[ExecuTorch][llm] Fuse w1+w3 into single GEMM in quantized_moe_ffn

[digantdesai]

Stack from ghstack (oldest at bottom):

• [ExecuTorch][llm] Wire --use_moe_quantized_op through export pipeline and LLM config #21126
• [ExecuTorch][llm] Add MoE quantized custom op documentation #21125
• -> [ExecuTorch][llm] Fuse w1+w3 into single GEMM in quantized_moe_ffn #21124
• [ExecuTorch][llm] Add numerical kernel tests for quantized_moe_ffn #21123
• [ExecuTorch][llm] Add source-transform and Meta-kernel tests for MoE op #21122
• [ExecuTorch][llm] Add MoE source transformation with INT4 quantized packing #21121
• [ExecuTorch][llm] Add Meta kernel and registration probe for quantized_moe_ffn #21120
• [ExecuTorch][llm] Add quantized_moe_ffn custom op with C++ kernel #21119

Fuse the up-projection (w1) and gate-projection (w3) into a single [2F, D] GEMM per expert. This halves the number of torchao activation quantizations per expert (from 2 to 1) and reduces total GEMM calls from 3 to 2 per active expert.

At AOT time, w1 and w3 are concatenated before packing: pack_fn(cat([w1, w3], dim=0)). At runtime, a single expert_linear_dispatch produces [m_e, 2F], then a fused swiglu_and_compact pass reads the interleaved h1/h3 and writes [m_e, F] for the w2 down-projection.

Schema changes from (packed_w1, packed_w3, packed_w2) to (packed_w13, packed_w2) — one fewer tensor arg (14 -> 13).

Differential Revision: D102799854

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/21124

• 📄 Preview Python docs built from this PR

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• Unrelated MacOS MPS test jobs time out (30 min) due to stale cpp_extension lock file poisoning self-hosted runners

❌ 1 New Failure, 10 Pending, 2 Unrelated Failures

As of commit a4989f1 with merge base 21554e5 ():

NEW FAILURE - The following job has failed:

• pull / unittest-editable / macos / macos-job (gh)
  extension/llm/custom_ops/test_quantized_moe.py::TestQuantizedMoeFfnOp::test_sigmoid_with_bias_route_scale_2p5

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

• pull / test-arm-backend-no-driver (test_pytest_models_tosa) / linux-job (gh) (matched linux rule in flaky-rules.json)
  The runner has received a shutdown signal. This can happen when the runner service is stopped, or a manually started runner is canceled.
• pull / unittest / macos / macos-job (gh) (detected as infra flaky with no log or failing log classifier)

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.