Optimize w8a8 pallas kernel #9473
Merged
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
vanbasten23
reviewed
Jul 11, 2025
vanbasten23
reviewed
Jul 11, 2025
torch_xla/experimental/pallas_kernels/quantized_matmul_kernel.py
Outdated
Show resolved
Hide resolved
vanbasten23
reviewed
Jul 11, 2025
vanbasten23
reviewed
Jul 11, 2025
vanbasten23
reviewed
Jul 11, 2025
torch_xla/experimental/pallas_kernels/quantized_matmul_kernel.py
Outdated
Show resolved
Hide resolved
vanbasten23
reviewed
Jul 11, 2025
vanbasten23
reviewed
Jul 11, 2025
c246fa6 to
9a965b0
Compare
vanbasten23
approved these changes
Jul 15, 2025
Collaborator
vanbasten23
left a comment
There was a problem hiding this comment.
Looks good. Thanks @kyuyeunk
Will merge once the TPU CI finishes.
Collaborator
|
The TPU CI seems blocking the merge. |
Contributor
Author
Do you know what is wrong & how to resolve it? |
Adds following optimizations - Minimize run-time branching logic by creating multiple functions during compile time - Use bf16 during input quantization - If possible, cache input quantization result for later use to reduce re-quantization - Only save accumulation output to scratch memory when necessary - Only create scratch memory when necessary
9a965b0 to
bdc7787
Compare
3 tasks
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Outdated suggestions cannot be applied.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.
This PR optimizes performance of quantized matmul kernel using following optimizations