enable FP_Qaunt on xpu, validated#11
Merged
Godofnothing merged 1 commit intoIST-DASLab:masterfrom Oct 25, 2025
Merged
Conversation
Signed-off-by: Yao, Matrix <matrix.yao@intel.com>
Contributor
Author
|
@BlackSamorez @Godofnothing , pls help review, thx very much. |
Godofnothing
approved these changes
Oct 25, 2025
Member
Godofnothing
left a comment
Godofnothing
left a comment
There was a problem hiding this comment.
Thank you very much for your contribution.
I have reviewed the code at the changes to make sense.
Therefore, I am approving the PR.
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
2 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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Since 2.6, PyTorch introduce
torch.acceleratorto extend built-in device support beyond CUDA GPU. Other accelerators like Intel XPU already supported through this mechanism officially by PyTorch.In this PR, we want to extend FP_Quant support to Intel XPU, too. Hugging Face transformers FP-Quant psedo-quant(w/ triton backend) test cases are all passed on XPU. For QUBLAS backend, since current gen of Intel XPU doesn't support MIXFP4/NVFP4 in hardware yet, we will start to support once our next Gen XPU Crescent Island is out. But it will not impact this PR.
The code is back-compatible, when users install a earlier torch which doesn't support, it will fallback to directly use cuda.
Thx very much.