fix performance issue on AllIntersections #14
Merged
Conversation
This file contains 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
gilbsgilbs
force-pushed
the
fix-intersect-perf
branch
3 times, most recently
from
d053367 to
88a564a
Compare
|
Codecov ReportAll modified and coverable lines are covered by tests ✅
❗ Your organization needs to install the Codecov GitHub app to enable full functionality. Additional details and impacted files@@ Coverage Diff @@
## main #14 +/- ##
==========================================
- Coverage 95.52% 94.88% -0.64%
==========================================
Files 6 6
Lines 670 861 +191
==========================================
+ Hits 640 817 +177
- Misses 18 32 +14
Partials 12 12 ☔ View full report in Codecov by Sentry. |
gilbsgilbs
force-pushed
the
fix-intersect-perf
branch
from
88a564a to
37b87f8
Compare
This commit addresses a performance issue in the AllIntersections function, where the conditional for traversing the right side of the tree was evaluating to true most of the time. This led to unnecessary visits of right nodes, resulting in linear time complexity. More background on why this fix is correct: - Left is safe to prune when all lower nodes end before the start of the lookup interval (`Left.MaxEnd < start`). (Note: this case was already correct before this commit, but I reversed the inequality as I find it easier to reason about) - Right is safe to prune when all higher nodes start after the end of lookup interval (`end < Current.Start`). If the current node starts after the end of the lookup interval, we're guaranteed all its right nodes will also because they all start after the current node by definition. Fuzz testing yielded consistent results with original implementation. Benchmark showed ~30% speed improvement on average case (random intervals), and over 99.9% improvement in worst case (lookup interval at the far left of the tree). As expected, new implementation explores fewer nodes that can be evicted just as safely in advance.
gilbsgilbs
force-pushed
the
fix-intersect-perf
branch
from
37b87f8 to
22514f2
Compare
|
Hey @gilbsgilbs sorry for the late review. It's been a tough quarter at work. Your contribution makes total sense to me. Thank you! |
|
v1.4.1 published with the changes. Thank you again. |
|
No need to apologize. Thanks for the review and release 🙂. |
|
Just wanted to share some results after upgrading to v1.4.1. Great job!
|
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.
This commit addresses a performance issue in the AllIntersections function, where the conditional for traversing the right side of the tree was evaluating to true most of the time. This led to unnecessary visits of right nodes, resulting in linear time complexity.
More background on why this fix is correct:
Left.MaxEnd < start). (Note: this case was already correct before this commit, but I reversed the inequality as I find it easier to reason about)end < Current.Start). If the current node starts after the end of the lookup interval, we're guaranteed all its right nodes will also because they all start after the current node by definition.Fuzz testing yielded consistent results with original implementation. Benchmark showed ~30% speed improvement on average case (random intervals), and over 99.9% improvement in worst case (lookup interval at the far left of the tree). As expected, new implementation explores fewer nodes that can be evicted just as safely in advance.