Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Updating cutlass to v3.1.0, give easy speedboosts to HeteroLinear #235

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 3 commits into from
May 31, 2023
Merged

Updating cutlass to v3.1.0, give easy speedboosts to HeteroLinear #235

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
3c7ecee
updating cutlass to v3.1.0, give easy speedboosts to heterolinear (up…
puririshi98 May 30, 2023
cf950ab
[pre-commit.ci] auto fixes from pre-commit.com hooks
pre-commit-ci[bot] May 30, 2023
9ca0358
update
rusty1s May 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
1 change: 1 addition & 0 deletions CHANGELOG.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,7 @@ The format is based on [Keep a Changelog](http://keepachangelog.com/en/1.0.0/).
- Added `metis` partitioning ([#229](https://github.com/pyg-team/pyg-lib/pull/229))
- Enable `hetero_neighbor_samplee` to work in parallel ([#211](https://github.com/pyg-team/pyg-lib/pull/211))
### Changed
- Updated `cutlass` version for speed boosts in `segment_matmul` and `grouped_matmul` ([#235](https://github.com/pyg-team/pyg-lib/pull/235))
- Drop nested tensor wrapper for `grouped_matmul` implementation ([#226](https://github.com/pyg-team/pyg-lib/pull/226))
- Fixed TorchScript support in `grouped_matmul` ([#220](https://github.com/pyg-team/pyg-lib/pull/220))
### Removed
Expand Down
15 changes: 5 additions & 10 deletions pyg_lib/csrc/ops/cuda/matmul_kernel.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -147,8 +147,7 @@ void grouped_matmul_out_kernel(const at::TensorList input,
float, 1, float, float>, //
cutlass::gemm::threadblock:: // Swizzling Operator
GemmIdentityThreadblockSwizzle<8>, //
2, // Stages
cutlass::arch::OpMultiplyAdd // Operation
2 // Stages
>::GemmKernel;
run_grouped_gemm<GemmKernel_Volta>(input, other, out, segment);
} else {
Expand Down Expand Up @@ -184,8 +183,7 @@ void grouped_matmul_out_kernel(const at::TensorList input,
float, 1, float, float>, //
cutlass::gemm::threadblock:: // Swizzling Operator
GemmIdentityThreadblockSwizzle<8>, //
3, // Stages
cutlass::arch::OpMultiplyAdd // Operation
3 // Stages
>::GemmKernel;
int grouped_shared_mem =
shared_memory_for_kernel<DefaultGemmKernel_TF32>();
Expand Down Expand Up @@ -217,8 +215,7 @@ void grouped_matmul_out_kernel(const at::TensorList input,
float, 1, float, float>, //
cutlass::gemm::threadblock:: // Swizzling Operator
GemmIdentityThreadblockSwizzle<8>, //
3, // Stages
cutlass::arch::OpMultiplyAdd // Operation
3 // Stages
>::GemmKernel;
run_grouped_gemm<SmallGemmKernel_TF32>(input, other, out, segment);
}
Expand Down Expand Up @@ -246,8 +243,7 @@ void grouped_matmul_out_kernel(const at::TensorList input,
float, 1, float, float>, //
cutlass::gemm::threadblock:: // Swizzling Operator
GemmIdentityThreadblockSwizzle<8>, //
3, // Stages
cutlass::arch::OpMultiplyAdd // Operation
3 // Stages
>::GemmKernel;
int grouped_shared_mem =
shared_memory_for_kernel<DefaultGemmKernel_FP32>();
Expand Down Expand Up @@ -279,8 +275,7 @@ void grouped_matmul_out_kernel(const at::TensorList input,
float, 1, float, float>, //
cutlass::gemm::threadblock:: // Swizzling Operator
GemmIdentityThreadblockSwizzle<8>, //
3, // Stages
cutlass::arch::OpMultiplyAdd // Operation
3 // Stages
>::GemmKernel;
run_grouped_gemm<SmallGemmKernel_FP32>(input, other, out, segment);
}
Expand Down
2 changes: 1 addition & 1 deletion setup.cfg
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,7 @@ classifiers=
test=pytest

[tool:pytest]
addopts=--capture=no
addopts=--capture=no --ignore=third_party

[flake8]
ignore=E731
Expand Down
2 changes: 1 addition & 1 deletion third_party/cutlass
Open in desktop
Submodule cutlass updated 1817 files