Skip to content

[Inference/Kernel] Feat rmsnorm baichuan2 #5611

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
Apr 19, 2024
+9 −3
Merged

[Inference/Kernel] Feat rmsnorm baichuan2 #5611

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
2 changes: 1 addition & 1 deletion examples/inference/benchmark_ops/benchmark_rmsnorm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -35,7 +35,7 @@
styles=[("red", "-"), ("blue", "-"), ("yellow", "-"), ("red", "--"), ("blue", "--"), ("yellow", "--")],
ylabel="ms",
plot_name=f"RMSNorm benchmarking results",
args={"HIDDEN_SIZE": 1024},
args={"HIDDEN_SIZE": 5120},
)
]

Expand Down
6 changes: 6 additions & 0 deletions extensions/csrc/cuda/rms_layernorm_kernel.cu
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -277,6 +277,9 @@ void rms_layernorm(
case 2:
RMSNORM_LAUNCHER(2, block);
break;
case 3:
RMSNORM_LAUNCHER(3, block);
break;
case 4:
RMSNORM_LAUNCHER(4, block);
break;
Expand Down Expand Up @@ -321,6 +324,9 @@ void fused_add_rms_layernorm(
case 2:
FUSED_ADD_RMSNORM_LAUNCHER(2, block);
break;
case 3:
FUSED_ADD_RMSNORM_LAUNCHER(3, block);
break;
case 4:
FUSED_ADD_RMSNORM_LAUNCHER(4, block);
break;
Expand Down
4 changes: 2 additions & 2 deletions tests/test_infer/test_ops/cuda/test_rms_layernorm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@


@pytest.mark.parametrize("M", [2, 4, 8, 16])
@pytest.mark.parametrize("N", [64, 128, 512])
@pytest.mark.parametrize("N", [64, 128, 512, 5120])
def test_rms_layernorm(M: int, N: int):
torch.manual_seed(123)
torch.cuda.empty_cache()
Expand Down Expand Up @@ -48,4 +48,4 @@ def test_rms_layernorm(M: int, N: int):


if __name__ == "__main__":
test_rms_layernorm(16, 512)
test_rms_layernorm(16, 5120)
Loading