FP4 grouped refactor #4934
Closed
FP4 grouped refactor #4934
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
Summary: X-link: facebookresearch/FBGEMM#1956 as title Differential Revision: D83083612
) Summary: X-link: facebookresearch/FBGEMM#1955 We would reuse this kernel as a base to add support for NV/MX FP4. As a first step, shuffle it into it's own file. Differential Revision: D83151150
Summary: X-link: facebookresearch/FBGEMM#1957 Split some clean-up/refactors from the core FP4 Torch API support to make the next diff more focused. - Removed `zero_start_index_M` as it's unused - Removed passing `G` into the kernel directly as it can be inferred - Rename `ElementComputeEpilogue` -> `ElementScale` - Add `namespace fbgemm_gpu` in `f4f4bf16_grouped_common.cuh` - Removed `num_x_scale_per_group` and `num_w_scale_per_group` as they are both unused - Removed un-neccesary cutlass headers in `f4f4bf16_grouped.cu` Differential Revision: D83166227
✅ Deploy Preview for pytorch-fbgemm-docs ready!
To edit notification comments on pull requests, go to your Netlify project configuration. |
|
This pull request has been merged in d9b6e01. |
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Outdated suggestions cannot be applied.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.
Summary:
X-link: https://github.com/facebookresearch/FBGEMM/pull/1957
Split some clean-up/refactors from the core FP4 Torch API support to make the next diff more focused.
zero_start_index_Mas it's unusedGinto the kernel directly as it can be inferredElementComputeEpilogue->ElementScalenamespace fbgemm_gpuinf4f4bf16_grouped_common.cuhnum_x_scale_per_groupandnum_w_scale_per_groupas they are both unusedf4f4bf16_grouped.cuDifferential Revision: D83166227