add support for setting sm count in nvcuda backend#384
Merged
Conversation
Collaborator
Author
|
This currently creates the green context after compiling the kernel. This is not necessary, and I would like to change this to right before benchmarking to be in line with how we handle special nvml parameters. |
|
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Very experimental, but the idea is to be able to tune the active SM percentage in combination with other parameters. This can be controlled in CUDA using "Green Contexts". It is relatively involved to setup, so I have only implemented it in the cuda-python backend as of now.
The idea is that this parameter is interesting for energy efficiency tuning. In particular bandwidth bound kernels may disable some SMs to reduce energy use of cores that are idling most of the time anyway.
We could extend it to other CUDA backends as well probably.