-
Notifications
You must be signed in to change notification settings - Fork 16
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
[Flang][OpenMP] Add support for lastprivate clause lowering. #1593
base: fir-dev
Are you sure you want to change the base?
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Thanks @arnamoy10 for working on lastprivate.
Could you add an explanation for the approach that you have taken? Particularly, how is it ensured that it is the last iteration's value that is copied back to the original variable?
Thanks for the comment @kiranchandramohan . From this C++ OpenMP example, looking at the diff between with and without the use of It will be great if you can point to the part of the IR that ensures it. |
@arnamoy10 Can you read the discussion in https://discourse.llvm.org/t/rfc-privatisation-in-openmp-dialect/3526? I think the lastprivate handling is not done in the body of the loop in the llvm it generated but it is in the footer or exit portion of the loop. We do not have a representation for the footer/exit portion of the loop in I can explain more if required. |
Q: is there a related phabricator review to this PR? Thanks for working on this. |
@schweitzpgi Thanks for asking, unfortunately, privatization implementation, for now, has to be done in |
@schweitzpgi Is a phabricator review what you would prefer for OpenMP work? |
See Steve's announcement here. https://discourse.llvm.org/t/nvidia-transition-from-fir-dev/61947 I think transitioning to working upstream, in so far as it is possible to do so, would be a good direction. |
No description provided.