-
Notifications
You must be signed in to change notification settings - Fork 2.9k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
TSAN race in //tests:export_back_compat_test_cpu under free-threading #26272
Comments
Another similar race report Click
To reproduce: bazel test --test_env=JAX_NUM_GENERATED_CASES=$JAX_NUM_GENERATED_CASES --test_env=JAX_ENABLE_X64=$JAX_ENABLE_X64 --test_env=JAX_SKIP_SLOW_TESTS=$JAX_SKIP_SLOW_TESTS --test_env=PYTHON_GIL=0 --test_env=TSAN_OPTIONS=halt_on_error=1,suppressions=$PWD/.github/workflows/tsan-suppressions.txt --test_env=JAX_TEST_NUM_THREADS=8 --test_output=all --local_test_jobs=32 --test_timeout=600 --nocache_test_results --runs_per_test=100 --test_filter=CompatTest //tests:export_back_compat_test_cpu |
A perhaps silly question: when we compile the MLIR C++ code, do we compile it with threading enabled? I seem to recall some locking there is only enabled if you build with threading enabled. Perhaps also we need to enable threading on our MLIR contexts? IIRC right now we disable it because we don't want MLIR to create thread pools, but that's orthogonal to "is thread safe". |
Description
This 313t free-threading CI run:
https://github.com/jax-ml/jax/actions/runs/13111654401/job/36576699494?pr=26261
contains this race in MLIR code:
PTAL?
System info (python version, jaxlib version, accelerator, etc.)
Python 3.13t
The text was updated successfully, but these errors were encountered: