From 0928594e309c691baffc674409fb879106685446 Mon Sep 17 00:00:00 2001 From: gasoonjia Date: Fri, 9 Oct 2026 05:29:34 -0700 Subject: [PATCH] [rocm] Raise the unittest-rocm-gfx950 timeout to 150 minutes Since #23603 the QuantizedGemmFamily kernels compile and run on gfx950 instead of failing fast at the Triton compile, and unittest-rocm-gfx950 went from ~28 to ~80 minutes (job 113766563170: 10:02-11:22, pytest 67.6 min). About 37 of those minutes are Int4AutotunePickTest's CPU-saturated AOTI compile, which is the point of that test. At a 90-minute timeout one run was already cancelled. Raise the job's timeout to 150 minutes rather than drop or split suites. The other gfx950 jobs in this workflow use 180. --- .github/workflows/rocm.yml | 5 ++++- 1 file changed, 4 insertions(+), 1 deletion(-) diff --git a/.github/workflows/rocm.yml b/.github/workflows/rocm.yml index 154454eb145..263d9ff2bcb 100644 --- a/.github/workflows/rocm.yml +++ b/.github/workflows/rocm.yml @@ -175,7 +175,10 @@ jobs: id-token: write contents: read with: - timeout: 90 + # Since #23603 the quantized GEMM kernels compile and run here instead of + # failing fast; the job takes ~80 min, ~37 of them Int4AutotunePickTest's + # CPU-saturated AOTI compile. + timeout: 150 no-sudo: true runner: linux.rocm.gpu.gfx950.1 gpu-arch-type: rocm