Skip to content

Conversation

@alekstheod
Copy link
Collaborator

Test fix_jax_ut

alekstheod and others added 8 commits October 30, 2025 15:09
* Switch to use rbe workers

* Introduce rocm rbe pools

* First check for multigpu tag

* Fix buildifier issue

* Use valid platform name

* Fix docker url

* Run multigpu tests locally

* Build mutigpu tests locally

* Ignore numa related leaks

* Limit the supported archs

* Build locally run remotely

* Enable remote and disk_cache for jax tests

* Execute iota_test locally

* Mark failing tests as local

* Ignore iota_test as it is flaky and large

* Make dot operation local and exclude flaky test

* Bump to rocm7 container

* Disable failing test on tsan

* Switch rbe image

* Fixing hermetic build

* Force tsan builds to run locally

* Set proper multigpu tag for all_reduce_test

* Run flaky on rbe tests locally

* Make collective ops test local
* Assign gpu pools only to test actions

* Ignore flaky tests
@alekstheod alekstheod changed the title Ci fix jax build test failures [Do not merge just a test] Ci fix jax build test failures Nov 4, 2025
@alekstheod alekstheod force-pushed the ci_fix_jax_build_test_failures branch from 8fefdbd to 6ed8b63 Compare November 4, 2025 16:03
@alekstheod alekstheod force-pushed the ci_fix_jax_build_test_failures branch from ce6d8e8 to 65ede86 Compare November 4, 2025 17:17
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants