Skip to content

Parallelization transformation#1010

Draft
majosm wants to merge 5 commits into
inducer:mainfrom
majosm:parallelize
Draft

Parallelization transformation#1010
majosm wants to merge 5 commits into
inducer:mainfrom
majosm:parallelize

Conversation

@majosm
Copy link
Copy Markdown
Contributor

@majosm majosm commented May 21, 2026

Adds some transformations for parallelization. Based on @kaushikcfd's work in inducer/arraycontext#216. This version generalizes that code to work with fused loops.

The basic procedure is:

  1. Collect instructions that share inames into sets.
  2. Find the "outer" inames (i.e., inames shared between all instructions) in each of those sets.
  3. Parallelize over those.

cc @lukeolson

TODO:

  • ⚠️ Re-run profiling on Tuolumne after porting over to loopy to make sure I haven't messed anything up
  • ⚠️ Add docs

@majosm
Copy link
Copy Markdown
Contributor Author

majosm commented May 22, 2026

Never mind, I'm going to add this in arraycontext instead (inducer/arraycontext#358). Leaving this open for the moment because I still want to look into the test failures in CI.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant