POC GPU Assembly #6612
Draft
POC GPU Assembly #6612
+1,919
−486
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
multitalentloes
added
the
manual:new-feature
multitalentloes
force-pushed
the
start_porting_assembly_to_gpu
branch
from
1c3c202 to
1f483ec
Compare
Member
Author
multitalentloes
force-pushed
the
start_porting_assembly_to_gpu
branch
from
1f483ec to
cf6557d
Compare
currently only supports GPU, but groundwork is layed for handling OpenMP parallelization in the same code via the wrapper.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
NOT intended for merging.
This PR shows current a POC for having the matrix assemble happen on the GPU. Many shortcuts and simplifications have been made, and very little effort has been spent on writing high-quality code. Performance results will be shared later when all the required elements for a proper SPE11C simulation case has been implemented and some effort has been spent into optimizing it. Currently I have a separate executable that will always use GPU assembly.
I will create separate PRs for subsets of files and get it merged partially over time.