[ExecuTorch][WebGPU] Add certified scoped output suppression

[JCNTH]

Stack from ghstack (oldest at bottom):

• [ExecuTorch][WebGPU] Preserve Linear and Q4 embedding params across dynamic resize #21483
• [ExecuTorch][WebGPU] Generate logical and arithmetic binary shader variants #21451
• [ExecuTorch][WebGPU] Generate extrema and unary shader variants #21450
• [ExecuTorch][WebGPU] Generate typed to_copy shader variants #21449
• [ExecuTorch][WebGPU] Make WGSL generation rollback-safe #21448
• [ExecuTorch][WebGPU] Add Qwen3 K16 attention and fp16 KV cache #21136
• [ExecuTorch][WebGPU] Add HuggingFace rotate-half RoPE #21135
• [ExecuTorch][WebGPU] Repair dynamic SDPA routing, add attestation #21134
• [ExecuTorch][WebGPU] Select emdawnwebgpu for Emscripten builds #21133
• [ExecuTorch][WebGPU] Add Llama K16 online causal attention #21132
• [ExecuTorch][WebGPU] Replace broad QKV fusion with BK64 kernel #21131
• [ExecuTorch][WebGPU] Route exact prefill shapes to the BK64 GEMM #21130
• [ExecuTorch][WebGPU] Fuse the SwiGLU elementwise chain #21129
• [ExecuTorch][WebGPU] Coordinate dynamic dispatch routes #21128
• -> [ExecuTorch][WebGPU] Add certified scoped output suppression #21127

When a caller discards a model's terminal output tensor, the backend still
dispatches the final quantized-linear range and reads it back, wasting a
dispatch and a device-to-host copy on every execution. This adds a default-off,
certificate-bound execution option that suppresses exactly one explicitly
certified terminal output dispatch range and its readback. The certificate must
bind the exact PTE and method and prove a single delegate, no portable nodes,
and a unique leaf output resolving to the caller's host buffer; anything
ambiguous, aliased, portable, or uncertified fails closed and executes
normally. No Vulkan analogue (WebGPU runtime output-contract change).

Key changes:

• WebGPUExecutionOptions.{h,cpp}: certificate struct, RAII
  ScopedWebGPUExecutionOptions, and plan_webgpu_execution suppression planner.
• WebGPUGraph.{cpp,h}: resolve the certified output ordinal and skip its
  dispatch range and readback when suppression is active.
• WebGPUBackend.cpp: thread the scoped option through execution.
  @exported-using-ghexport

Differential Revision: D113171751

Differential Revision: D113171751

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21127

• 📄 Preview Python docs built from this PR

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

✅ No Failures

As of commit 32d2543 with merge base 8200b9e ():
💚 Looks good so far! There are no failures yet. 💚

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.