, int w, int h, int stride, Mat& m, Allocator* allocator)
{
m.create(w, h, 1, 4u, allocator);
if (m.empty())
return -100;
const int wgap = stride - w;
if (wgap == 0)
{
w = w * h;
h = 1;
}
float* ptr = m;
for (int y ... | subq $0x78, %rsp
movq %rdi, 0x50(%rsp)
movl %esi, 0x4c(%rsp)
movl %edx, 0x48(%rsp)
movl %ecx, 0x44(%rsp)
movq %r8, 0x38(%rsp)
movq %r9, 0x30(%rsp)
movq 0x38(%rsp), %rdi
movl 0x4c(%rsp), %esi
movl 0x48(%rsp), %edx
movq 0x30(%rsp), %r9
movl $0x1, %ecx
movl $0x4, %r8d
callq 0x38aa0
movq 0x38(%rsp), %rax
movq %rax, 0x60(%r... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::to_rgb2rgba(ncnn::Mat const&, unsigned char*, int) | static void to_rgb2rgba(const Mat& m, unsigned char* rgba, int stride)
{
int w = m.w;
int h = m.h;
const int wgap = stride - w * 4;
if (wgap == 0)
{
w = w * h;
h = 1;
}
const float* ptr0 = m.channel(0);
const float* ptr1 = m.channel(1);
const float* ptr2 = m.channel... | subq $0x3b8, %rsp # imm = 0x3B8
movq %rdi, 0x1b8(%rsp)
movq %rsi, 0x1b0(%rsp)
movl %edx, 0x1ac(%rsp)
movq 0x1b8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1a8(%rsp)
movq 0x1b8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x1a4(%rsp)
movl 0x1ac(%rsp), %eax
movl 0x1a8(%rsp), %ecx
shll $0x2, %ecx
subl %ecx,... | /Tencent[P]ncnn/src/mat_pixel.cpp |
ncnn::resize_bilinear_c4(unsigned char const*, int, int, int, unsigned char*, int, int, int) | void resize_bilinear_c4(const unsigned char* src, int srcw, int srch, int srcstride, unsigned char* dst, int w, int h, int stride)
{
const int INTER_RESIZE_COEF_BITS = 11;
const int INTER_RESIZE_COEF_SCALE = 1 << INTER_RESIZE_COEF_BITS;
double scale_x = (double)srcw / w;
double scale_y = (double)srch /... | pushq %rbx
subq $0x310, %rsp # imm = 0x310
movl 0x328(%rsp), %eax
movl 0x320(%rsp), %eax
movq %rdi, 0x248(%rsp)
movl %esi, 0x244(%rsp)
movl %edx, 0x240(%rsp)
movl %ecx, 0x23c(%rsp)
movq %r8, 0x230(%rsp)
movl %r9d, 0x22c(%rsp)
movl $0xb, 0x228(%rsp)
movl $0x800, 0x224(%rsp) # imm = 0x800
cvtsi2sdl 0x244(%rsp)... | /Tencent[P]ncnn/src/mat_pixel_resize.cpp |
ncnn::ModelBin::load(int, int, int) const | Mat ModelBin::load(int w, int h, int type) const
{
Mat m = load(w * h, type);
if (m.empty())
return m;
return m.reshape(w, h);
} | subq $0x138, %rsp # imm = 0x138
movq %rdi, 0x30(%rsp)
movq %rdi, %rax
movq %rax, 0x38(%rsp)
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movl %edx, 0xb4(%rsp)
movl %ecx, 0xb0(%rsp)
movl %r8d, 0xac(%rsp)
movq 0xb8(%rsp), %rsi
movl 0xb4(%rsp), %edx
imull 0xb0(%rsp), %edx
movl 0xac(%rsp), %ecx
movq (%rsi), %rax
... | /Tencent[P]ncnn/src/modelbin.cpp |
ncnn::ModelBin::load(int, int, int, int, int) const | Mat ModelBin::load(int w, int h, int d, int c, int type) const
{
Mat m = load(w * h * d * c, type);
if (m.empty())
return m;
return m.reshape(w, h, d, c);
} | subq $0x138, %rsp # imm = 0x138
movq %rdi, 0x30(%rsp)
movq %rdi, %rax
movq %rax, 0x38(%rsp)
movl 0x140(%rsp), %eax
movq %rdi, 0xc0(%rsp)
movq %rsi, 0xb8(%rsp)
movl %edx, 0xb4(%rsp)
movl %ecx, 0xb0(%rsp)
movl %r8d, 0xac(%rsp)
movl %r9d, 0xa8(%rsp)
movq 0xb8(%rsp), %rsi
movl 0xb4(%rsp), %edx
imull 0xb0(%rsp), ... | /Tencent[P]ncnn/src/modelbin.cpp |
ncnn::NetPrivate::do_forward_layer(ncnn::Layer const*, std::vector<ncnn::Mat, std::allocator<ncnn::Mat>>&, ncnn::Option const&) const | int NetPrivate::do_forward_layer(const Layer* layer, std::vector<Mat>& blob_mats, const Option& opt) const
{
if (layer->one_blob_only)
{
int bottom_blob_index = layer->bottoms[0];
int top_blob_index = layer->tops[0];
Mat& bottom_blob_ref = blob_mats[bottom_blob_index];
Mat botto... | subq $0x708, %rsp # imm = 0x708
movq %rdi, 0x3c0(%rsp)
movq %rsi, 0x3b8(%rsp)
movq %rdx, 0x3b0(%rsp)
movq %rcx, 0x3a8(%rsp)
movq 0x3c0(%rsp), %rax
movq %rax, 0x1a0(%rsp)
movq 0x3b8(%rsp), %rax
testb $0x1, 0x8(%rax)
je 0x62dc8
movq 0x3b8(%rsp), %rdi
addq $0x70, %rdi
xorl %eax, %eax
movl %eax, %esi
callq 0x6f5... | /Tencent[P]ncnn/src/net.cpp |
ncnn::Net::register_custom_layer(int, ncnn::Layer* (*)(void*), void (*)(ncnn::Layer*, void*), void*) | int Net::register_custom_layer(int index, layer_creator_func creator, layer_destroyer_func destroyer, void* userdata)
{
int custom_index = index & ~LayerType::CustomBit;
if (index == custom_index)
{
NCNN_LOGE("overwrite built-in layer type %d", index);
for (size_t i = 0; i < d->overwrite_bu... | subq $0xc8, %rsp
movq %rdi, 0xb8(%rsp)
movl %esi, 0xb4(%rsp)
movq %rdx, 0xa8(%rsp)
movq %rcx, 0xa0(%rsp)
movq %r8, 0x98(%rsp)
movq 0xb8(%rsp), %rax
movq %rax, 0x40(%rsp)
movl 0xb4(%rsp), %eax
andl $0xfffffeff, %eax # imm = 0xFFFFFEFF
movl %eax, 0x94(%rsp)
movl 0xb4(%rsp), %eax
cmpl 0x94(%rsp), %eax
jne 0x666e8
jm... | /Tencent[P]ncnn/src/net.cpp |
ncnn::Net::load_param_bin(char const*) | int Net::load_param_bin(const char* protopath)
{
FILE* fp = fopen(protopath, "rb");
if (!fp)
{
NCNN_LOGE("fopen %s failed", protopath);
return -1;
}
int ret = load_param_bin(fp);
fclose(fp);
return ret;
} | subq $0x38, %rsp
movq %rdi, 0x28(%rsp)
movq %rsi, 0x20(%rsp)
movq 0x28(%rsp), %rax
movq %rax, 0x8(%rsp)
movq 0x20(%rsp), %rdi
leaq 0x1e2c143(%rip), %rsi # 0x1e98907
callq 0x241b0
movq %rax, 0x18(%rsp)
cmpq $0x0, 0x18(%rsp)
jne 0x6c817
jmp 0x6c7d8
movq 0x1eba801(%rip), %rax # 0x1f26fe0
movq (%rax), %rdi
movq 0x20(%r... | /Tencent[P]ncnn/src/net.cpp |
ncnn::Net::custom_layer_to_index(char const*) | int Net::custom_layer_to_index(const char* type)
{
const size_t custom_layer_registry_entry_count = d->custom_layer_registry.size();
for (size_t i = 0; i < custom_layer_registry_entry_count; i++)
{
if (strcmp(type, d->custom_layer_registry[i].name) == 0)
return static_cast<int>(i);
}... | subq $0x38, %rsp
movq %rdi, 0x28(%rsp)
movq %rsi, 0x20(%rsp)
movq 0x28(%rsp), %rax
movq %rax, 0x8(%rsp)
movq 0x48(%rax), %rdi
addq $0x98, %rdi
callq 0x6f980
movq %rax, 0x18(%rsp)
movq $0x0, 0x10(%rsp)
movq 0x10(%rsp), %rax
cmpq 0x18(%rsp), %rax
jae 0x6cd73
movq 0x8(%rsp), %rax
movq 0x20(%rsp), %rcx
movq %rcx, (%rsp)
mo... | /Tencent[P]ncnn/src/net.cpp |
ncnn::Extractor::input(char const*, ncnn::Mat const&) | int Extractor::input(const char* blob_name, const Mat& in)
{
int blob_index = d->net->find_blob_index_by_name(blob_name);
if (blob_index == -1)
{
NCNN_LOGE("Try");
const std::vector<const char*>& input_names = d->net->input_names();
for (size_t i = 0; i < input_names.size(); i++)
... | subq $0x58, %rsp
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movq %rdx, 0x38(%rsp)
movq 0x48(%rsp), %rax
movq %rax, 0x18(%rsp)
movq 0x8(%rax), %rax
movq (%rax), %rdi
movq 0x40(%rsp), %rsi
callq 0x69640
movl %eax, 0x34(%rsp)
cmpl $-0x1, 0x34(%rsp)
jne 0x6d573
jmp 0x6d48f
movq 0x1eb9b4a(%rip), %rax # 0x1f26fe0
movq (%r... | /Tencent[P]ncnn/src/net.cpp |
ncnn::Extractor::extract(char const*, ncnn::Mat&, int) | int Extractor::extract(const char* blob_name, Mat& feat, int type)
{
int blob_index = d->net->find_blob_index_by_name(blob_name);
if (blob_index == -1)
{
NCNN_LOGE("Try");
const std::vector<const char*>& output_names = d->net->output_names();
for (size_t i = 0; i < output_names.size(... | subq $0x58, %rsp
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movq %rdx, 0x38(%rsp)
movl %ecx, 0x34(%rsp)
movq 0x48(%rsp), %rax
movq %rax, 0x18(%rsp)
movq 0x8(%rax), %rax
movq (%rax), %rdi
movq 0x40(%rsp), %rsi
callq 0x69640
movl %eax, 0x30(%rsp)
cmpl $-0x1, 0x30(%rsp)
jne 0x6d917
jmp 0x6d833
movq 0x1eb97a6(%rip), %rax ... | /Tencent[P]ncnn/src/net.cpp |
ncnn::PoolAllocator::fastMalloc(unsigned long) | void* PoolAllocator::fastMalloc(size_t size)
{
d->budgets_lock.lock();
// find free budget
std::list<std::pair<size_t, void*> >::iterator it = d->budgets.begin(), it_max = d->budgets.begin(), it_min = d->budgets.begin();
for (; it != d->budgets.end(); ++it)
{
size_t bs = it->first;
... | subq $0x108, %rsp # imm = 0x108
movq %rdi, 0xd8(%rsp)
movq %rsi, 0xd0(%rsp)
movq 0xd8(%rsp), %rax
movq %rax, 0x40(%rsp)
movq 0x8(%rax), %rdi
callq 0x7d560
movq 0x40(%rsp), %rax
movq 0x8(%rax), %rdi
addq $0x60, %rdi
callq 0x7d450
movq %rax, %rcx
movq 0x40(%rsp), %rax
movq %rcx, 0xc8(%rsp)
movq 0x8(%rax), %rdi... | /Tencent[P]ncnn/src/allocator.cpp |
ncnn::cpu_support_arm_neon() | int cpu_support_arm_neon()
{
try_initialize_global_cpu_info();
#if __aarch64__
return 1;
#elif __arm__
#if defined _WIN32
return g_cpu_support_arm_neon;
#elif defined __ANDROID__ || defined __linux__
return g_hwcaps & HWCAP_NEON;
#elif __APPLE__
return g_hw_cputype == CPU_TYPE_ARM && g_hw_cpusubtype... | pushq %rax
callq 0x7e150
xorl %eax, %eax
popq %rcx
retq
nopw (%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_support_arm_asimdhp() | int cpu_support_arm_asimdhp()
{
try_initialize_global_cpu_info();
#if __aarch64__
#if defined _WIN32
return g_cpu_support_arm_asimdhp;
#elif defined __ANDROID__ || defined __linux__
return g_hwcaps & HWCAP_ASIMDHP;
#elif __APPLE__
return g_hw_optional_arm_FEAT_FP16
|| g_hw_cpufamily == CPUFAM... | pushq %rax
callq 0x7e150
xorl %eax, %eax
popq %rcx
retq
nopw (%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_support_arm_bf16() | int cpu_support_arm_bf16()
{
try_initialize_global_cpu_info();
#if __aarch64__
#if defined _WIN32
return g_cpu_support_arm_bf16;
#elif defined __ANDROID__ || defined __linux__
return g_hwcaps2 & HWCAP2_BF16;
#elif __APPLE__
return g_hw_optional_arm_FEAT_BF16
|| g_hw_cpufamily == CPUFAMILY_ARM... | pushq %rax
callq 0x7e150
xorl %eax, %eax
popq %rcx
retq
nopw (%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_support_arm_sve() | int cpu_support_arm_sve()
{
try_initialize_global_cpu_info();
#if __aarch64__
#if defined _WIN32
return g_cpu_support_arm_sve;
#elif defined __ANDROID__ || defined __linux__
return g_hwcaps & HWCAP_SVE;
#elif __APPLE__
return 0; // no known apple cpu support armv8.6 sve
#else
return 0;
#endif
#else
... | pushq %rax
callq 0x7e150
xorl %eax, %eax
popq %rcx
retq
nopw (%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_support_x86_avx() | int cpu_support_x86_avx()
{
try_initialize_global_cpu_info();
#if defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64)
return g_cpu_support_x86_avx;
#else
return 0;
#endif
} | pushq %rax
callq 0x7e150
movl 0x1ea9324(%rip), %eax # 0x1f27570
popq %rcx
retq
nop
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_support_x86_xop() | int cpu_support_x86_xop()
{
try_initialize_global_cpu_info();
#if defined(__i386__) || defined(__x86_64__) || defined(_M_IX86) || defined(_M_X64)
return g_cpu_support_x86_xop;
#else
return 0;
#endif
} | pushq %rax
callq 0x7e150
movl 0x1ea930c(%rip), %eax # 0x1f27578
popq %rcx
retq
nop
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::cpu_riscv_vlenb() | int cpu_riscv_vlenb()
{
#if C906
// FIXME xuantie qemu reports all zero auxv flags
return 16;
#endif
try_initialize_global_cpu_info();
#if __riscv
if (!cpu_support_riscv_v())
return 0;
int a = 0;
asm volatile(
".word 0xc22026f3 \n" // csrr a3, vlenb
"mv %0, a3 ... | pushq %rax
callq 0x7e150
xorl %eax, %eax
popq %rcx
retq
nopw (%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
ncnn::set_flush_denormals(int) | int set_flush_denormals(int flush_denormals)
{
if (flush_denormals < 0 || flush_denormals > 3)
{
NCNN_LOGE("denormals_zero %d not supported", flush_denormals);
return -1;
}
#if defined(__SSE3__)
if (flush_denormals == 0)
{
_MM_SET_DENORMALS_ZERO_MODE(_MM_DENORMALS_ZERO_OFF);
... | pushq %rax
movl %edi, (%rsp)
cmpl $0x0, (%rsp)
jl 0x7e790
cmpl $0x3, (%rsp)
jle 0x7e7cf
jmp 0x7e792
movq 0x1ea8847(%rip), %rax # 0x1f26fe0
movq (%rax), %rdi
movl (%rsp), %edx
leaq 0x1e1a6b1(%rip), %rsi # 0x1e98e57
movb $0x0, %al
callq 0x24200
movq 0x1ea882c(%rip), %rax # 0x1f26fe0
movq (%rax), %rdi
leaq 0x1e1988f... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_avx() | static int get_cpu_support_x86_avx()
{
unsigned int cpu_info[4] = {0};
x86_cpuid(0, cpu_info);
int nIds = cpu_info[0];
if (nIds < 1)
return 0;
x86_cpuid(1, cpu_info);
// check AVX XSAVE OSXSAVE
if (!(cpu_info[2] & (1u << 28)) || !(cpu_info[2] & (1u << 26)) || !(cpu_info[2] & (1u <<... | subq $0x28, %rsp
leaq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
leaq 0x10(%rsp), %rsi
xorl %edi, %edi
callq 0x7fe70
movl 0x10(%rsp), %eax
movl %eax, 0xc(%rsp)
cmpl $0x1, 0xc(%rsp)
jge 0x7ee0a
movl $0x0, 0x24(%rsp)
jmp 0x7ee6c
leaq 0x10(%rsp), %rsi
movl $0x1, %edi
callq 0x7fe70
movl 0x18(%rsp), %ea... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_xop() | static int get_cpu_support_x86_xop()
{
unsigned int cpu_info[4] = {0};
x86_cpuid(0x80000000, cpu_info);
if (cpu_info[0] < 0x80000001)
return 0;
x86_cpuid(0x80000001, cpu_info);
return cpu_info[2] & (1u << 11);
} | subq $0x18, %rsp
movq %rsp, %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
movq %rsp, %rsi
movl $0x80000000, %edi # imm = 0x80000000
callq 0x7fe70
cmpl $0x80000001, (%rsp) # imm = 0x80000001
jae 0x7ef63
movl $0x0, 0x14(%rsp)
jmp 0x7ef7d
movq %rsp, %rsi
movl $0x80000001, %edi # imm = 0x80000001
call... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_avx_vnni_int8() | static int get_cpu_support_x86_avx_vnni_int8()
{
unsigned int cpu_info[4] = {0};
x86_cpuid(0, cpu_info);
int nIds = cpu_info[0];
if (nIds < 7)
return 0;
x86_cpuid(1, cpu_info);
// check AVX XSAVE OSXSAVE
if (!(cpu_info[2] & (1u << 28)) || !(cpu_info[2] & (1u << 26)) || !(cpu_info[2... | subq $0x28, %rsp
leaq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
leaq 0x10(%rsp), %rsi
xorl %edi, %edi
callq 0x7fe70
movl 0x10(%rsp), %eax
movl %eax, 0xc(%rsp)
cmpl $0x7, 0xc(%rsp)
jge 0x7f1aa
movl $0x0, 0x24(%rsp)
jmp 0x7f223
leaq 0x10(%rsp), %rsi
movl $0x1, %edi
callq 0x7fe70
movl 0x18(%rsp), %ea... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_avx_ne_convert() | static int get_cpu_support_x86_avx_ne_convert()
{
unsigned int cpu_info[4] = {0};
x86_cpuid(0, cpu_info);
int nIds = cpu_info[0];
if (nIds < 7)
return 0;
x86_cpuid(1, cpu_info);
// check AVX XSAVE OSXSAVE
if (!(cpu_info[2] & (1u << 28)) || !(cpu_info[2] & (1u << 26)) || !(cpu_info[... | subq $0x28, %rsp
leaq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
leaq 0x10(%rsp), %rsi
xorl %edi, %edi
callq 0x7fe70
movl 0x10(%rsp), %eax
movl %eax, 0xc(%rsp)
cmpl $0x7, 0xc(%rsp)
jge 0x7f32a
movl $0x0, 0x24(%rsp)
jmp 0x7f3a3
leaq 0x10(%rsp), %rsi
movl $0x1, %edi
callq 0x7fe70
movl 0x18(%rsp), %ea... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_avx512() | static int get_cpu_support_x86_avx512()
{
#if __APPLE__
return get_hw_capability("hw.optional.avx512f")
&& get_hw_capability("hw.optional.avx512bw")
&& get_hw_capability("hw.optional.avx512cd")
&& get_hw_capability("hw.optional.avx512dq")
&& get_hw_capability("hw.optional... | subq $0x28, %rsp
leaq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
leaq 0x10(%rsp), %rsi
xorl %edi, %edi
callq 0x7fe70
movl 0x10(%rsp), %eax
movl %eax, 0xc(%rsp)
cmpl $0x7, 0xc(%rsp)
jge 0x7f3ed
movl $0x0, 0x24(%rsp)
jmp 0x7f4f0
leaq 0x10(%rsp), %rsi
movl $0x1, %edi
callq 0x7fe70
movl 0x18(%rsp), %ea... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_support_x86_avx512_fp16() | static int get_cpu_support_x86_avx512_fp16()
{
#if __APPLE__
return get_hw_capability("hw.optional.avx512fp16");
#else
unsigned int cpu_info[4] = {0};
x86_cpuid(0, cpu_info);
int nIds = cpu_info[0];
if (nIds < 7)
return 0;
x86_cpuid(1, cpu_info);
// check AVX XSAVE OSXSAVE
if (... | subq $0x28, %rsp
leaq 0x10(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
leaq 0x10(%rsp), %rsi
xorl %edi, %edi
callq 0x7fe70
movl 0x10(%rsp), %eax
movl %eax, 0xc(%rsp)
cmpl $0x7, 0xc(%rsp)
jge 0x7f6dd
movl $0x0, 0x24(%rsp)
jmp 0x7f770
leaq 0x10(%rsp), %rsi
movl $0x1, %edi
callq 0x7fe70
movl 0x18(%rsp), %ea... | /Tencent[P]ncnn/src/cpu.cpp |
get_cpu_level3_cachesize() | static int get_cpu_level3_cachesize()
{
int size = 0;
#if defined _WIN32
typedef BOOL(WINAPI * LPFN_GLPI)(PSYSTEM_LOGICAL_PROCESSOR_INFORMATION, PDWORD);
LPFN_GLPI glpi = (LPFN_GLPI)GetProcAddress(GetModuleHandle(TEXT("kernel32")), "GetLogicalProcessorInformation");
if (glpi != NULL)
{
DWORD... | pushq %rax
movl $0x0, 0x4(%rsp)
movl $0x3, %edi
callq 0x80030
movl %eax, 0x4(%rsp)
cmpl $0x0, 0x4(%rsp)
jg 0x7f82c
movl $0xc2, %edi
callq 0x24640
movl %eax, 0x4(%rsp)
movl 0x4(%rsp), %eax
popq %rcx
retq
nopw %cs:(%rax,%rax)
| /Tencent[P]ncnn/src/cpu.cpp |
x86_cpuid_sublevel(int, int, unsigned int*) | static inline void x86_cpuid_sublevel(int level, int sublevel, unsigned int out[4])
{
#if defined(_MSC_VER)
__cpuidex((int*)out, level, sublevel);
#elif defined(__clang__) || defined(__GNUC__)
__cpuid_count(level, sublevel, out[0], out[1], out[2], out[3]);
#else
NCNN_LOGE("x86_cpuid_sublevel is unknown for ... | movl %edi, -0x4(%rsp)
movl %esi, -0x8(%rsp)
movq %rdx, -0x10(%rsp)
movq -0x10(%rsp), %r9
movq -0x10(%rsp), %rdi
movq -0x10(%rsp), %rax
movq %rax, -0x18(%rsp)
movq -0x10(%rsp), %rax
movq %rax, -0x20(%rsp)
movl -0x4(%rsp), %eax
movl -0x8(%rsp), %ecx
xchgq %rbx, %r8
cpuid
xchgq %rbx, %r8
movl %eax, %r10d
movq -0x20(%rsp),... | /Tencent[P]ncnn/src/cpu.cpp |
get_data_cache_size(int, int) | static int get_data_cache_size(int cpuid, int level)
{
char path[256];
// discover sysfs cache entry
int indexid = -1;
for (int i = 0;; i++)
{
// check level
{
sprintf(path, "/sys/devices/system/cpu/cpu%d/cache/index%d/level", cpuid, i);
FILE* fp = fopen(path... | subq $0x358, %rsp # imm = 0x358
movl %edi, 0x350(%rsp)
movl %esi, 0x34c(%rsp)
movl $0xffffffff, 0x23c(%rsp) # imm = 0xFFFFFFFF
movl $0x0, 0x238(%rsp)
leaq 0x240(%rsp), %rdi
movl 0x350(%rsp), %edx
movl 0x238(%rsp), %ecx
leaq 0x1e18f38(%rip), %rsi # 0x1e99040
movb $0x0, %al
callq 0x24160
leaq 0x240(%rsp), %r... | /Tencent[P]ncnn/src/cpu.cpp |
ncnn::create_layer_naive(int) | Layer* create_layer_naive(int index)
{
if (index < 0 || index >= layer_registry_entry_count)
return 0;
layer_creator_func layer_creator = layer_registry[index].creator;
if (!layer_creator)
return 0;
Layer* layer = layer_creator(0);
layer->typeindex = index;
return layer;
} | subq $0x28, %rsp
movl %edi, 0x1c(%rsp)
cmpl $0x0, 0x1c(%rsp)
jl 0x895b6
cmpl $0x6b, 0x1c(%rsp)
jl 0x895c1
movq $0x0, 0x20(%rsp)
jmp 0x89613
movslq 0x1c(%rsp), %rcx
leaq 0x1e92113(%rip), %rax # 0x1f1b6e0
shlq $0x4, %rcx
addq %rcx, %rax
movq 0x8(%rax), %rax
movq %rax, 0x10(%rsp)
cmpq $0x0, 0x10(%rsp)
jne 0x895f0
movq $... | /Tencent[P]ncnn/src/layer.cpp |
ncnn::Layer_final::create_pipeline(ncnn::Option const&) | virtual int create_pipeline(const Option& opt)
