-
Notifications
You must be signed in to change notification settings - Fork 140
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
Multideterminant Batch code [Part 3 of Offload] #3892
Conversation
Test this please |
… into MSD-Offload_Part3
Test this please |
All requests from @ye-luo addressed. Once this is merged, It will be the baseline for comparing performance. |
Test this please |
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.
LGTM need non-ANL review/approval
Test this please |
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.
Very very minor request: I understand that the ///PRAGMA OFFLOAD comments are for part 4, but please tidy:
///This function needs really to be ported...
and any similar commentary in the touched source code. These have a habit of remaining for a long time and lose context.
@prckent : Comments are removed |
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 Anouar.
Test this please |
Please review the developer documentation
on the wiki of this project that contains help and requirements.
Proposed changes
All variables in Batched Multideterminant implementation have a dual view, host and device.
Multiple functions related to Phi->evaluateXXX do not have a batched or GPU implementations and required to send only the host view. Following those calls require either to send the result to the device using the
XXX.updateTo()
function or later, for those values to be computed on the GPU directly.This PR belongs to the general effort aiming at generating a GPU implementation of the multideterminant code. #3830
This is part 3: All Variables needed by Table method have dual View Host/Device.
What type(s) of changes does this code introduce?
Delete the items that do not apply
Refactoring (no functional changes, no api changes)
Other (please describe): GPU capability
Does this introduce a breaking change?
What systems has this change been tested on?
Laptop
Checklist
Update the following with a yes where the items apply. If you're unsure about any of them, don't hesitate to ask. This is
simply a reminder of what we are going to look for before merging your code.