Rewrite needle_map.CompactMap() for more efficient memory usage
#6813
Conversation
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
Has no measurable performance impact, but greatly increases memory efficiency when out-of-order needle IDs are written. TBD what's the impact on memory fragmentation, though it will be very platform dependent.
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Outdated suggestions cannot be applied.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.
What problem are we solving?
#6804
How are we solving the problem?
CompactMap()relies on fixed-size storage buckets of 10,000 elements, and a separate storage for overflow (i.e. out of order) needle IDs. The problem is that most of that fixed storage bucket space ends up allocated but never used; this is particularly true for out-of-order needle IDs, but happens also with perfectly ordered needle writes.This MR reworks
CompactMap()so:copy()instead of ad-hoc loops.ifandforsections are unrolled, whenever possible.The result is a dramatic improvement in memory usage, at the cost of slightly increased memory fragmentation - whose impact will be entirely platform-dependent.
#6804 has more details, but this MR improves memory usage of
weedprocesses by ~95% in best-case scenarios, and ~99% in worst-case scenarios, with no measurable performance impact.How is the PR tested?
Every single existing unit and integration test should pass without issues; the existing API and behavior for
CompactMap()is unchanged.Checks