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.

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

Destroy CUDA stream when finished #2921

Merged
merged 1 commit into from
Aug 14, 2015
Merged

Conversation

buaaliyi
Copy link
Contributor

Add missing cudaStreamDestroy in BasePrefetchingDataLayer::InternalThreadEntry. cudaStreamCreate and cudaStreamDestroy function call always come in pairs.

@ronghanghu ronghanghu added the bug label Aug 14, 2015
@ronghanghu
Copy link
Member

@cypof I am not an expert in CUDA myself, but I believe cudaStreamDestroy was missing in #2903 (originally #2114). Although it shouldn't really matter since prefetch thread is now persistent, it is still better to call cudaStreamDestroy.

I think this PR should be the right fix. Please correct me if I'm wrong.

@cypof
Copy link
Member

cypof commented Aug 14, 2015

Ah yes, that's correct. Good catch.

ronghanghu added a commit that referenced this pull request Aug 14, 2015
Destroy CUDA stream when finished
@ronghanghu ronghanghu merged commit 65c7fa6 into BVLC:master Aug 14, 2015
@buaaliyi buaaliyi deleted the multi_gpu branch August 15, 2015 04:15
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants