leavelet/cutlass
1
0
Fork 0
Code Issues Pull Requests Actions 1 Packages Projects Releases Wiki Activity

CUTLASS 3.6.0 (#1850)

Browse Source 
* v3.6

* update changelog

* update readme

* fix typo

* fixing typos

* hopper gemm with weight prefetch

---------

Co-authored-by: yuzhai <yuzhai@nvidia.com>
Co-authored-by: Haicheng Wu <haichengw@nvidia.com>
This commit is contained in:
Yujia Zhai
2024-10-09 12:33:27 -07:00
committed by GitHub
parent 0837a2a00a
commit cc3c29a81a
354 changed files with 105943 additions and 8203 deletions
Download Patch File Download Diff File Expand all files Collapse all files

42
tools/profiler/src/performance_report.cpp
View File

@@ -94,7 +94,7 @@ PerformanceReport::PerformanceReport(
if (options_.report.append) {
std::ifstream test_output_file(op_file_name_);
if (test_output_file.is_open()) {
print_header = false;
test_output_file.close();
@@ -145,7 +145,7 @@ void PerformanceReport::append_result(PerformanceResult result) {
if (options_.report.verbose) {
std::cout << "\n";
print_result_pretty_(std::cout, result) << std::flush;
print_result_pretty_(std::cout, result) << std::flush;
}
if (junit_output_file_.is_open()) {
@@ -237,7 +237,7 @@ static const char *disposition_status_color(Disposition disposition) {
/// Prints the result in human readable form
std::ostream & PerformanceReport::print_result_pretty_(
std::ostream &out,
std::ostream &out,
PerformanceResult const &result,
bool use_shell_coloring) {
@@ -251,14 +251,14 @@ std::ostream & PerformanceReport::print_result_pretty_(
int column_idx = 0;
for (auto const & tag : options_.report.pivot_tags) {
out << (column_idx++ ? "," : "") << tag.first << ":" << tag.second;
}
}
out << "\n";
}
std::string shell_color_bright = use_shell_coloring ? SHELL_COLOR_BRIGHT() : "";
std::string shell_color_end = use_shell_coloring ? SHELL_COLOR_END() : "";
auto _disposition_status_color = [&](Disposition d) -> const char * {
auto _disposition_status_color = [&](Disposition d) -> const char * {
return use_shell_coloring ? disposition_status_color(d) : "";
};
@@ -277,7 +277,7 @@ std::ostream & PerformanceReport::print_result_pretty_(
static int const indent_spaces = 16;
for(auto & m : result.verification_map) {
out << std::right << std::setw(indent_spaces) << library::to_string(m.first, true) << ": " << to_string(m.second, true) << "\n";
out << std::right << std::setw(indent_spaces) << library::to_string(m.first, true) << ": " << to_string(m.second, true) << "\n";
}
}
@@ -287,7 +287,7 @@ std::ostream & PerformanceReport::print_result_pretty_(
int column_idx = 0;
for (auto const &arg : result.arguments) {
if (!arg.second.empty()) {
out << " --" << arg.first << "=" << arg.second;
out << " --" << arg.first << "=" << arg.second;
column_idx += int(4 + arg.first.size() + arg.second.size());
if (column_idx > 98) {
out << " \\\n ";
@@ -297,7 +297,7 @@ std::ostream & PerformanceReport::print_result_pretty_(
}
out << "\n\n";
out
out
<< " Bytes: " << result.bytes << " bytes\n"
<< " FLOPs: " << result.flops << " flops\n"
<< " FLOPs/Byte: " << (result.flops / result.bytes) << "\n\n";
@@ -325,7 +325,7 @@ std::ostream & PerformanceReport::print_csv_header_(
out << (column_idx++ ? "," : "") << tag.first;
}
out
out
<< (column_idx ? "," : "") << "Problem,Provider"
<< ",OperationKind,Operation,Disposition,Status";
@@ -333,7 +333,7 @@ std::ostream & PerformanceReport::print_csv_header_(
out << "," << arg_name;
}
out
out
<< ",Bytes"
<< ",Flops"
<< ",Flops/Byte"
@@ -347,7 +347,7 @@ std::ostream & PerformanceReport::print_csv_header_(
/// Print the result in CSV output
std::ostream & PerformanceReport::print_result_csv_(
std::ostream &out,
std::ostream &out,
PerformanceResult const &result) {
int column_idx = 0;
@@ -357,8 +357,8 @@ std::ostream & PerformanceReport::print_result_csv_(
out << (column_idx++ ? "," : "") << tag.second;
}
out
<< (column_idx ? "," : "")
out
<< (column_idx ? "," : "")
<< result.problem_index
<< "," << to_string(result.provider, true)
<< "," << to_string(result.op_kind)
@@ -370,7 +370,7 @@ std::ostream & PerformanceReport::print_result_csv_(
out << "," << arg.second;
}
out
out
<< "," << result.bytes
<< "," << result.flops
<< "," << result.flops / result.bytes
@@ -387,7 +387,7 @@ std::ostream & PerformanceReport::print_result_csv_(
else {
out << std::string(2
, ','
);
);
}
return out;
@@ -451,25 +451,25 @@ std::ostream & PerformanceReport::print_junit_result_(std::ostream &out, Perform
case Disposition::kNotSupported:
skipped = true;
break;
case Disposition::kPassed:
case Disposition::kPassed:
case Disposition::kNotVerified:
break;
case Disposition::kFailed:
case Disposition::kFailed:
case Disposition::kIncorrect:
failed = true;
failed = true;
break;
case Disposition::kInvalidProblem:
case Disposition::kInvalid:
error = true;
break;
};
if (skipped) {
out << "status=\"notrun\"";
} else {
out << "status=\"run\"";
}
out << ">" << std::endl;
if (failed) {
@@ -488,7 +488,7 @@ std::ostream & PerformanceReport::print_junit_result_(std::ostream &out, Perform
out << " </testcase>" << std::endl;
return out;
return out;
}