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

Fix FA tutorial #485

Merged
merged 1 commit into from
Jan 25, 2024
Merged

Fix FA tutorial #485

merged 1 commit into from
Jan 25, 2024

Conversation

zhanglx13
Copy link

  • Check correctness for fp8 inputs only when torch supports it
  • Only run benchmark in fp16

@zhanglx13
Fix FA tutorial
c9e0b1e 
- Check correctness for fp8 inputs only when torch supports it
- Only run benchmark in fp16
@xiaohuguo2023
Copy link
Member

passed on my test on MI250:
collected 28 items
Running 28 items in this shard: tutorials/06-fused-attention.py::test_op_fwd[False-4-48-1024-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-1024-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-2048-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-2048-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-4096-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-4096-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-1024-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-1024-128-fp8], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-2048-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-2048-128-fp8], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-4096-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[False-4-48-4096-128-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-1024-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-1024-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-2048-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-2048-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-4096-64-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-4096-64-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-1024-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-1024-128-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-2048-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-2048-128-fp8], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-4096-128-fp16], tutorials/06-fused-attention.py::test_op_fwd[True-4-48-4096-128-fp8], tutorials/06-fused-attention.py::test_op_bwd[4-48-1024-64], tutorials/06-fused-attention.py::test_op_bwd[4-48-2048-64], tutorials/06-fused-attention.py::test_op_bwd[4-48-4096-64], tutorials/06-fused-attention.py::test_op_bwd[1-16-8192-64]

06-fused-attention.py::test_op_fwd[False-4-48-1024-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-1024-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[False-4-48-2048-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-2048-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[False-4-48-4096-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-4096-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[False-4-48-1024-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-1024-128-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[False-4-48-2048-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-2048-128-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[False-4-48-4096-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[False-4-48-4096-128-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-1024-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-1024-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-2048-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-2048-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-4096-64-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-4096-64-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-1024-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-1024-128-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-2048-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-2048-128-fp8] SKIPPED
06-fused-attention.py::test_op_fwd[True-4-48-4096-128-fp16] PASSED
06-fused-attention.py::test_op_fwd[True-4-48-4096-128-fp8] SKIPPED
06-fused-attention.py::test_op_bwd[4-48-1024-64] PASSED
06-fused-attention.py::test_op_bwd[4-48-2048-64] PASSED
06-fused-attention.py::test_op_bwd[4-48-4096-64] PASSED
06-fused-attention.py::test_op_bwd[1-16-8192-64] PASSED

================= 16 passed, 12 skipped in 340.48s (0:05:40) =================

xiaohuguo2023
xiaohuguo2023 approved these changes Jan 25, 2024
View reviewed changes
Copy link
Member

@xiaohuguo2023 xiaohuguo2023 left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

good to go

@zhanglx13 zhanglx13 merged commit c631824 into triton-mlir Jan 25, 2024
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@xiaohuguo2023 xiaohuguo2023 xiaohuguo2023 approved these changes

@vgokhale vgokhale Awaiting requested review from vgokhale

@scxiao scxiao Awaiting requested review from scxiao

Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
2 participants
@zhanglx13 @xiaohuguo2023