Fix in pytorch do_bench_using_profiling #154766
Conversation
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/154766
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 1 Cancelled Job, 2 Unrelated FailuresAs of commit dfc3fb5 with merge base 0f3db20 ( CANCELLED JOB - The following job was cancelled. Please retry:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
BROKEN TRUNK - The following job failed but was present on the merge base:👉 Rebase onto the `viable/strict` branch to avoid these failures
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
|
|
This pull request was exported from Phabricator. Differential Revision: D75698662 |
This PR needs a
|
7e190b1 to
d027c08
Compare
|
This pull request was exported from Phabricator. Differential Revision: D75698662 |
d027c08 to
0d6dfa3
Compare
0d6dfa3 to
18bdbbc
Compare
|
This pull request was exported from Phabricator. Differential Revision: D75698662 |
18bdbbc to
4ffe9ad
Compare
|
This pull request was exported from Phabricator. Differential Revision: D75698662 |
4ffe9ad to
27decd3
Compare
Summary: Pull Request resolved: pytorch#154766 Local testing with benchmark demonstrated a decrease in the error message frequency: "Failed to divide all profiling events into #repeat groups. " "#CUDA events: %d, #repeats: %s" However, this is not a comprehensive solution and does not solve the problem every time. Test Plan: Please refer to the next diff. Differential Revision: D75698662
27decd3 to
3cbf635
Compare
Summary: Pull Request resolved: pytorch#154766 Local testing with benchmark demonstrated a decrease in the error message frequency: "Failed to divide all profiling events into #repeat groups. " "#CUDA events: %d, #repeats: %s" However, this is not a comprehensive solution and does not solve the problem every time. Test Plan: Please refer to the next diff. Differential Revision: D75698662
|
This pull request was exported from Phabricator. Differential Revision: D75698662 |
3cbf635 to
dfc3fb5
Compare
|
Looks like this PR hasn't been updated in a while so we're going to go ahead and mark this as |
|
@JonathanHuangg , can you describe more about the problem in a github issue? |
Summary:
Local testing with benchmark demonstrated a decrease in the error message frequency:
"Failed to divide all profiling events into #repeat groups. "
"#CUDA events: %d, #repeats: %s"
However, this is not a comprehensive solution and does not solve the problem every time.
Test Plan: Please refer to the next diff.
Differential Revision: D75698662
cc @voznesenskym @penguinwu @EikanWang @jgong5 @Guobing-Chen @XiaobingSuper @zhuhaozhe @blzheng @wenzhe-nrv @jiayisunx @ipiszy @chenyang78 @kadeng @muchulee8 @amjames @chauhang @aakhundov @coconutruben