Enable CUTLASS grouped GEMM for pretraining wgrad on GB200 and H100 (resubmit) · pytorch/FBGEMM@87dd0dd

Triggered via pull request September 22, 2025 21:37

opened #4913

Status Failure

Total duration 10m 3s

Artifacts 1

fbgemm_gpu_benchmark_cpu.yml

on: pull_request

Matrix: build_artifact

Matrix: benchmark_artifact

2 errors

The job was not acquired by Runner of type self-hosted even after multiple attempts

Internal server error. Correlation ID: 25c0162d-a887-4d27-85ee-6806c3cdcb5f

Produced during runtime

Name	Size	Digest
fbgemm_gpu_nightly_cpu_arm_gcc_py3.13.whl	4.2 MB	`sha256:5b09056a48b9aa033079114e26598b8f52f0f3bdae4ed7b75eb5190736da038b`