Skip to content
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.

Sign up for GitHub

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

Jump to bottom

Fix a bug in static_multimap: add early exit for empty input #151

Merged
merged 1 commit into from
May 4, 2022
Merged

Fix a bug in static_multimap: add early exit for empty input #151

merged 1 commit into from
May 4, 2022

Conversation

PointKernel
Copy link
Member

Related to rapidsai/cudf#10572

This PR adds an early exit for all bulk APIs in static_multimap. It can get rid of the cudaErrorInvalidConfiguration error since the current code will try to launch a gridDim.x == 0 kernel if the input is empty.

@PointKernel PointKernel added type: bug Something isn't working helps: rapids Helps or needed by RAPIDS labels May 2, 2022
vyasr
vyasr approved these changes May 4, 2022
View reviewed changes
Copy link
Collaborator

@vyasr vyasr left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

LGTM. Errors seem to be unrelated CI issues with sccache configuration.

@PointKernel
Copy link
Member Author

rerun tests

jrhemstad
jrhemstad approved these changes May 4, 2022
View reviewed changes
Copy link
Collaborator

@jrhemstad jrhemstad left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

One of the many things to add to my list for cuco refactoring is a kernel launch wrapper that will automate skipping kernel launches for 0 threads so we can avoid having to put these checks everywhere.

@jrhemstad jrhemstad merged commit 8b15f06 into NVIDIA:dev May 4, 2022
@PointKernel
Copy link
Member Author

One of the many things to add to my list for cuco refactoring is a kernel launch wrapper that will automate skipping kernel launches for 0 threads so we can avoid having to put these checks everywhere.

I will file an issue for this.

This was referenced May 4, 2022
Open
Merged
rapids-bot bot pushed a commit to rapidsai/cudf that referenced this pull request May 5, 2022
@PointKernel
Update cuco git tag (#10788)
d994905 
Closes #10572

This PR fetches the changes in NVIDIA/cuCollections#151 to avoid launching kernels with 0 threads.

Authors:
  - Yunsong Wang (https://github.com/PointKernel)

Approvers:
  - Mark Harris (https://github.com/harrism)

URL: #10788
@PointKernel PointKernel mentioned this pull request Jun 19, 2022
Open
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Reviewers

@vyasr vyasr vyasr approved these changes

@jrhemstad jrhemstad jrhemstad approved these changes

Assignees
No one assigned
Labels
helps: rapids Helps or needed by RAPIDS type: bug Something isn't working
Projects
None yet
Milestone
No milestone
Development

Successfully merging this pull request may close these issues.
3 participants
@PointKernel @vyasr @jrhemstad