{
set_layer_properties();
#if NCNN_VULKAN
if (layer_vulkan)
{
if (vkdev)
{
int ret = layer_vulkan->create_pipeline(opt);
get_layer_properties();
return ret;
}
... | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movq 0x20(%rsp), %rdi
movq %rdi, 0x8(%rsp)
callq 0x89980
movq 0x8(%rsp), %rax
movq 0xd0(%rax), %rdi
movq 0x18(%rsp), %rsi
movq (%rdi), %rax
callq *0x20(%rax)
movq 0x8(%rsp), %rdi
movl %eax, 0x14(%rsp)
callq 0x89a40
movl 0x14(%rsp), %eax
addq $0x28, %rsp
retq
... | /Tencent[P]ncnn/src/layer.cpp |
ncnn::AbsVal::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int AbsVal::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 size = w * h;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* pt... | subq $0x188, %rsp # imm = 0x188
movq %rdi, 0xb8(%rsp)
movq %rsi, 0xb0(%rsp)
movq %rdx, 0xa8(%rsp)
movq 0xb0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xb0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xa0(%rsp)
movq 0xb0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x9c(%rsp)
movl 0xa4(%rs... | /Tencent[P]ncnn/src/layer/absval.cpp |
ncnn::BatchNorm::BatchNorm() | BatchNorm::BatchNorm()
{
one_blob_only = true;
support_inplace = true;
} | subq $0x128, %rsp # imm = 0x128
movq %rdi, 0x50(%rsp)
movq 0x50(%rsp), %rdi
movq %rdi, 0x30(%rsp)
callq 0x88680
movq 0x30(%rsp), %rax
leaq 0x1e93629(%rip), %rcx # 0x1f1d950
addq $0x10, %rcx
movq %rcx, (%rax)
addq $0xd8, %rax
movq %rax, 0x58(%rsp)
movq 0x58(%rsp), %rax
movq %rax, 0x38(%rsp)
movq $0x0, (%rax... | /Tencent[P]ncnn/src/layer/batchnorm.cpp |
ncnn::BatchNorm::load_model(ncnn::ModelBin const&) | int BatchNorm::load_model(const ModelBin& mb)
{
slope_data = mb.load(channels, 1);
if (slope_data.empty())
return -100;
mean_data = mb.load(channels, 1);
if (mean_data.empty())
return -100;
var_data = mb.load(channels, 1);
if (var_data.empty())
return -100;
bias_da... | subq $0x4e8, %rsp # imm = 0x4E8
movq %rdi, 0x228(%rsp)
movq %rsi, 0x220(%rsp)
movq 0x228(%rsp), %rax
movq %rax, 0xd8(%rsp)
movq 0x220(%rsp), %rsi
movl 0xd0(%rax), %edx
movq (%rsi), %rax
leaq 0x1d8(%rsp), %rdi
movl $0x1, %ecx
callq *0x10(%rax)
movq 0xd8(%rsp), %rax
addq $0xd8, %rax
movq %rax, 0x2e8(%rsp)
leaq... | /Tencent[P]ncnn/src/layer/batchnorm.cpp |
ncnn::BatchNorm_x86::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int BatchNorm_x86::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int dims = bottom_top_blob.dims;
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int c = bottom_top_blob.c;
int elempack = bottom_top_blob.elempack;
if (dims == 1)
{
... | subq $0x598, %rsp # imm = 0x598
movq %rdi, 0x208(%rsp)
movq %rsi, 0x200(%rsp)
movq %rdx, 0x1f8(%rsp)
movq 0x208(%rsp), %rax
movq %rax, 0x78(%rsp)
movq 0x200(%rsp), %rax
movl 0x28(%rax), %eax
movl %eax, 0x1f4(%rsp)
movq 0x200(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x1f0(%rsp)
movq 0x200(%rsp), %rax
movl... | /Tencent[P]ncnn/src/layer/x86/batchnorm_x86.cpp |
ncnn::BatchNorm_x86_fma::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int BatchNorm_x86_fma::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int dims = bottom_top_blob.dims;
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int c = bottom_top_blob.c;
int elempack = bottom_top_blob.elempack;
if (dims == 1)
{
... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0xa20, %rsp # imm = 0xA20
movq %rdi, 0x3e8(%rsp)
movq %rsi, 0x3e0(%rsp)
movq %rdx, 0x3d8(%rsp)
movq 0x3e8(%rsp), %rax
movq %rax, 0x108(%rsp)
movq 0x3e0(%rsp), %rax
movl 0x28(%rax), %eax
movl %eax, 0x3d4(%rsp)
movq 0x3e0(%rsp), %rax
movl 0x2c(%rax), %eax
movl... | /Tencent[P]ncnn/build_O0/src/layer/x86/batchnorm_x86_fma.cpp |
ncnn::Bias::load_model(ncnn::ModelBin const&) | int Bias::load_model(const ModelBin& mb)
{
bias_data = mb.load(bias_data_size, 1);
if (bias_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(... | /Tencent[P]ncnn/src/layer/bias.cpp |
ncnn::Bias_x86::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int Bias_x86::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int channels = bottom_top_blob.c;
int size = w * h * d;
const float* bias_ptr = bias_data;
#pragma omp parallel for num_threads(opt... | subq $0x238, %rsp # imm = 0x238
movq %rdi, 0x108(%rsp)
movq %rsi, 0x100(%rsp)
movq %rdx, 0xf8(%rsp)
movq 0x108(%rsp), %rax
movq 0x100(%rsp), %rcx
movl 0x2c(%rcx), %ecx
movl %ecx, 0xf4(%rsp)
movq 0x100(%rsp), %rcx
movl 0x30(%rcx), %ecx
movl %ecx, 0xf0(%rsp)
movq 0x100(%rsp), %rcx
movl 0x34(%rcx), %ecx
movl %e... | /Tencent[P]ncnn/src/layer/x86/bias_x86.cpp |
ncnn::BNLL_x86_avx512::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int BNLL_x86_avx512::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int channels = bottom_top_blob.c;
int elempack = bottom_top_blob.elempack;
int size = w * h * d * elempack;
#pragma omp para... | pushq %rbp
movq %rsp, %rbp
andq $-0x40, %rsp
subq $0x4f80, %rsp # imm = 0x4F80
movq %rdi, 0x400(%rsp)
movq %rsi, 0x3f8(%rsp)
movq %rdx, 0x3f0(%rsp)
movq 0x3f8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x3ec(%rsp)
movq 0x3f8(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x3e8(%rsp)
movq 0x3f8(%rsp), %rax
mov... | /Tencent[P]ncnn/build_O0/src/layer/x86/bnll_x86_avx512.cpp |
ncnn::Convolution::load_param(ncnn::ParamDict const&) | int Convolution::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);
p... | 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 0x78d90
movl 0x34(%rsp), %edx
movl %eax, %ecx
movq 0x38(%rsp), %rax
movl %ecx, 0xd0(%rax)
movq 0xe8(%rsp), ... | /Tencent[P]ncnn/src/layer/convolution.cpp |
ncnn::Convolution_x86::create_pipeline_int8_x86(ncnn::Option const&) | int Convolution_x86::create_pipeline_int8_x86(const Option& opt)
{
const int maxk = kernel_w * kernel_h;
const int num_input = weight_data_size / maxk / num_output;
bool prefer_winograd = (opt.use_winograd23_convolution || opt.use_winograd43_convolution) && (num_input > 8 || num_output > 8);
if (opt.u... | subq $0xa8, %rsp
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movq 0x48(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movl 0xd4(%rcx), %eax
