Skip to content
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

Jump to bottom

[ROCM] Jax/Pallas unit test, limiting tile to 64 in for pallas kernel in multihead attention test. #25043

Open
wants to merge 1 commit into
base: main
Choose a base branch
from
+6 −2
Open

[ROCM] Jax/Pallas unit test, limiting tile to 64 in for pallas kernel in multihead attention test. #25043

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
8 changes: 6 additions & 2 deletions tests/pallas/gpu_ops_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -284,8 +284,12 @@ def test_fused_attention_bwd(
segment_ids = None

def f(q, k, v):
return attention.mha(q, k, v, segment_ids, causal=causal,
interpret=self.INTERPRET).sum()
if jtu.is_device_rocm():
return attention.mha(q, k, v, segment_ids, 1.0, causal, 64, 64,
interpret=self.INTERPRET).sum()
else:
return attention.mha(q, k, v, segment_ids, causal=causal,
interpret=self.INTERPRET).sum()

def f_ref(q, k, v):
return attention.mha_reference(q, k, v, segment_ids, causal=causal).sum()
Expand Down