[Gasoonjia]

test-muse-glimmer-cuda-e2e-kquant-17gb-dflash-image-batching has been cancelled at the 240-minute job timeout on every run since it was added in #24066 (for example job 130041481693: the export started at 19:12 and was still compiling at 22:50).

Why

The export is CPU-bound. All measurements below are on an A100 devserver with cold Inductor and Triton caches:

	44 CPUs	11-CPU cgroup quota, 125 GiB (CI runner)
forward_1	5m47s	6m24s
forward_8	6m16s	7m00s
forward_16	7m10s	9m26s
forward_32	8m22s	11m45s
whole export	2h27m	about 2h50m

Memory and swap stay flat under the quota. The quantized-GEMM compile pool is sized by os.cpu_count(), which inside a container reports the whole host, so it starts 32 compile workers on 11 vCPU. CI was slower still (about 2h53m to reach forward_others, against 1h26m on the devserver).

Change

• Size the quantized-GEMM compile pool by the CPUs the process may use: the affinity mask and the cgroup v2 cpu.max quota. On the CI runner that is 11 workers instead of 32. Hosts without a quota are unchanged (capped at 32).
• Export only forward_{1,4,16,64} for the CI e2e, instead of all seven static sizes. The CUDA executor picks from whatever forward methods the program has and pads each step to the next exported width (test_step_plan.cpp already covers sparse sets like {1,4}), so the runtime path under test is the same.

Test plan

• The pool sizing returns 11 for a cpu.max of 1100000 100000, and 32 for max or no cgroup file.
• CI: this PR touches .ci/scripts/export_model_artifact.sh, so the job should run here; it needs to finish within 240 minutes.

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/25018

• 📄 Preview Python docs built from this PR

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

❌ 2 New Failures, 4 Unrelated Failures, 3 Unclassified Failures

As of commit 814619c with merge base d01021b ():

NEW FAILURES - The following jobs have failed:

• trunk / test-models-linux-aarch64 (mv2, xnnpack-quantization-delegation, mt-l-arm64g4-16-62) / linux-job (gh)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 1
• trunk / unittest-release / linux / linux-job (gh)
  devtools/inspector/tests/inspector_test.py::TestInspector::test_transformer_block_xnnpack_numeric_gap_within_tolerance

UNCLASSIFIED FAILURES - DrCI could not classify the following jobs because the workflow did not run on the merge base. The failures may be pre-existing on trunk or introduced by this PR:

• Test CUDA Builds / test-model-cuda-e2e-Qwen3-0.6B-non-quantized / linux-job (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  ##[error]fatal: Fetched in submodule path 'backends/mlx/third-party/mlx', but it did not contain 1f8e74e3f12f31365464a6867c6579f0e9b29d85. Direct fetching of that commit failed.
• Test CUDA Builds / test-muse-glimmer-cuda-e2e-kquant-17gb-dflash-image-batching / linux-job (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  ##[error]received SIGINT
• Test ROCm AOTI / test-voxtral-realtime-rocm-gfx950-rocm7.2 / linux-job (gh) (this job did not run on the merge base, so DrCI cannot tell whether the failure is pre-existing)
  ##[error]The operation was canceled.

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

• trunk / Arm suite / test-arm-backend-ethos-u (test_pytest_models_ethos_u85) / linux-job (gh) (similar failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 2
• trunk / Arm suite / test-arm-backend-ethos-u (test_pytest_ops_ethos_u55) / linux-job (gh) (similar failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 2

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

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

• trunk / Arm suite / test-arm-backend-ethos-u (test_pytest_models_ethos_u55) / linux-job (gh) (trunk failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 2
• trunk / Arm suite / test-arm-backend-ethos-u (test_pytest_ops_ethos_u85) / linux-job (gh) (trunk failure)
  RuntimeError: Command bash /__w/_temp/exec_script failed with exit code 2

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]

Closing in favor of #24889. Measurement from this PR: with the cgroup-aware compile pool and --forward-sizes 1,4,16,64, the kquant-17gb dflash-image-batching export started 22 minutes into the job and was still compiling the draft methods when the 240-minute timeout hit, so the 480-minute timeout in #24889 is still needed. The forward-sizes change is now commit 8731344 on #24889.