Set correct threads_per_warp to 64 for AMD GPU#170
Open
yaomingamd wants to merge 9 commits into
Open
Conversation
TODO: need to find way setit to 32 for CUDA/NVIDIA GPU
TODO: need to set warp_size 32 for CUDA//NVIDIA GPU
…lly set warp_size =32 for AMD GPU it is a temporary solution
Collaborator
|
Hi @yaomingamd you can close this PR due to #177 |
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
3 participants
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
In xla0.5.0, some algorithms depends on warp_size 32, so that thereads_per_warp has been manually set to 32. This PR tries to reset threads_per_warp to 64 correctly for current AMD GPU and at the same, fix unittest failure/numerical issues due to this change. I have run unittest of jax, it shows that unittests have much less failure/numerical issues than original/base branch rocm-jaxlib-v0.5.0. For example, pytest tests/pallas pass all unittests.
=============== 1907 passed, 3307 skipped in 1281.18s (0:21:21) ================