fix: raise container nofile ulimit to prevent FD exhaustion in AMD multi-node jobs (#2140)

hongxiayang · web-flow · commit b2c32a1a64fb · 2026-07-09T15:21:51.000-07:00
Signed-off-by: Hongxia Yang &lt;hongxia.yang@amd.com&gt;
diff --git a/benchmarks/multi_node/amd_utils/job.slurm b/benchmarks/multi_node/amd_utils/job.slurm
@@ -573,6 +573,7 @@ fi
     --device=/dev/infiniband/uverbs7 \
     --ulimit memlock=-1 \
     --ulimit stack=67108864 \
+    --ulimit nofile=1048576:1048576 \
     --network host \
     --ipc host \
     --group-add video \
