This repository has been archived by the owner on Aug 16, 2024. It is now read-only.
L2-friendly chunking and twiddle persistence for batched NTTs and batched NTT+bitreverse sequences #31
Conversation
This file contains 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
mcarilli
force-pushed
the
mc-ntt-persistence
branch
from
a25d5f8 to
60dbec7
Compare
4 tasks
mcarilli
force-pushed
the
mc-ntt-persistence
branch
from
60dbec7 to
2bc2b81
Compare
Merged
4 tasks
mcarilli
force-pushed
the
mc-ntt-persistence
branch
from
a1daa63 to
f7e7c54
Compare
robik75
pushed a commit
to matter-labs/era-boojum-cuda
that referenced
this pull request
Feb 13, 2024
Required by matter-labs/era-shivini#31 ## Checklist - [x] PR title corresponds to the body of PR (we generate changelog entries from PRs). - [x] Tests for the changes have been added / updated. - [x] Documentation comments have been added / updated. - [x] Code has been formatted via `cargo fmt` and `cargo lint`.
mcarilli
changed the title
robik75
approved these changes
Feb 13, 2024
Sign up for free
to subscribe to this conversation on GitHub.
Already have an account?
Sign in.
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Outdated suggestions cannot be applied.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.
What ❔
Batched NTT (+bitrev) operations launch a sequence of several kernels. This PR splits batches into chunks small enough to persist in the L2 cache across kernel launches, ie, we do the whole NTT (+bitrev) sequence for the first chunk, then the second chunk, and so on.
Why ❔
Leveraging L2 persistence this way reduces gmem traffic and improves performance.
Checklist
zk fmtandzk lint.