-
Notifications
You must be signed in to change notification settings - Fork 26.1k
Closed
Labels
module: flaky-testsProblem is a flaky test in CIProblem is a flaky test in CImodule: macosMac OS related issuesMac OS related issuesmodule: mpsRelated to Apple Metal Performance Shaders frameworkRelated to Apple Metal Performance Shaders frameworkskippedDenotes a (flaky) test currently skipped in CI.Denotes a (flaky) test currently skipped in CI.triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module
Description
Platforms: mac, macos
This test was disabled because it is failing in CI. See recent examples and the most recent trunk workflow logs.
Over the past 3 hours, it has been determined flaky in 4 workflow(s) with 4 failures and 4 successes.
Debugging instructions (after clicking on the recent samples link):
DO NOT ASSUME THINGS ARE OKAY IF THE CI IS GREEN. We now shield flaky tests from developers so CI will thus be green but it will be harder to parse the logs.
To find relevant log snippets:
- Click on the workflow logs linked above
- Click on the Test step of the job so that it is expanded. Otherwise, the grepping will not work.
- Grep for
test_sdpa_mask_fp16_L6_S17_NH23_HS121 - There should be several instances run (as flaky tests are rerun in CI) from which you can study the logs.
Sample error message
Traceback (most recent call last):
File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/test_mps.py", line 9547, in test_sdpa_mask_fp16_L6_S17_NH23_HS121
self._test_sdpa_mask(torch.float16, 7, 17, 23, 121)
File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/test_mps.py", line 9535, in _test_sdpa_mask
self._compare_tensors(y.cpu(), y_ref)
File "/Users/ec2-user/runner/_work/pytorch/pytorch/test/test_mps.py", line 9459, in _compare_tensors
self.assertLess(err, 0.01)
File "/Users/ec2-user/runner/_work/_temp/conda_environment_11490952968/lib/python3.9/unittest/case.py", line 1223, in assertLess
self.fail(self._formatMessage(msg, standardMsg))
File "/Users/ec2-user/runner/_work/_temp/conda_environment_11490952968/lib/python3.9/unittest/case.py", line 676, in fail
raise self.failureException(msg)
AssertionError: nan not less than 0.01
To execute this test, run the following from the base repo dir:
python test/test_mps.py TestSDPA.test_sdpa_mask_fp16_L6_S17_NH23_HS121
This message can be suppressed by setting PYTORCH_PRINT_REPRO_ON_FAILURE=0
Test file path: test_mps.py
cc @clee2000 @malfet @albanD @kulinseth @DenisVieriu97 @jhavukainen
Metadata
Metadata
Assignees
Labels
module: flaky-testsProblem is a flaky test in CIProblem is a flaky test in CImodule: macosMac OS related issuesMac OS related issuesmodule: mpsRelated to Apple Metal Performance Shaders frameworkRelated to Apple Metal Performance Shaders frameworkskippedDenotes a (flaky) test currently skipped in CI.Denotes a (flaky) test currently skipped in CI.triagedThis issue has been looked at a team member, and triaged and prioritized into an appropriate moduleThis issue has been looked at a team member, and triaged and prioritized into an appropriate module