[ET-VK] Store weights transposed for int8 linear #9765
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
## Context The weight tensor of a linear layer is usually stored in a transposed manner, such that when computing the matrix multiplication, the reduction traverses along the rows of the weight tensor as opposed to the columns. This results in a better memory access pattern for CPUs. However, for GPUs, I have found that "un-transposing" the weight tensors result in better performance. This is likely due to the fact since GPUs can compute multiple output elements in parallel, reading along the columns allows for coalescing memory loads among threads in a work group. ## Changes * Introduce the ability to transpose height and weight dims when transferring tensor data to the GPU. * Prepackthe weight tensor "un-transposed" for the int8 quantized linear operator Differential Revision: [D72066588](https://our.internmc.facebook.com/intern/diff/D72066588/) [ghstack-poisoned]
🔗 Helpful Links🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/executorch/9765
Note: Links to docs will display an error until the docs builds have been completed. ❌ 1 New Failure, 1 Cancelled Job, 1 Unrelated FailureAs of commit 1f29600 with merge base 2aa7748 ( NEW FAILURE - The following job has failed:
CANCELLED JOB - The following job was cancelled. Please retry:
FLAKY - The following job failed but was likely due to flakiness present on trunk:
This comment was automatically generated by Dr. CI and updates every 15 minutes. |
facebook-github-bot
added
the
CLA Signed
|
This pull request was exported from Phabricator. Differential Revision: D72066588 |
This PR needs a
|
## Context The weight tensor of a linear layer is usually stored in a transposed manner, such that when computing the matrix multiplication, the reduction traverses along the rows of the weight tensor as opposed to the columns. This results in a better memory access pattern for CPUs. However, for GPUs, I have found that "un-transposing" the weight tensors result in better performance. This is likely due to the fact since GPUs can compute multiple output elements in parallel, reading along the columns allows for coalescing memory loads among threads in a work group. ## Changes * Introduce the ability to transpose height and weight dims when transferring tensor data to the GPU. * Prepackthe weight tensor "un-transposed" for the int8 quantized linear operator Differential Revision: [D72066588](https://our.internmc.facebook.com/intern/diff/D72066588/) [ghstack-poisoned]
|
This pull request was exported from Phabricator. Differential Revision: D72066588 |
trivedivivek
approved these changes
Apr 1, 2025
facebook-github-bot
merged commit 439d66d
into
gh/SS-JIA/204/base
79 of 84 checks passed
facebook-github-bot
temporarily deployed
to
cherry-pick-bot
GitHub Actions
Inactive
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.
Stack from ghstack (oldest at bottom):
Context
The weight tensor of a linear layer is usually stored in a transposed manner, such that when computing the matrix multiplication, the reduction traverses along the rows of the weight tensor as opposed to the columns. This results in a better memory access pattern for CPUs.
However, for GPUs, I have found that "un-transposing" the weight tensors result in better performance. This is likely due to the fact since GPUs can compute multiple output elements in parallel, reading along the columns allows for coalescing memory loads among threads in a work group.
Changes
Differential Revision: D72066588