Add Triton INT4 dense kernels with dequant prefill path for Qwen3.5 MoE

[digantdesai]

Stack from ghstack (oldest at bottom):

• Add structured stats reporting and GPU memory tracking to Qwen3.5 MoE runner #19190
• -> Add Triton INT4 dense kernels with dequant prefill path for Qwen3.5 MoE #19188
• Add W4A8 INT8 activation kernels for batched MoE prefill #19187

Add three new Triton kernels for dense W4A16 linear projections that
replace tinygemm's tiled INT4 format with simple [N, K//2] packed weights
(same format as MoE experts):

• int4_matmul: fused dequant+tl.dot GEMM for medium M (prefill crossover)
• int4_matvec: bandwidth-optimized vec-mat for M=1 decode
• dequant_w4_to_bf16: weight dequant for large-M prefill via Inductor mm

W4DequantLinear wraps these with dual decode/prefill dispatch:

• Decode (M=1): int4_matvec (73 tok/s, ~35% slower than tinygemm)
• Prefill (M>1): dequant+F.linear via Inductor (3400 tok/s at 3K tokens,
  +67% over tinygemm baseline)

Single 18GB weight blob (no duplication). Decode perf regression is a
known trade-off for uniform weight format — to be revisited with a
CUDA C++ matvec kernel.

Also adds INT8 dynamic-activation MoE tests and comprehensive correctness
tests (48 tests, all passing at rtol=0.01).

Co-authored-by: Claude noreply@anthropic.com

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/19188

• 📄 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:

• Rolling out OSDC (ARC) runners on pull & trunk workflows in PyTorch main

❌ 8 New Failures, 1 Cancelled Job, 1 Pending, 1 Unrelated Failure

As of commit 3e518f0 with merge base cb4e5ae ():

NEW FAILURES - The following jobs have failed:

• MLX / backend-tester (models) / test-mlx-backend-models (gh)
  RuntimeError: Command bash /Users/runner/work/_temp/exec_script failed with exit code 1
• MLX / test-mlx-qwen35-moe / test-mlx-qwen35-moe (gh)
  AttributeError: 'Namespace' object has no attribute 'moe_moe_activation_dtype'. Did you mean: 'moe_activation_dtype'?
• pull / android / run-emulator (gh)
  The process '/usr/bin/sh' failed with exit code 1
• pull / test-multimodal-linux (gemma3-4b) / linux-job (gh)
  RuntimeError: Command docker exec -t f441f8162ae9c3d1274790c828ac088625fc961b972e0920442034bb51fac453 /exec failed with exit code 139
• Test CUDA Windows Export and E2E / export-model-cuda-windows-artifact (facebook, dinov2-small-imagenet1k-1-layer, non-quantized) / linux-job (gh)
  RuntimeError: Command docker exec -t b09c5324e7129bcca35eba9380520cab29fb59678a09a7c1c2c585f51b413815 /exec failed with exit code 1
• Test CUDA Windows Export and E2E / export-model-cuda-windows-artifact (nvidia, parakeet-tdt, non-quantized) / linux-job (gh)
  RuntimeError: Command docker exec -t 0a0bd53522b7d897be5d4b2c35fe6be3173de09c4cfd833e7e374a317a4270da /exec failed with exit code 1
• Test CUDA Windows Export and E2E / export-model-cuda-windows-artifact (nvidia, parakeet-tdt, quantized-int4-weight-only) / linux-job (gh)
  RuntimeError: Command docker exec -t 91d8a417748365f3830dbda60cd383f810ba45796209f448823ae28eba31ccf8 /exec failed with exit code 1
• Test Metal Backend / test-metal-qwen35-moe-tiny / macos-job (gh)
  AttributeError: 'Namespace' object has no attribute 'moe_moe_activation_dtype'. Did you mean: 'moe_activation_dtype'?

CANCELLED JOB - The following job was cancelled. Please retry:

• Test CoreML Backend / test-coreml / test-backend-macos (coreml, operators) / macos-job (gh)
  ##[error]The operation was canceled.

BROKEN TRUNK - The following job failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

• MLX / test-mlx-voxtral / test-mlx-voxtral (gh) (trunk failure)

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.

[Gasoonjia]

Please update ci to cover the new changes

[Gasoonjia]

why moe_moe_moe_activation_dtype ? what does the triple MoEs mean?

[digantdesai]

Claude going rogue