imull 0xd8(%rcx), %eax
movl %eax, 0x3c(%rsp)
movl 0x104(%rcx), %eax
cltd
idivl 0x3c(%rsp)
cltd
idivl 0xd0(%rcx)
movl %eax, 0x38(%rsp)
movq 0x40(%rsp), %rax
testb $0x1, 0x37(%rax)
jne 0xe6d9e
mo... | /Tencent[P]ncnn/src/layer/x86/convolution_x86.cpp |
ncnn::conv3x3s1_winograd23(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd23(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, const Mat& bias, int nT, const Option& opt)
{
int outw = top_blob.w;
int outh = top_blob.h;
// pad to 2n+2, winograd F(2,3)
int w_tiles = (outw + 1) / 2;
int h_tiles = (outh + 1) / 2;
int tiles = w_tiles * h_t... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movq %rcx, 0x6f8(%rsp)
movl %r8d, 0x6f4(%rsp)
movq %r9, 0x6e8(%rsp)
movq 0x708(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x6e4(%rsp)
movq 0x708(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x6e0(%rsp)
movl... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd.h |
ncnn::conv3x3s1_winograd43(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd43(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, const Mat& bias, int nT, const Option& opt)
{
int outw = top_blob.w;
int outh = top_blob.h;
// pad to 4n+2, winograd F(4,3)
int w_tiles = (outw + 3) / 4;
int h_tiles = (outh + 3) / 4;
int tiles = w_tiles * h_t... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movq %rcx, 0x6f8(%rsp)
movl %r8d, 0x6f4(%rsp)
movq %r9, 0x6e8(%rsp)
movq 0x708(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x6e4(%rsp)
movq 0x708(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x6e0(%rsp)
movl... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd.h |
ncnn::convolution_im2col_gemm(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, int, int, int, int, int, int, ncnn::Option const&) | static int convolution_im2col_gemm(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, const Mat& bias, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h, int nT, const Option& opt)
{
const int maxk = kernel_w * kernel_h;
const int M = top_blob.c * top_blob.elempack;
... | pushq %rbp
pushq %r14
pushq %rbx
subq $0xb00, %rsp # imm = 0xB00
movq 0xb48(%rsp), %rax
movl 0xb40(%rsp), %eax
movl 0xb38(%rsp), %eax
movl 0xb30(%rsp), %eax
movl 0xb28(%rsp), %eax
movl 0xb20(%rsp), %eax
movq %rdi, 0x508(%rsp)
movq %rsi, 0x500(%rsp)
movq %rdx, 0x4f8(%rsp)
movq %rcx, 0x4f0(%rsp)
movl %r8d, 0x4... | /Tencent[P]ncnn/src/layer/x86/convolution_im2col_gemm.h |
ncnn::conv3x3s1_winograd23_int8(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd23_int8(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, int nT, const Option& opt)
{
#if NCNN_RUNTIME_CPU && NCNN_AVX512VNNI && __AVX512F__ && !__AVX512VNNI__
if (ncnn::cpu_support_x86_avx512_vnni())
{
return conv3x3s1_winograd23_int8_avx512vnni(bottom_blob, top_blob, ... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movl %ecx, 0x6fc(%rsp)
movq %r8, 0x6f0(%rsp)
callq 0x7e260
cmpl $0x0, %eax
je 0x1390b0
movq 0x710(%rsp), %rdi
movq 0x708(%rsp), %rsi
movq 0x700(%rsp), %rdx
movl 0x6fc(%rsp), %ecx
movq 0x6f0(%rsp), %r8
callq 0... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd_int8.h |
ncnn::convolution_im2col_input_tile(ncnn::Mat const&, ncnn::Mat&, int, int, int, int, int, int, int, int, int, int) | static void convolution_im2col_input_tile(const Mat& bottom_blob, Mat& B, int j, int max_jj, int k, int max_kk, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h)
{
if (kernel_w == 1 && kernel_h == 1 && dilation_w == 1 && dilation_h == 1 && stride_w == 1 && stride_h == 1)
{
... | pushq %rbp
pushq %r14
pushq %rbx
subq $0x50, %rsp
movl 0x98(%rsp), %eax
movl 0x90(%rsp), %eax
movl 0x88(%rsp), %eax
movl 0x80(%rsp), %eax
movl 0x78(%rsp), %eax
movl 0x70(%rsp), %eax
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movl %edx, 0x3c(%rsp)
movl %ecx, 0x38(%rsp)
movl %r8d, 0x34(%rsp)
movl %r9d, 0x30(%rsp)
cmpl $... | /Tencent[P]ncnn/src/layer/x86/convolution_im2col_gemm.h |
ncnn::convolution_im2col_pack_A_tile_int8(ncnn::Mat const&, ncnn::Mat&, int, int, int, int) | static void convolution_im2col_pack_A_tile_int8(const Mat& A, Mat& AT, int i, int max_ii, int k, int max_kk)
{
// A = (pa, maxk, inch/pa), outch
#if NCNN_RUNTIME_CPU && __AVX512F__
Gemm_x86_avx512_utility::pack_A_tile_int8(A, AT, i, max_ii, k, max_kk);
#elif NCNN_RUNTIME_CPU && __FMA__
Gemm_x86_fma_utility... | subq $0x28, %rsp
movq %rdi, 0x20(%rsp)
movq %rsi, 0x18(%rsp)
movl %edx, 0x14(%rsp)
movl %ecx, 0x10(%rsp)
movl %r8d, 0xc(%rsp)
movl %r9d, 0x8(%rsp)
movq 0x20(%rsp), %rdi
movq 0x18(%rsp), %rsi
movl 0x14(%rsp), %edx
movl 0x10(%rsp), %ecx
movl 0xc(%rsp), %r8d
movl 0x8(%rsp), %r9d
callq 0x1763f30
addq $0x28, %rsp
retq
nopw ... | /Tencent[P]ncnn/src/layer/x86/convolution_im2col_gemm_int8.h |
ncnn::convolution_im2col_input_tile_int8(ncnn::Mat const&, ncnn::Mat&, int, int, int, int, int, int, int, int, int, int) | static void convolution_im2col_input_tile_int8(const Mat& bottom_blob, Mat& B, int j, int max_jj, int k, int max_kk, int kernel_w, int kernel_h, int dilation_w, int dilation_h, int stride_w, int stride_h)
{
#if NCNN_RUNTIME_CPU && NCNN_AVX512VNNI && __AVX512F__ && !__AVX512VNNI__
if (ncnn::cpu_support_x86_avx512_vn... | pushq %rbp
pushq %r14
pushq %rbx
subq $0x50, %rsp
movl 0x98(%rsp), %eax
movl 0x90(%rsp), %eax
movl 0x88(%rsp), %eax
movl 0x80(%rsp), %eax
movl 0x78(%rsp), %eax
movl 0x70(%rsp), %eax
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movl %edx, 0x3c(%rsp)
movl %ecx, 0x38(%rsp)
movl %r8d, 0x34(%rsp)
movl %r9d, 0x30(%rsp)
cmpl $... | /Tencent[P]ncnn/src/layer/x86/convolution_im2col_gemm_int8.h |
ncnn::Convolution_x86_avx512::create_pipeline_int8_x86(ncnn::Option const&) | int Convolution_x86_avx512::create_pipeline_int8_x86(const Option& opt)
{
const int maxk = kernel_w * kernel_h;
const int num_input = weight_data_size / maxk / num_output;
bool prefer_winograd = (opt.use_winograd23_convolution || opt.use_winograd43_convolution) && (num_input > 8 || num_output > 8);
if... | subq $0xa8, %rsp
movq %rdi, 0x48(%rsp)
movq %rsi, 0x40(%rsp)
movq 0x48(%rsp), %rcx
movq %rcx, 0x20(%rsp)
movl 0xd4(%rcx), %eax
