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

WIP: keyword arguments for parallel_for and synchronize #186

Closed
wants to merge 1 commit into from

Conversation

PhilipFackler
Copy link
Collaborator

No description provided.

@PhilipFackler
Copy link
Collaborator Author

Test this please

function JACC.synchronize(::AMDGPUBackend; stream = default_stream())
AMDGPU.synchronize(stream)
end

function JACC.parallel_for(
Copy link
Collaborator

@williamfgc williamfgc Feb 10, 2025

Choose a reason for hiding this comment

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

Is this absorbing the generic API? Would it make sense to overload rather than extend? Since this is a specialized case and keyword arguments can add overhead to quick kernel launching. Would be good to test in apps like Lulesh with those characteristics. Thoughts?

Copy link
Collaborator Author

Choose a reason for hiding this comment

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

Not sure how this question applies to synchronize. But for parallel_for: yes I included these keywords in the current implementation. I could implement a separate API, but the keyword arguments would need to be handled differently (it would conflict with the vararg x...). This would be easy enough. I'll try it that way on a separate branch so we can compare.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants