Skip to content

[libc] Add N Threads Benchmark Helper #99834

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jul 22, 2024
Merged

[libc] Add N Threads Benchmark Helper #99834

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
5 changes: 1 addition & 4 deletions libc/benchmarks/gpu/LibcGpuBenchmark.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -151,10 +151,7 @@ void Benchmark::run_benchmarks() {
all_results.reset();

gpu::sync_threads();
if (!b->flags ||
((b->flags & BenchmarkFlags::SINGLE_THREADED) && id == 0) ||
((b->flags & BenchmarkFlags::SINGLE_WAVE) &&
id < gpu::get_lane_size())) {
if (b->num_threads == static_cast<uint32_t>(-1) || id < b->num_threads) {
auto current_result = b->run();
all_results.update(current_result);
}
Expand Down
26 changes: 14 additions & 12 deletions libc/benchmarks/gpu/LibcGpuBenchmark.h
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -74,21 +74,20 @@ struct BenchmarkResult {
clock_t total_time = 0;
};

enum BenchmarkFlags { SINGLE_THREADED = 0x1, SINGLE_WAVE = 0x2 };

BenchmarkResult benchmark(const BenchmarkOptions &options,
cpp::function<uint64_t(void)> wrapper_func);

class Benchmark {
const cpp::function<uint64_t(void)> func;
const cpp::string_view suite_name;
const cpp::string_view test_name;
const uint8_t flags;
const uint32_t num_threads;

public:
Benchmark(cpp::function<uint64_t(void)> func, char const *suite_name,
char const *test_name, uint8_t flags)
: func(func), suite_name(suite_name), test_name(test_name), flags(flags) {
char const *test_name, uint32_t num_threads)
: func(func), suite_name(suite_name), test_name(test_name),
num_threads(num_threads) {
add_benchmark(this);
}

Expand All @@ -108,18 +107,21 @@ class Benchmark {
} // namespace benchmarks
} // namespace LIBC_NAMESPACE_DECL

// Passing -1 indicates the benchmark should be run with as many threads as
// allocated by the user in the benchmark's CMake.
#define BENCHMARK(SuiteName, TestName, Func) \
LIBC_NAMESPACE::benchmarks::Benchmark SuiteName##_##TestName##_Instance( \
Func, #SuiteName, #TestName, 0)
Func, #SuiteName, #TestName, -1)

#define SINGLE_THREADED_BENCHMARK(SuiteName, TestName, Func) \
#define BENCHMARK_N_THREADS(SuiteName, TestName, Func, NumThreads) \
LIBC_NAMESPACE::benchmarks::Benchmark SuiteName##_##TestName##_Instance( \
Func, #SuiteName, #TestName, \
LIBC_NAMESPACE::benchmarks::BenchmarkFlags::SINGLE_THREADED)
Func, #SuiteName, #TestName, NumThreads)

#define SINGLE_THREADED_BENCHMARK(SuiteName, TestName, Func) \
BENCHMARK_N_THREADS(SuiteName, TestName, Func, 1)

#define SINGLE_WAVE_BENCHMARK(SuiteName, TestName, Func) \
LIBC_NAMESPACE::benchmarks::Benchmark SuiteName##_##TestName##_Instance( \
Func, #SuiteName, #TestName, \
LIBC_NAMESPACE::benchmarks::BenchmarkFlags::SINGLE_WAVE)
BENCHMARK_N_THREADS(SuiteName, TestName, Func, \
LIBC_NAMESPACE::gpu::get_lane_size())

#endif
Loading