imull 0xd8(%rcx), %eax
movl %eax, 0x3c(%rsp)
movl 0x104(%rcx), %eax
cltd
idivl 0x3c(%rsp)
cltd
idivl 0xd0(%rcx)
movl %eax, 0x38(%rsp)
movq 0x40(%rsp), %rax
testb $0x1, 0x37(%rax)
jne 0x1c828e
m... | /Tencent[P]ncnn/build_O0/src/layer/x86/convolution_x86_avx512.cpp |
ncnn::convolution_im2col_gemm_transform_kernel(ncnn::Mat const&, ncnn::Mat&, int, int, int, int, ncnn::Option const&) | static void convolution_im2col_gemm_transform_kernel(const Mat& kernel, Mat& AT, int inch, int outch, int kernel_w, int kernel_h, const Option& opt)
{
// NCNN_LOGE("convolution_im2col_gemm_transform_kernel");
const int maxk = kernel_w * kernel_h;
const int M = outch;
const int K = inch * maxk;
int... | subq $0x618, %rsp # imm = 0x618
movq 0x620(%rsp), %rax
movq %rdi, 0x308(%rsp)
movq %rsi, 0x300(%rsp)
movl %edx, 0x2fc(%rsp)
movl %ecx, 0x2f8(%rsp)
movl %r8d, 0x2f4(%rsp)
movl %r9d, 0x2f0(%rsp)
movl 0x2f4(%rsp), %eax
imull 0x2f0(%rsp), %eax
movl %eax, 0x2ec(%rsp)
movl 0x2f8(%rsp), %eax
movl %eax, 0x2e8(%rsp)
... | /Tencent[P]ncnn/src/layer/x86/convolution_im2col_gemm.h |
ncnn::conv3x3s1_winograd23(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd23(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, const Mat& bias, int nT, const Option& opt)
{
int outw = top_blob.w;
int outh = top_blob.h;
// pad to 2n+2, winograd F(2,3)
int w_tiles = (outw + 1) / 2;
int h_tiles = (outh + 1) / 2;
int tiles = w_tiles * h_t... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movq %rcx, 0x6f8(%rsp)
movl %r8d, 0x6f4(%rsp)
movq %r9, 0x6e8(%rsp)
movq 0x708(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x6e4(%rsp)
movq 0x708(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x6e0(%rsp)
movl... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd.h |
ncnn::conv3x3s1_winograd63(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd63(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, const Mat& bias, int nT, const Option& opt)
{
int outw = top_blob.w;
int outh = top_blob.h;
// pad to 6n+2, winograd F(6,3)
int w_tiles = (outw + 5) / 6;
int h_tiles = (outh + 5) / 6;
int tiles = w_tiles * h_t... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movq %rcx, 0x6f8(%rsp)
movl %r8d, 0x6f4(%rsp)
movq %r9, 0x6e8(%rsp)
movq 0x708(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x6e4(%rsp)
movq 0x708(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0x6e0(%rsp)
movl... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd.h |
ncnn::conv3x3s1_winograd23_transform_kernel_int8(ncnn::Mat const&, ncnn::Mat&, int, int, ncnn::Option const&) | static void conv3x3s1_winograd23_transform_kernel_int8(const Mat& kernel, Mat& AT, int inch, int outch, const Option& opt)
{
#if NCNN_RUNTIME_CPU && NCNN_AVX2 && __AVX__ && !__AVX2__ && !__AVXVNNI__ && !__AVX512VNNI__
if (ncnn::cpu_support_x86_avx2())
{
conv3x3s1_winograd23_transform_kernel_int8_avx2(ke... | subq $0x498, %rsp # imm = 0x498
movq %rdi, 0x220(%rsp)
movq %rsi, 0x218(%rsp)
movl %edx, 0x214(%rsp)
movl %ecx, 0x210(%rsp)
movq %r8, 0x208(%rsp)
movl 0x210(%rsp), %eax
movl %eax, 0x204(%rsp)
movl 0x214(%rsp), %eax
movl %eax, 0x200(%rsp)
movl $0x10, 0x1fc(%rsp)
movl 0x204(%rsp), %edi
movl 0x200(%rsp), %edx
m... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd_int8.h |
ncnn::conv3x3s1_winograd23_int8(ncnn::Mat const&, ncnn::Mat&, ncnn::Mat const&, int, ncnn::Option const&) | static int conv3x3s1_winograd23_int8(const Mat& bottom_blob, Mat& top_blob, const Mat& AT, int nT, const Option& opt)
{
#if NCNN_RUNTIME_CPU && NCNN_AVX512VNNI && __AVX512F__ && !__AVX512VNNI__
if (ncnn::cpu_support_x86_avx512_vnni())
{
return conv3x3s1_winograd23_int8_avx512vnni(bottom_blob, top_blob, ... | subq $0xf98, %rsp # imm = 0xF98
movq %rdi, 0x710(%rsp)
movq %rsi, 0x708(%rsp)
movq %rdx, 0x700(%rsp)
movl %ecx, 0x6fc(%rsp)
movq %r8, 0x6f0(%rsp)
callq 0x7e2e0
cmpl $0x0, %eax
je 0x294070
movq 0x710(%rsp), %rdi
movq 0x708(%rsp), %rsi
movq 0x700(%rsp), %rdx
movl 0x6fc(%rsp), %ecx
movq 0x6f0(%rsp), %r8
callq 0... | /Tencent[P]ncnn/src/layer/x86/convolution_3x3_winograd_int8.h |
ncnn::ELU::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int ELU::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int channels = bottom_top_blob.c;
int size = w * h * d;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channe... | subq $0x198, %rsp # imm = 0x198
movq %rdi, 0xc8(%rsp)
movq %rsi, 0xc0(%rsp)
movq %rdx, 0xb8(%rsp)
movq 0xc8(%rsp), %rax
movq %rax, 0x38(%rsp)
movq 0xc0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xb4(%rsp)
movq 0xc0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xb0(%rsp)
movq 0xc0(%rsp), %rax
movl 0x34(%ra... | /Tencent[P]ncnn/src/layer/elu.cpp |
ncnn::Embed::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Embed::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
int words = static_cast<int>(bottom_blob.total());
top_blob.create(num_output, words, 4u, opt.blob_allocator);
if (top_blob.empty())
return -100;
#if NCNN_INT8
if (int8_scale_term)
{
embed_int8(bot... | subq $0x58, %rsp
movq %rdi, 0x30(%rsp)
movq %rsi, 0x28(%rsp)
movq %rdx, 0x20(%rsp)
movq %rcx, 0x18(%rsp)
movq 0x30(%rsp), %rax
movq %rax, (%rsp)
movq 0x28(%rsp), %rcx
movq %rcx, 0x50(%rsp)
movq 0x50(%rsp), %rdx
movq 0x40(%rdx), %rcx
movslq 0x38(%rdx), %rdx
imulq %rdx, %rcx
movl %ecx, 0x14(%rsp)
movq 0x20(%rsp), %rdi
mo... | /Tencent[P]ncnn/src/layer/embed.cpp |
ncnn::Exp::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int Exp::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 size = w * h;
if (base == -1.f)
{
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < chann... | subq $0x2d8, %rsp # imm = 0x2D8
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movq 0x140(%rsp), %rax
movq %rax, 0x50(%rsp)
movq 0x138(%rsp), %rcx
movl 0x2c(%rcx), %ecx
movl %ecx, 0x12c(%rsp)
movq 0x138(%rsp), %rcx
movl 0x30(%rcx), %ecx
movl %ecx, 0x128(%rsp)
movq 0x138(%rsp), %rcx
movl... | /Tencent[P]ncnn/src/layer/exp.cpp |
ncnn::InnerProduct::load_model(ncnn::ModelBin const&) | int InnerProduct::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;
}
#if NCNN_INT8
