Most of E2E models failed with torch._inductor.exc.InductorError: RuntimeError: Triton Error [ZE]: 0x78000011 on BMG windows.

[libohao1201]

🐛 Describe the bug

python benchmarks\dynamo\huggingface.py --accuracy -d xpu -n10 --backend=inductor --cold-start-latency --inference --float32                   --output=C:\logs\inductor-huggingface-accuracy-inference-float32.csv  

Versions

Machine	BMG - Win11
stock pytorch	PyTorch branch: https://github.com/etaf/pytorch-inductor-xpu/tree/xpu_inductor_windows (branch: xpu_inductor_windows)
triton	Commit: intel/intel-xpu-backend-for-triton@4d7fc60
Driver	internal driver: 32.0.101.6559
Conda	python 3.10

[Stonepia]

Link for reproducer:
intel/intel-xpu-backend-for-triton#3481

[Stonepia]

It should be a bug of the driver. Triton team is working on that. We will return to this issue once driver update.