ciflow/trunk/197326: Map the 1-GPU runners to the 4xlarge-sized OSDC labels
- PyTorch: 1404 events in the last 90 days
- PyTorch: 1386th Release in the last 90 days
- Previous: earlier the same day · ciflow/trunk/198426
What happened
linux.g5.4xlarge.nvidia.gpu and linux.g6.4xlarge.experimental.nvidia.gpu were pointing at the 29 vCPU / 113Gi pods, which occupy a whole 8xlarge node. The 4xlarge-sized pods fit the instance these labels name. linux.g5.4xlarge.nvidia.gpu l-x86aavx2-29-113-a10g -> l-x86aavx2-11-41-a10g linux.g6.4xlarge.experimental.nvidia.gpu l-x86aavx2-29-113-l4 -> l-x86aavx2-11-41-l4 periodic.yml pins the label directly for aoti_cr…
Summary assembled by rule from the sources below