DArray/stencil: Allocate HaloArray outside spawn_datadeps#703
Open
jpsamaroo wants to merge 7 commits into
Open
DArray/stencil: Allocate HaloArray outside spawn_datadeps#703jpsamaroo wants to merge 7 commits into
jpsamaroo wants to merge 7 commits into
Conversation
jpsamaroo
force-pushed
the
jps/stencil-outer-alloc
branch
from
1965160 to
7a1416b
Compare
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
This PR changes the stencil code to allocate
HaloArrayobjects outside of thespawn_datadepsregion, allowing Datadeps to properly parallelize stencil tasks accessing them (no more serial dependency because ofUnknownAliasing). This appears to improve performance by about 3x on a 512x512 benchmark with 8 threads, which is pretty nice! I expect that improvement to scale as we increase threads as well.This PR also caches and reuses allocated
HaloArrayobjects on the same task (caching byDArrayand per-chunk-index), netting another 2x performance benefit.This PR also includes changes to reduce task spawns and memory allocations, hopefully improving ROCm CI (which invokes the OOM killer).
Written by Claude Sonnet and Cursor Composer