Fix diffusion dask OOM: pass scalar diffusivity directly to chunks#1117
Merged
Conversation
Parallel subagent triage + ralph-loop workflow for auditing all xrspatial modules for performance bottlenecks, OOM risk under 30TB dask workloads, and backend-specific anti-patterns.
7 tasks covering command scaffold, module scoring, parallel subagent dispatch, report merging, ralph-loop generation, and smoke tests.
For scalar diffusivity, the dask chunk function now receives the float value directly instead of a full-raster numpy array captured in every task closure. This eliminates the O(H*W) eager allocation and the per-task serialization overhead. For DataArray diffusivity, the dask path passes the dask array as a second argument to map_overlap so each chunk gets only its own slice.
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
map_overlapso each chunk gets only its own slicenp.full(agg.shape, ...)allocation (line 261) anddiffusivity.valuesmaterialization (line 255) that caused OOM on large dask inputsContext
Found during performance sweep triage (#1116). On 512x512: dask used 35 MB vs numpy's 27 MB, with the 8 MB gap from the full-raster alpha_arr. At 30TB this allocation alone would OOM before any diffusion work starts.
Test plan
test_dask_matches_numpyconfirms dask scalar path matches numpy output