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.

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

Compat: fix float16(32)-renderable tests #4148

Merged
merged 3 commits into from
Jan 17, 2025

Conversation

shrekshao
Copy link
Contributor

Fix by requesting float16-renderable and float32-renderable for related tests in compat mode.

This CL only fixes a subset of the related test (operations)

@shrekshao shrekshao requested a review from greggman January 15, 2025 01:41
Copy link
Contributor

@greggman greggman left a comment

Choose a reason for hiding this comment

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

See comments.

I think the whole selectDeviceOrSkipTestCase system needs to be refactored.

Instead, off the top of my head,

  1. If no adapter has been requested then request one
  2. For each "selectXOrSkipTestCase" it should check the adapter, if no then skip, otherwise, add the requirements to some deviceDescriptor
  3. Only the end when the test actually starts (GPUTestSubcaseBatchState.postInit?) should it actually make the device request.

As it is, only one feature can be selected.

I'm not sure that's the best design. Maybe you have other ideas. But, somehow the requirements need to be accumulated.

src/webgpu/gpu_test.ts Outdated Show resolved Hide resolved
src/webgpu/api/operation/sampling/filter_mode.spec.ts Outdated Show resolved Hide resolved
src/webgpu/api/operation/sampling/filter_mode.spec.ts Outdated Show resolved Hide resolved
src/webgpu/api/operation/sampling/filter_mode.spec.ts Outdated Show resolved Hide resolved
src/webgpu/api/operation/sampling/filter_mode.spec.ts Outdated Show resolved Hide resolved
src/webgpu/api/operation/sampling/filter_mode.spec.ts Outdated Show resolved Hide resolved
Copy link
Contributor

@greggman greggman left a comment

Choose a reason for hiding this comment

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

with #4150 in this PR should work now

@shrekshao shrekshao force-pushed the float16-32-feature-fix branch from 6e707f8 to c02fc90 Compare January 17, 2025 02:56
@shrekshao
Copy link
Contributor Author

Thanks for the refactor!
I did the rename and revert some workaround needed before the fix.

@shrekshao shrekshao merged commit 1a434ee into gpuweb:main Jan 17, 2025
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants