WebGPU: Add indirect dispatch for flash attention graph capture

[feich-ms]

Summary

• Adds `PrepareIndirectDispatchProgram` to compute dispatch group sizes on GPU from `seqlen_k` when total_sequence_length is unavailable on CPU
• Extends indirect dispatch to kv_empty layers (shared KV layers where `total_sequence_length_=0` would cause `dispatch(0)` crashes)
• Enables graph capture for models with dynamic sequence lengths (e.g., Gemma4), achieving ~25% decode throughput improvement on WebGPU

Changes

• `flash_attention.cc`:
  • New `PrepareIndirectDispatchProgram::GenerateShaderCode` for GPU-side dispatch sizing on kv_empty layers
  • Extracted `AppendNormalizeDispatchShader()` helper shared by `CopyKVCacheProgram` and `PrepareIndirectDispatchProgram` to eliminate duplicated tile-count + `normalize_dispatch_group_size` logic
  • `use_indirect_dispatch` expressed as `use_seqlen_k && (share_buffer || kv_empty)` to make the subset relationship explicit
  • Tightened `use_seqlen_k` / `use_indirect_dispatch` guards from `== 0` to `<= 0` for defensive handling of negative `total_sequence_length_`
• `flash_attention.h`: Added `PrepareIndirectDispatchProgram` class declaration

Test plan

• Verified with Gemma4 INT4 model: ~95-105 tok/s with GC=ON vs ~75 tok/s with GC=OFF
• Tested multiple prompts (short/long) with graph capture enabled — no crashes, correct output
• Verified warmup (capture run) followed by replay produces consistent results

[Copilot]

Pull request overview

This PR improves the WebGPU FlashAttention decode path’s compatibility with graph capture and dynamic sequence lengths by enabling GPU-side computation of indirect dispatch group sizes from seqlen_k (including the kv_empty/shared-KV case that previously could produce dispatch(0)).

Changes:

• Added a dedicated PrepareIndirectDispatchProgram shader to compute normalized indirect dispatch dimensions on GPU from seqlen_k.
• Expanded use_seqlen_k / use_indirect_dispatch gating so indirect dispatch is also available when total_sequence_length_ is unavailable on CPU (including kv_empty layers).
• Ensured the kv_empty path prepares the indirect dispatch buffer even when CopyKVCache is skipped.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
onnxruntime/contrib_ops/webgpu/bert/flash_attention.h	Declares PrepareIndirectDispatchProgram and its uniform interface.
onnxruntime/contrib_ops/webgpu/bert/flash_attention.cc	Implements the new program and wires indirect-dispatch preparation into the kv_empty path and broadened gating logic.

[github-actions]

You can commit the suggested changes from lintrunner.