if (int8_scale_term)... | subq $0x688, %rsp # imm = 0x688
movq %rdi, 0x368(%rsp)
movq %rsi, 0x360(%rsp)
movq 0x368(%rsp), %rax
movq %rax, 0xf8(%rsp)
movq 0x360(%rsp), %rsi
movl 0xd8(%rax), %edx
movq (%rsi), %rax
leaq 0x318(%rsp), %rdi
xorl %ecx, %ecx
callq *0x10(%rax)
movq 0xf8(%rsp), %rax
addq $0x130, %rax # imm = 0x130
m... | /Tencent[P]ncnn/src/layer/innerproduct.cpp |
ncnn::InnerProduct_x86_avx512::create_pipeline_fp16s(ncnn::Option const&) | int InnerProduct_x86_avx512::create_pipeline_fp16s(const Option& opt)
{
const int num_input = weight_data_size / num_output;
innerproduct_transform_kernel_fp16s_sse(weight_data, weight_data_tm, num_input, num_output, opt);
if (opt.lightmode)
weight_data.release();
return 0;
} | subq $0x48, %rsp
movq %rdi, 0x28(%rsp)
movq %rsi, 0x20(%rsp)
movq 0x28(%rsp), %rcx
movq %rcx, 0x10(%rsp)
movl 0xd8(%rcx), %eax
cltd
idivl 0xd0(%rcx)
movl %eax, %ecx
movq 0x10(%rsp), %rax
movl %ecx, 0x1c(%rsp)
movq %rax, %rdi
addq $0x130, %rdi # imm = 0x130
movq %rax, %rsi
addq $0x258, %rsi # imm =... | /Tencent[P]ncnn/build_O0/src/layer/x86/innerproduct_x86_avx512.cpp |
ncnn::Pooling::make_padding(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | void Pooling::make_padding(const Mat& bottom_blob, Mat& bottom_blob_bordered, const Option& opt) const
{
int w = bottom_blob.w;
int h = bottom_blob.h;
bottom_blob_bordered = bottom_blob;
float pad_value = 0.f;
if (pooling_type == PoolMethod_MAX)
{
pad_value = bottom_blob.elemsize == 1 ... | subq $0x1d8, %rsp # imm = 0x1D8
movq %rdi, 0x198(%rsp)
movq %rsi, 0x190(%rsp)
movq %rdx, 0x188(%rsp)
movq %rcx, 0x180(%rsp)
movq 0x198(%rsp), %rax
movq %rax, 0x40(%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... | /Tencent[P]ncnn/src/layer/pooling.cpp |
ncnn::pooling3x3s2_max_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void pooling3x3s2_max_pack4_sse(const Mat& bottom_blob, Mat& top_blob, const Option& opt)
{
int w = bottom_blob.w;
int inch = bottom_blob.c;
int outw = top_blob.w;
int outh = top_blob.h;
const int tailstep = (w - 2 * outw + w) * 4;
#pragma omp parallel for num_threads(opt.num_threads)
... | subq $0x918, %rsp # imm = 0x918
movq %rdi, 0x338(%rsp)
movq %rsi, 0x330(%rsp)
movq %rdx, 0x328(%rsp)
movq 0x338(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x324(%rsp)
movq 0x338(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x320(%rsp)
movq 0x330(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x31c(%rsp)
movq... | /Tencent[P]ncnn/src/layer/x86/pooling_3x3_pack4.h |
ncnn::pooling2x2s2_max_pack4_sse(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void pooling2x2s2_max_pack4_sse(const Mat& bottom_blob, Mat& top_blob, const Option& opt)
{
int w = bottom_blob.w;
int inch = bottom_blob.c;
int outw = top_blob.w;
int outh = top_blob.h;
const int tailstep = (w - 2 * outw + w) * 4;
#pragma omp parallel for num_threads(opt.num_threads)
... | subq $0x3f8, %rsp # imm = 0x3F8
movq %rdi, 0x1b0(%rsp)
movq %rsi, 0x1a8(%rsp)
movq %rdx, 0x1a0(%rsp)
movq 0x1b0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x19c(%rsp)
movq 0x1b0(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x198(%rsp)
movq 0x1a8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x194(%rsp)
movq... | /Tencent[P]ncnn/src/layer/x86/pooling_2x2_pack4.h |
ncnn::pooling3x3s2_max_pack8_avx(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) | static void pooling3x3s2_max_pack8_avx(const Mat& bottom_blob, Mat& top_blob, const Option& opt)
{
int w = bottom_blob.w;
int inch = bottom_blob.c;
int outw = top_blob.w;
int outh = top_blob.h;
const int tailstep = (w - 2 * outw + w) * 8;
#pragma omp parallel for num_threads(opt.num_threads)
... | pushq %rbp
movq %rsp, %rbp
andq $-0x20, %rsp
subq $0x1cc0, %rsp # imm = 0x1CC0
movq %rdi, 0x980(%rsp)
movq %rsi, 0x978(%rsp)
movq %rdx, 0x970(%rsp)
movq 0x980(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x96c(%rsp)
movq 0x980(%rsp), %rax
movl 0x38(%rax), %eax
movl %eax, 0x968(%rsp)
movq 0x978(%rsp), %rax
mov... | /Tencent[P]ncnn/src/layer/x86/pooling_3x3_pack8.h |
ncnn::Power::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int Power::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 size = w * h;
#pragma omp parallel for num_threads(opt.num_threads)
for (int q = 0; q < channels; q++)
{
float* ptr... | subq $0x188, %rsp # imm = 0x188
movq %rdi, 0xb8(%rsp)
movq %rsi, 0xb0(%rsp)
movq %rdx, 0xa8(%rsp)
movq 0xb8(%rsp), %rax
movq %rax, 0x28(%rsp)
movq 0xb0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0xa4(%rsp)
movq 0xb0(%rsp), %rax
movl 0x30(%rax), %eax
movl %eax, 0xa0(%rsp)
movq 0xb0(%rsp), %rax
movl 0x38(%ra... | /Tencent[P]ncnn/src/layer/power.cpp |
ncnn::generate_anchors(int, ncnn::Mat const&, ncnn::Mat const&) | static Mat generate_anchors(int base_size, const Mat& ratios, const Mat& scales)
{
int num_ratio = ratios.w;
int num_scale = scales.w;
Mat anchors;
anchors.create(4, num_ratio * num_scale);
const float cx = base_size * 0.5f;
const float cy = base_size * 0.5f;
for (int i = 0; i < num_ratio... | subq $0x138, %rsp # imm = 0x138
movq %rdi, 0x38(%rsp)
movq %rdi, %rax
movq %rax, 0x40(%rsp)
movq %rdi, 0xb8(%rsp)
movl %esi, 0xb4(%rsp)
movq %rdx, 0xa8(%rsp)
movq %rcx, 0xa0(%rsp)
movq 0xa8(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x9c(%rsp)
movq 0xa0(%rsp), %rax
movl 0x2c(%rax), %eax
movl %eax, 0x98(%rs... | /Tencent[P]ncnn/src/layer/proposal.cpp |
ncnn::Reduction::forward(ncnn::Mat const&, ncnn::Mat&, ncnn::Option const&) const | int Reduction::forward(const Mat& bottom_blob, Mat& top_blob, const Option& opt) const
{
int dims = bottom_blob.dims;
int axes_flag[4] = {0};
bool reduce_w = false;
bool reduce_h = false;
bool reduce_d = false;
bool reduce_c = false;
if (reduce_all)
{
reduce_w = true;
re... | subq $0x88, %rsp
movq %rdi, 0x78(%rsp)
movq %rsi, 0x70(%rsp)
movq %rdx, 0x68(%rsp)
movq %rcx, 0x60(%rsp)
movq 0x78(%rsp), %rax
movq %rax, 0x18(%rsp)
movq 0x70(%rsp), %rax
movl 0x28(%rax), %eax
movl %eax, 0x5c(%rsp)
leaq 0x40(%rsp), %rdi
xorl %esi, %esi
movl $0x10, %edx
callq 0x24090
movq 0x18(%rsp), %rax
movb $0x0, 0x3... | /Tencent[P]ncnn/src/layer/reduction.cpp |
ncnn::reduction(float, float const*, int, int) | static float reduction(float v0, const float* ptr, int size, int op_type)
{
if (op_type == Reduction::ReductionOp_SUM) return reduction<reduction_op_add>(v0, ptr, size);
if (op_type == Reduction::ReductionOp_ASUM) return reduction<reduction_op_asum>(v0, ptr, size);
if (op_type == Reduction::ReductionOp_SUMS... | subq $0x18, %rsp
movss %xmm0, 0x10(%rsp)
movq %rdi, 0x8(%rsp)
movl %esi, 0x4(%rsp)
movl %edx, (%rsp)
cmpl $0x0, (%rsp)
jne 0xe78fcb
movss 0x10(%rsp), %xmm0
movq 0x8(%rsp), %rdi
movl 0x4(%rsp), %esi
callq 0xe79520
movss %xmm0, 0x14(%rsp)
jmp 0xe790a9
cmpl $0x1, (%rsp)
jne 0xe78ff0
movss 0x10(%rsp), %xmm0
movq 0x8(%rsp),... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_min>(float, float const*, int, int) | static float reduction(float v0, const float* ptr, int size, int stride)
{
Op op;
float sum = v0;
for (int i = 0; i < size; i++)
{
sum = op(sum, *ptr);
ptr += stride;
}
return sum;
} | subq $0x28, %rsp
movss %xmm0, 0x24(%rsp)
movq %rdi, 0x18(%rsp)
movl %esi, 0x14(%rsp)
movl %edx, 0x10(%rsp)
movss 0x24(%rsp), %xmm0
movss %xmm0, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x14(%rsp), %eax
jge 0xe79b24
movq 0x18(%rsp), %rdx
leaq 0xf(%rsp), %rdi
leaq 0x8(%rsp), %rsi
callq 0xe7a880
movss %xmm... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_sumexp>(float, float const*, int, int) | static float reduction(float v0, const float* ptr, int size, int stride)
{
Op op;
float sum = v0;
for (int i = 0; i < size; i++)
{
sum = op(sum, *ptr);
ptr += stride;
}
return sum;
} | subq $0x28, %rsp
movss %xmm0, 0x24(%rsp)
movq %rdi, 0x18(%rsp)
movl %esi, 0x14(%rsp)
movl %edx, 0x10(%rsp)
movss 0x24(%rsp), %xmm0
movss %xmm0, 0x8(%rsp)
movl $0x0, 0x4(%rsp)
movl 0x4(%rsp), %eax
cmpl 0x14(%rsp), %eax
jge 0xe79ba4
movq 0x18(%rsp), %rdx
leaq 0xf(%rsp), %rdi
leaq 0x8(%rsp), %rsi
callq 0xe7a8b0
movss %xmm... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_add>(float, float const*, int, int, int) | static float reduction(float v0, const float* ptr, int size0, int size1, int stride1)
{
Op op;
float sum = v0;
for (int i = 0; i < size1; i++)
{
for (int j = 0; j < size0; j++)
{
sum = op(sum, ptr[j]);
}
ptr += stride1;
}
return sum;
} | subq $0x38, %rsp
movss %xmm0, 0x34(%rsp)
movq %rdi, 0x28(%rsp)
movl %esi, 0x24(%rsp)
movl %edx, 0x20(%rsp)
movl %ecx, 0x1c(%rsp)
movss 0x34(%rsp), %xmm0
movss %xmm0, 0x14(%rsp)
movl $0x0, 0x10(%rsp)
movl 0x10(%rsp), %eax
cmpl 0x20(%rsp), %eax
jge 0xe79c53
movl $0x0, 0xc(%rsp)
movl 0xc(%rsp), %eax
cmpl 0x24(%rsp), %eax
... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_sumsq>(float, float const*, int, int, int) | static float reduction(float v0, const float* ptr, int size0, int size1, int stride1)
{
Op op;
float sum = v0;
for (int i = 0; i < size1; i++)
{
for (int j = 0; j < size0; j++)
{
sum = op(sum, ptr[j]);
}
ptr += stride1;
}
return sum;
} | subq $0x38, %rsp
movss %xmm0, 0x34(%rsp)
movq %rdi, 0x28(%rsp)
movl %esi, 0x24(%rsp)
movl %edx, 0x20(%rsp)
movl %ecx, 0x1c(%rsp)
movss 0x34(%rsp), %xmm0
movss %xmm0, 0x14(%rsp)
movl $0x0, 0x10(%rsp)
movl 0x10(%rsp), %eax
cmpl 0x20(%rsp), %eax
jge 0xe79db3
movl $0x0, 0xc(%rsp)
movl 0xc(%rsp), %eax
cmpl 0x24(%rsp), %eax
... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_mul>(float, float const*, int, int, int) | static float reduction(float v0, const float* ptr, int size0, int size1, int stride1)
{
Op op;
float sum = v0;
for (int i = 0; i < size1; i++)
{
for (int j = 0; j < size0; j++)
{
sum = op(sum, ptr[j]);
}
ptr += stride1;
}
return sum;
} | subq $0x38, %rsp
movss %xmm0, 0x34(%rsp)
movq %rdi, 0x28(%rsp)
movl %esi, 0x24(%rsp)
movl %edx, 0x20(%rsp)
movl %ecx, 0x1c(%rsp)
movss 0x34(%rsp), %xmm0
movss %xmm0, 0x14(%rsp)
movl $0x0, 0x10(%rsp)
movl 0x10(%rsp), %eax
cmpl 0x20(%rsp), %eax
jge 0xe79e63
movl $0x0, 0xc(%rsp)
movl 0xc(%rsp), %eax
cmpl 0x24(%rsp), %eax
... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_sumsq>(float, float const*, int, int, int, int) | static float reduction(float v0, const float* ptr, int size0, int stride0, int size1, int stride1)
{
Op op;
float sum = v0;
for (int i = 0; i < size1; i++)
{
const float* ptr0 = ptr;
for (int j = 0; j < size0; j++)
{
sum = op(sum, *ptr0);
ptr0 += stride0;... | subq $0x48, %rsp
movss %xmm0, 0x44(%rsp)
movq %rdi, 0x38(%rsp)
movl %esi, 0x34(%rsp)
movl %edx, 0x30(%rsp)
movl %ecx, 0x2c(%rsp)
movl %r8d, 0x28(%rsp)
movss 0x44(%rsp), %xmm0
movss %xmm0, 0x20(%rsp)
movl $0x0, 0x1c(%rsp)
movl 0x1c(%rsp), %eax
cmpl 0x2c(%rsp), %eax
jge 0xe7a2e5
movq 0x38(%rsp), %rax
movq %rax, 0x10(%rsp... | /Tencent[P]ncnn/src/layer/reduction.cpp |
float ncnn::reduction<ncnn::reduction_op_max>(float, float const*, int, int, int, int) | static float reduction(float v0, const float* ptr, int size0, int stride0, int size1, int stride1)
{
Op op;
float sum = v0;
for (int i = 0; i < size1; i++)
{
const float* ptr0 = ptr;
for (int j = 0; j < size0; j++)
{
sum = op(sum, *ptr0);
ptr0 += stride0;... | subq $0x48, %rsp
movss %xmm0, 0x44(%rsp)
movq %rdi, 0x38(%rsp)
movl %esi, 0x34(%rsp)
movl %edx, 0x30(%rsp)
movl %ecx, 0x2c(%rsp)
movl %r8d, 0x28(%rsp)
movss 0x44(%rsp), %xmm0
movss %xmm0, 0x20(%rsp)
movl $0x0, 0x1c(%rsp)
movl 0x1c(%rsp), %eax
cmpl 0x2c(%rsp), %eax
jge 0xe7a485
movq 0x38(%rsp), %rax
movq %rax, 0x10(%rsp... | /Tencent[P]ncnn/src/layer/reduction.cpp |
ncnn::ReLU::forward_inplace(ncnn::Mat&, ncnn::Option const&) const | int ReLU::forward_inplace(Mat& bottom_top_blob, const Option& opt) const
{
int w = bottom_top_blob.w;
int h = bottom_top_blob.h;
int d = bottom_top_blob.d;
int channels = bottom_top_blob.c;
int size = w * h * d;
if (slope == 0.f)
{
#pragma omp parallel for num_threads(opt.num_thread... | subq $0x2d8, %rsp # imm = 0x2D8
movq %rdi, 0x140(%rsp)
movq %rsi, 0x138(%rsp)
movq %rdx, 0x130(%rsp)
movq 0x140(%rsp), %rax
movq %rax, 0x50(%rsp)
movq 0x138(%rsp), %rcx
movl 0x2c(%rcx), %ecx
movl %ecx, 0x12c(%rsp)
movq 0x138(%rsp), %rcx
movl 0x30(%rcx), %ecx
movl %ecx, 0x128(%rsp)
movq 0x138(%rsp), %rcx
movl... | /Tencent[P]ncnn/src/layer/relu.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.