CUDA fix #103
Open
CUDA fix #103
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
|
Added another CUDA function to CUDA::Kernel which can be used to control GPU occupancy of a kernel. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
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Outdated suggestions cannot be applied.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.
Just a small fix I forgot to add a while ago. In OpenCL the global work size is the number of all work items, but in CUDA the grid size is the number of all thread blocks (workgroups in OpenCL). When I added the CUDA example (which was adapted from the OpenCL example) I forgot to make this change. Doesn't cause any errors but CUDA just launches a bunch of extra threads that immediately exit. So I'm fixing this for future ACE users.
Funny story, for the longest time the KINC CUDA kernel was always just slightly slower than the OpenCL kernel... then I finally fixed this in the KINC code and then the CUDA version was slightly faster... as it should be.