-
Notifications
You must be signed in to change notification settings - Fork 21.4k
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
[CUBLAS][CUDA GRAPHS] (re-re-re-re-open of #83461) Explicitly set the workspace for cuBLAS handles #86645
Conversation
Summary: This was causing crashes in the cuda tests inside Meta. Test Plan: Already verified this works inside Meta. Rely on OSS CI. Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
Summary: Avoid multiple traversals of the map with correct std::map usage. Test Plan: Rely on CI. Reviewers: Subscribers: Tasks: Tags: [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/86645
Note: Links to docs will display an error until the docs builds have been completed. ✅ No FailuresAs of commit 0e4ca01: This comment was automatically generated by Dr. CI and updates every 15 minutes. |
|
For the CLA, maybe try squashing all the commits in the eqy:cublasworkspace_graphs branch so that there is only a single commit authored by you. It looks like it is getting confused at 071ebae |
@xuzhao9 heads up I'm landing this PR, as my understanding is there are workarounds for torchbench. |
@pytorchbot merge -g |
Merge startedYour change will be merged once all checks on your PR pass since you used the green (-g) flag (ETA: 0-4 Hours). Learn more about merging in the wiki. Questions? Feedback? Please reach out to the PyTorch DevX Team |
Hey @eqy. |
@xuzhao9 the link for the corresponding If the memory leak check (pytorch/benchmark#1228) is added, additional changes like the ones in this PR would also be needed. |
re-opening (again) in hopes of working around failed/stuck CLA check
CC @ptrblck @ngimel @huydhn