MaheshRavishankar wrote: If it passes tests, id say its OK to land. If the nvidia-bot fails again, we can chalk it down to flake. I dont know who owns the nvidia-bot but they will have to help triage the error https://github.com/llvm/llvm-project/pull/142827