Skip to content

fix benchmark GeluBackward #102

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
liqiangxl merged 2 commits into from
Mar 31, 2023
Merged

fix benchmark GeluBackward #102

Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
3693fc8
fix benchmark GeluBackward
liqiangxl Mar 30, 2023
3884236
Merge branch 'main' into llu/fix_benchmark_gelu
liqiangxl Mar 31, 2023
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
10 changes: 5 additions & 5 deletions benchmark/gelu_backward.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -146,11 +146,11 @@ static void GeluBackward_Compile(benchmark::State& benchmark_state) {
// inputs
std::vector<c10::IValue> inputs = setupInputs();

schedulePointwise(&fusion, c10::ArrayRef<c10::IValue>(inputs));
auto lparams = schedulePointwise(&fusion, c10::ArrayRef<c10::IValue>(inputs));

for (auto _ : benchmark_state) {
FusionExecutor executor;
executor.compileFusion(&fusion, inputs);
executor.compileFusion(&fusion, inputs, lparams);
}
}

Expand All @@ -173,7 +173,7 @@ static void GeluBackward_RunFusion(benchmark::State& benchmark_state) {
auto lparams = schedulePointwise(&fusion, c10::ArrayRef<c10::IValue>(inputs));

FusionExecutor executor;
executor.compileFusion(&fusion, inputs);
executor.compileFusion(&fusion, inputs, lparams);

C10_CUDA_CHECK(cudaDeviceSynchronize());

Expand Down Expand Up @@ -204,7 +204,7 @@ static void GeluBackward_RunFusion_GpuOnly(benchmark::State& benchmark_state) {

FusionExecutor executor;
executor.setMeasureKernelTimeFlag(true);
executor.compileFusion(&fusion, inputs);
executor.compileFusion(&fusion, inputs, lparams);

C10_CUDA_CHECK(cudaDeviceSynchronize());

Expand Down Expand Up @@ -237,7 +237,7 @@ static void GeluBackward_RunFusion_CpuOnly(benchmark::State& benchmark_state) {

FusionExecutor executor;
executor.setExecuteKernelFlag(false);
executor.compileFusion(&fusion, inputs);
executor.compileFusion(&fusion, inputs, lparams);

for (auto _ : benchmark_state) {
outputs = executor.runFusion(c10::ArrayRef<c10::IValue>(inputs), lparams);
Expand Down