Plumb transposed cache config through export pipeline

[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

Benchmarking shows transposed KV cache [B, H, S, D] significantly outperforms
standard layout [B, S, H, D] in custom_sdpa, especially at longer cache fills:
1.64x at start_pos=1024, 1.14x at start_pos=512, 1.13x for prefill seq_len=512
(Llama 3 8B config, Apple M-series). The improvement comes from better memory
locality in the attn_score @ V GEMM where V stride along S_kv changes from
H*D to D.

This commit replaces the hardcoded is_seq_at_dim_2=True # hacking temporarily
values in sdpa.py and custom_kv_cache.py with a proper configurable parameter
threaded through the export pipeline:

• Add use_transposed_cache: bool = True to ModelConfig in llm_config.py
• Thread it through _get_source_transforms in export_llama_lib.py
• Add is_seq_at_dim_2 parameter to replace_kv_cache_with_custom_kv_cache
  and replace_sdpa_with_custom_op (defaulting to True for backward compat)

Also fixes:

• torchao aarch64:matmul BUCK: deps -> exported_deps for :macro, fixing
  transitive header visibility on arm64
• op_update_cache.cpp: %zd -> PRId64 for int64_t format strings

Authored with Claude.

Differential Revision: D99677679

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/18712

• 📄 Preview Python docs built from this PR

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

❌ 9 New Failures, 2 Cancelled Jobs

As of commit 6f31ee3 with merge base fb1618e ():

NEW FAILURES - The following jobs have failed:

• pull / test-lora-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 027585108d5aff25aab23065c56bc490431cb8376d49405f130e08c795d2fc4e /exec failed with exit code 1
• pull / test-lora-multimethod-linux / linux-job (gh)
  RuntimeError: Command docker exec -t 95d857e6d8ad9805ac679b4c85c6d3b3af90f1157700327cf51dfd2048c71c8d /exec failed with exit code 1
• pull / test-multimodal-linux (gemma3-4b) / linux-job (gh)
  RuntimeError: Command docker exec -t 1a3cc9695fca0e6bf8b57ecf0470f24f005b2243aefa734aa49f2f033553f747 /exec failed with exit code 134
• pull / unittest / linux / linux-job (gh)
  examples/models/llama/tests/test_replace_kv_cache.py::TestReplaceKVCache::test_replace_quantized_kv_cache_with_quantized_ring_kv_cache
• pull / unittest / macos / macos-job (gh)
  examples/models/llama/tests/test_replace_kv_cache.py::TestReplaceKVCache::test_replace_quantized_kv_cache_with_quantized_ring_kv_cache
• pull / unittest-buck / linux / linux-job (gh)
  RuntimeError: Command docker exec -t f573da8de17349530a07d3266eae88ac920390f0fae6cf61268575a588d9bc9b /exec failed with exit code 3
• pull / unittest-buck / macos / macos-job (gh)
  RuntimeError: Command bash /Users/ec2-user/runner/_work/_temp/exec_script failed with exit code 3
• pull / unittest-editable / linux / linux-job (gh)
  examples/models/llama/tests/test_replace_kv_cache.py::TestReplaceKVCache::test_replace_quantized_kv_cache_with_quantized_ring_kv_cache
• pull / unittest-editable / macos / macos-job (gh)
  examples/models/llama/tests/test_replace_kv_cache.py::TestReplaceKVCache::test_replace_quantized_kv_cache_with_quantized_ring_kv_cache

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

• 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.

[kimishpatel]

submitted by accident, not meant to land immedidately

[github-actions]

Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as Stale.
Feel free to remove the Stale label if you feel this was a mistake.
If you are unable to remove the Stale label please contact a maintainer in order to do so.
If you want the bot to never mark this PR stale again, add the no-stale label.
Stale pull requests will automatically be closed after 30 days of inactivity.