viable/strict/1790677893: [xpu] Correctly filter visiable device to align SYCL (#198336)
- PyTorch: 1641 events in the last 90 days
- PyTorch: 1622th Release in the last 90 days
- Previous: earlier the same day · trunk/7d96b234ccaec3ee31deffa704c2f488cb571010: [Dynamo] Support builtin fns on lazy constants (#191145)
What happened
Motivation In some platforms, iGPU is before dGPU. So we manually sort iGPU to the end to manually align to SYCL. Additional Context Fix intel/torch-xpu-ops#5015 Pull Request resolved: #198336 Approved by: https://github.com/gujinghui
Summary assembled by rule from the sources below