Add ONNX Runtime GQA-style SDPA benchmark

[kimishpatel]

Stack from ghstack (oldest at bottom):

• temp changes to enable benchmarking on device via adb #18719
• Add xplat cxx_binary target for bench_transposed_cache benchmark #18718
• Add non-flash SDPA path gated by ET_USE_UNFUSED_SDPA #18717
• Add ONNX Runtime GQA-style SDPA benchmark #18716
• Add GEMM-based standard SDPA benchmark and fix custom_sdpa_out signature #18715
• Support separate K/V seq dim in custom_sdpa op #18714
• Add --no_transposed_cache CLI flag for export pipeline #18713
• Plumb transposed cache config through export pipeline #18712
• Add benchmarks for transposed vs standard KV cache layout #18711
• temp hack #18710
• Add support for transposed kv cache in all variants of custom kv cache #18709
• Add contig requirement to aot binding for llm custom ops #18708
• Add support for seq at dim=2 #18707
• Make kv cache update op support transposed cache #18706
• Add xplat cxx_binary target for bench_sdpa #18705
• Use unfused SDPA for short sequences (q_len <= 128 or kv_len <= 128) #18651
• Limit softmax to causally-valid elements in cpu_sdpa #18650
• Add quantized input support to cpu_sdpa #18649
• [Executorch] Add non-flash SDPA for decode #18648
• -> Add ONNX Runtime GQA-style SDPA benchmark #18647
• Add GEMM-based standard SDPA benchmark #18646
• Fix SLEEF preprocessor macro name to match ATen vec headers #18645

Add run_onnx_gqa_sdpa() which faithfully ports the algorithm from
onnxruntime/contrib_ops/cpu/bert/gqa_attention_base.h:

• Scale baked into GEMM alpha (no separate scaling pass)
• Scores buffer padded to max_seq_len columns
• Causal mask: zero out future positions, softmax on valid window only
• Output always in [B, S, Hq, D] format

Extends validation to verify ONNX GQA output matches custom_sdpa_out
reference. Adds OnnxGQABenchFixture for benchmarking both layouts.

Differential Revision: D96044317

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18647

• 📄 Preview Python docs built from this PR

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

❌ 3 New Failures, 3 Cancelled Jobs

As of commit 8b9b1ae with merge base fb1618e ():

NEW FAILURES - The following jobs have failed:

• Lint / lintrunner (gh)
  >>> Lint for extension/llm/custom_ops/bench_sdpa.cpp:
• pull / test-multimodal-linux (gemma3-4b) / linux-job (gh)
  RuntimeError: Command docker exec -t 900b7704efcfefeab20abd4bc17d0b33d8f2b495a2012822f63d06ec735c7cc4 /exec failed with exit code 139
• pull / unittest / linux / linux-job (gh)
  examples/models/test/test_export.py::ExportTest::test_mv3_export_to_executorch

CANCELLED JOBS - The following jobs were cancelled. Please retry:

• pull / unittest / macos / macos-job (gh)
• pull / unittest / windows / windows-job (gh)
• pull / unittest-editable / windows / windows-job (gh)

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.

[digantdesai]

Review automatically exported from Phabricator review in Meta.