trunk/4581144c216b65e4e1b5c6eea26d4212239a231d: [ROCm] Re-enable skipped amdsmi tests (#200203)
Previously these tests were skipped because they fail on DPX.
Executive Summary
Previously these tests were skipped because they fail on DPX. It is confirmed that these functions of amdsmi is only available on the root partition, therefore unskipping them on the root partition. Example: Partition 0: https://github.com/pytorch/pytorch/actions/runs/37704922037/job/113089403200 TestCudaAllocator::test_temperature and TestCudaAllocator::test_clock_speed PASSED Partition 1: https://github.com/pytorch/pytorch/actions/runs/37704922037/job/113089403040 TestCudaAllocator::test_temperature and TestCudaAllocator::test_clock_speed SKIPPED with message "no full-GPU amdsmi metrics" Coauthored with Claude Pull Request resolved: #200203 Approved by:...
Hyperscale Open Source & Cloud Infrastructure Review
From an infrastructure engineering, cloud-native scale, and hyperscale governance standpoint: - **Hyperscale Provenance:** Open-sourcing internal frameworks subjects proprietary systems to rigorous public analysis. - **De-facto Standard Cohesion:** Publishing enterprise tooling establishes interoperable specifications across distributed execution. - **Vendor Decoupling:** Platform engineers can inspect underlying telemetry and memory layouts without black-box vendor lock-in.
Impact on the Open Ecosystem
Bridges enterprise engineering scale with independent, reproducible open-source software stacks.