Repository navigation
perf(leiden): bound memory of supernode aggregation - #80
Merged
Merged
Conversation
adsharma
self-requested a review
October 29, 2026 00:33
adsharma
reviewed
Oct 29, 2026
Supernodes with more than ICEBUG_VIEW_STAGE_CAP (default 2^16) staged entries are aggregated in a thread-local std::unordered_map instead of the edge-sized staging buffer. Neighbor vectors are now allocated at exact size to minimize memory usage.
- Wrap stage-cap comment to satisfy clang-format (100-col limit). - Hoist aggregationWork() out of aggregateNeighbors(): callers (computeNeighbors, ensureEagerCache) compute the O(members) estimate once and pass it through for the stage-cap branch and eager-cache eligibility, avoiding 2-3 passes over supernodeToOriginal[supernode]. Estimate is computed outside the cache-mutex since the mapping is read-only.
adsharma
force-pushed
the
pr1-view-memory
branch
from
October 29, 2026 02:20
c7e64ed to
693248b
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Supernodes with more than ICEBUG_VIEW_STAGE_CAP (default 2^16) staged entries are aggregated in a thread-local std::unordered_map instead of the edge-sized staging buffer, and neighbor vectors are now allocated at exact size to minimize memory usage.