preserve tensor striding during compute estimation #37
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
facebook-github-bot
added
the
CLA Signed
ezyang
approved these changes
Jul 12, 2025
Contributor
|
lgtm. and pushed a lint fix |
wconstab
added a commit
that referenced
this pull request
Jul 18, 2025
Turns out the previous PR #37 was not correct. It divided the wrong dim's stride. This PR divides the dim to the left of the one being sharded, which is what really happens. Note: that we have this util at all is worrying me. Why don't we just use dtensors to propagate?
fmassa
pushed a commit
that referenced
this pull request
Jul 18, 2025
Turns out the previous PR #37 was not correct. It divided the wrong dim's stride. This PR divides the dim to the left of the one being sharded, which is what really happens. Note: that we have this util at all is worrying me. Why don't we just use dtensors to propagate?
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Outdated suggestions cannot be applied.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.
the compute estimation in autoparallel uses
torch.emptyto allocate tensors, which is wrong if the input tensor has a specific striding.This came up trying to run llama3 with float8 quantization, because:
(1) float8linears desugar into a call to
aten._scaled_mm(2)
aten._scaled_mmrequires its second input to be column-major, and its assert was failing (code: https://github.com/pytorch/pytorch/blob/main/torch/_meta_registrations.py#L6453)repro code: checkout this branch pytorch/torchtitan#1378, and run: