Skip to content

tests: flex_attention NoValidChoicesError in test_pd_stream_admission CUDA cases on torch 2.10 #294

Closed

Description

@gspark-etri

Problem

With a GPU visible, three [cuda] cases in opt_prime/tests/test_pd_stream_admission.py fail on torch 2.10.0+cu128 with an inductor flex_attention NoValidChoicesError. CI (CPU only) never runs them.

Evidence

Found on optimus-rm 52ff463 on the 2x A100 development host (one A100 visible) while verifying #114 / PR #293. The failure is the same with and without TORCH_DEVICE_BACKEND_AUTOLOAD=0, so it is unrelated to the furiosa-torch autoload problem.

Scope

Find whether the flex_attention block mask / shapes used by the stream-admission path are outside what torch 2.10 inductor can compile on A100, and fix the call or fall back to eager / SDPA.

Done when

test_pd_stream_admission.py passes with a GPU visible on torch 2.10 (and 2.13), or the CUDA cases skip with a stated reason.

Owner

gspark-etri

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions