CTS: Color space conversion operation test for CopyToTexture

[shaoboyan]

This PR add cts to test copyExternalImageToTexture() color space
conversion ability.
It creates canvas with color space attr and copy to dst texture in
user defined color space through CopyExternalImageToTexture().

Issue: #913

---

Requirements for PR author:

• All missing test coverage is tracked with "TODO" or .unimplemented().
• New helpers are /** documented */ and new helper files are found in helper_index.txt.
• Test behaves as expected in a WebGPU implementation. (If not passing, explain above.)

Requirements for reviewer sign-off:

• Tests are properly located in the test tree.
• Test descriptions allow a reader to "read only the test plans and evaluate coverage completeness", and accurately reflect the test code.
• Tests provide complete coverage (including validation control cases). Missing coverage MUST be covered by TODOs.
• Helpers and types promote readability and maintainability.

When landing this PR, be sure to make any necessary issue status updates.

[shaoboyan]

The compatible Blink CL is here https://chromium-review.googlesource.com/c/chromium/src/+/3501779

[austinEng]

Some comments - though I did not actually review the test content. I think Kai would be a better reviewer here.

[shaoboyan]

@austinEng I try to upgrade tsc but got lots of errors. Maybe we should use a seperate PR to do the upgrading.

[kainino0x]

I'm working on a tsc upgrade, should be able to get it done today.

[kainino0x]

TypeScript update landed, please merge main into your branch.

[kainino0x]

LGTM w/ a few nits

[github-actions]

Previews, as seen when this build job started (8702c63):
Run tests | View tsdoc

[shaoboyan]

One problem here:
The test runs super slow(I think ~ a minute) with f16 formats.

[kainino0x]

The test runs super slow(I think ~ a minute) with f16 formats.

Do you know what causes this? If it's slowness in the implementation, we can mark the test as slow in our expectations (hopefully - though we should consider doing so before landing this PR so that the Chrome WebGPU Gardener won't have to debug it).
If it's slowness in the test we should see if we can optimize it. I'm guessing that the float16BitsToFloat32 helper is a bottleneck.

[github-actions]

Previews, as seen when this build job started (5a0b912):
Run tests | View tsdoc

[shaoboyan]

I'm guessing that the float16BitsToFloat32 helper is a bottleneck.

Yes, and the actual and expected arrays needs to reinterpret to Uin16Array first and then call float16ToFloat32 for each entries.
Since the other format works fine, I think this is the bottleneck.

[kainino0x]

Apparently I forgot to complete my re-review, sorry.

Since the other format works fine, I think this is the bottleneck.

OK. Since we know it's slow, do you think we could disable the f16 tests in this PR so we can land it, and re-enable them (hopefully with optimizations) in a separate PR?

[shaoboyan]

OK. Since we know it's slow, do you think we could disable the f16 tests in this PR so we can land it, and re-enable them (hopefully with optimizations) in a separate PR?

The f16 expect logic will cover more than color_space conversion cases. So I try to fix it by comparing the f16bits and handling ~0 values. PTAL again, thanks!

[github-actions]

Previews, as seen when this build job started (7481681):
Run tests | View tsdoc

[shaoboyan]

After this fixing, f16 format cases could finished in ~4s on my local machine.

[kainino0x]

Nice fix for both the tolerance and the performance! and simpler too