This is an archive of the discontinued LLVM Phabricator instance.

[AMDGPU][True16] Pre-commit addition tests.
ClosedPublic

Authored by kosarev on Jul 28 2023, 6:02 AM.

Diff Detail

Event Timeline

kosarev created this revision.Jul 28 2023, 6:02 AM
Herald added a project: Restricted Project. · View Herald TranscriptJul 28 2023, 6:02 AM
kosarev requested review of this revision.Jul 28 2023, 6:02 AM
Herald added a project: Restricted Project. · View Herald TranscriptJul 28 2023, 6:02 AM

After playing with it a bit, it feels like cloning whole test files is too much clutter and adding RUN lines along with functional changes without pre-committed tests is uncomfortably unsafe, so maybe doing it the usual way is still the best option.

foad added a comment.Jul 28 2023, 6:16 AM

After playing with it a bit, it feels like cloning whole test files is too much clutter and adding RUN lines along with functional changes without pre-committed tests is uncomfortably unsafe, so maybe doing it the usual way is still the best option.

Good, I agree.

Could we have a minimal 16-bit test for the bringup? One that tests all the basic 16-bit ops work on both paths, but maybe doesn't bother trying to handle all the vector types and other complexities

Could we have a minimal 16-bit test for the bringup? One that tests all the basic 16-bit ops work on both paths, but maybe doesn't bother trying to handle all the vector types and other complexities

Can you describe that a bit more? I would say the existing tests in tree have coverage of MC and Codegen for the fake16 and previous asic paths. And then it seems adding tests for the true16 path would mean a lot of them are xfail right now. But I do think more VOP2 true16 instructions than just add_f16 are functioning after the current patch stack. Perhaps we could add tests for those, such as mul_f16, sub_f16 etc.

Could we have a minimal 16-bit test for the bringup? One that tests all the basic 16-bit ops work on both paths, but maybe doesn't bother trying to handle all the vector types and other complexities

Can you describe that a bit more? I would say the existing tests in tree have coverage of MC and Codegen for the fake16 and previous asic paths. And then it seems adding tests for the true16 path would mean a lot of them are xfail right now. But I do think more VOP2 true16 instructions than just add_f16 are functioning after the current patch stack. Perhaps we could add tests for those, such as mul_f16, sub_f16 etc.

I mean like for wave32, we have test/CodeGen/AMDGPU/wave32.ll. It's not completely comprehensive, but it kept testing isolated when wave32 support was first added. At this point a lot of what's in it is probably redundant and duplicated in other tests

I mean like for wave32, we have test/CodeGen/AMDGPU/wave32.ll. It's not completely comprehensive, but it kept testing isolated when wave32 support was first added. At this point a lot of what's in it is probably redundant and duplicated in other tests

As Joe mentioned, we already have lots of t16 tests here, currently covering the support for fake instructions, but identical to what we would want for real ones, so feels like no need for special bring-up tests.

This revision is now accepted and ready to land.Aug 17 2023, 12:19 PM
This revision was automatically updated to reflect the committed changes.