Skip to content
This repository was archived by the owner on Mar 25, 2025. It is now read-only.

Avoid use of assert in the OpenACC kernel#720

Merged
pramodk merged 2 commits into
masterfrom
pramodk/psolve-direct
Aug 20, 2021
Merged

Avoid use of assert in the OpenACC kernel#720
pramodk merged 2 commits into
masterfrom
pramodk/psolve-direct

Conversation

@pramodk

@pramodk pramodk commented Aug 19, 2021

Copy link
Copy Markdown
Contributor

* Use of assert with atomic capture result
  into wrong results
* See details in BlueBrain/mod2c#68
@pramodk pramodk requested review from ohm314 and olupton August 19, 2021 15:15

@olupton olupton left a comment

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

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

Looks good to me, fixes the watchrange test locally...

@bbpbuildbot

Copy link
Copy Markdown
Collaborator

Logfiles from GitLab pipeline #14316 (:white_check_mark:) have been uploaded here!

Status and direct links:

@bbpbuildbot

Copy link
Copy Markdown
Collaborator

Logfiles from GitLab pipeline #14361 (:white_check_mark:) have been uploaded here!

Status and direct links:

@pramodk pramodk merged commit cf56a24 into master Aug 20, 2021
@pramodk pramodk deleted the pramodk/psolve-direct branch August 20, 2021 06:55
@olupton

olupton commented Aug 30, 2021

Copy link
Copy Markdown
Contributor

We reported a minimal reproducer of this problem (or of a similar problem, anyway) in https://forums.developer.nvidia.com/t/incorrect-gpu-results-with-pragma-acc-atomic-capture/187962.

JCGoran pushed a commit to neuronsimulator/nrn that referenced this pull request Mar 12, 2025
* Use of assert with atomic capture result into wrong results
* See details in BlueBrain/mod2c#68

NMODL Repo SHA: BlueBrain/nmodl@cf56a24
Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants