Enable backend test suite + x86 CI (#19964)

[JulianCloudNTH]

Summary:

Wires the WebGPU backend into the standard ExecuTorch backend test suite and adds an x86 Linux CI job, mirroring the Vulkan delegate: backends/test/suite/flows/webgpu.py plus a WebGPUTester, run by oss/.github/workflows/test-backend-webgpu.yml on SwiftShader (a software Vulkan adapter, via wgpu-native, minimal dependencies, no GPU).

Two fixes were needed for SwiftShader's downlevel limits: request the adapter's full requiredLimits at device creation (software adapters default storage-buffer limits to 0), and make the add op's workgroup size dynamic instead of a hardcoded constant. The WGSL now declares a pipeline-overridable override wg_size: u32 = 256 and the host clamps it to the device's maxComputeInvocationsPerWorkgroup (256 on real GPUs and lavapipe, 128 on SwiftShader), so SwiftShader's 128-invocation cap no longer forces a smaller workgroup size on real hardware. This mirrors the dynamic-workgroup-sizing approach in D107259348 and opens the door to selecting device/algorithm-optimal sizes later. The add op also validates its 1D dispatch count before allocating any GPU objects, against the device's queried maxComputeWorkgroupsPerDimension (falling back to the WebGPU spec-default floor of 65535 only when the limit query fails). Per Stephen's review, the workgroup-size clamp and the dispatch-count computation are factored into reusable inline helpers in runtime/WebGPUUtils.h (clamp_workgroup_size and compute_1d_workgroup_count, mirroring the Vulkan delegate's utils::div_up) so the other ops can share them rather than re-inlining the logic. The editable CMake build additionally marks the vulkan_schema subdirectory EXCLUDE_FROM_ALL so the WebGPU ALL build does not pull in targets that need glslc.
ghstack-source-id: 389222646
exported-using-ghexport

Differential Revision: D107288999

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19986

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❌ You can merge normally! (1 Unrelated Failure), 2 Unclassified Failures

As of commit 2310c6d with merge base 22a2daf ():

UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:

• Test WebGPU Backend / test-webgpu / test-backend-linux (webgpu, models) / linux-job (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  RuntimeError: Command docker exec -t 69fd10082d8df717b86f19fc4de9baa738133808de8394a51c2bade1df2ed0ad /exec failed with exit code 1
• Test WebGPU Backend / test-webgpu / test-backend-linux (webgpu, operators) / linux-job (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  RuntimeError: Command docker exec -t 9db1d29c566252c208ac1b9a76d6e1430266066462cfaf2bf67bcbe847411c99 /exec failed with exit code 1

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• trunk / test-arm-backend-ethos-u (test_smaller_stories_llama) / linux-job (gh) (trunk failure)

This comment was automatically generated by Dr. CI and updates every 15 minutes.

[github-actions]

This PR needs a release notes: label

If your change should be included in the release notes (i.e. would users of this library care about this change?), please use a label starting with release notes:. This helps us keep track and include your important work in the next release notes.

To add a label, you can comment to pytorchbot, for example
@pytorchbot label "release notes: none"

For more information, see
https://github.com/pytorch/pytorch/wiki/PyTorch-AutoLabel-Bot#why-categorize-for-release-notes-and-how-does-it-work.