Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Resize scheduler: enable vectorization #3694

Draft
wants to merge 4 commits into
base: resize_scheduler_reorder
Choose a base branch
from

Conversation

naoyam
Copy link
Collaborator

@naoyam naoyam commented Jan 10, 2025

Stacked on #3693

wujingyue and others added 2 commits January 9, 2025 22:00
This is for #3091 and #3092, limitations of mpi4py that are hard to get
rid of. This PR adds the bare minimum to get the sizes and the ranks.
The next PRs will expose more methods in Communicator.
@naoyam naoyam added the rope label Jan 10, 2025
naoyam and others added 2 commits January 10, 2025 10:56
This PR adds a scheduling primitive,
`cancelReshapeInLoopDomains(TensorView* from_tv)`, where all reshape
transforms appearing between `from_tv` and fusion outputs are
effectively cancelled in their loop domains. Please see the
[comment](https://github.com/NVIDIA/Fuser/pull/3679/files#diff-dc44235151285593f374bf60312da86dddebe6aed272e619001c088db507b783R72)
for a motivating example.

This could be used to remove the restriction of the interfering reshape
in reduction/normalization fusions.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants