Fix huge allocation overhead in UnstableRateCounter
#26976
Merged
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Outdated suggestions cannot be applied.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.
Overview
Currently
UnstableRateCounter
is allocationg a new array of size of judged objects on each judgement, which quickly becoming a problem for maps with many objects. By using Welford's online algorithm we can avoid selection and multiple passes over hitEvents and compute UR in a single loop.Benchmark
Code
Algorithm comparison
There is a
![diff](https://private-user-images.githubusercontent.com/22874522/301969875-d8b64f47-6628-4aeb-812c-d48df57dba6d.png?jwt=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.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.U8w7CPH6Bdnzs00sKXdVRvzHtXsVvk59xblH2B6y5l4)
UnstableRateTest
test scene but just in case I compared old and new algo manually, results are basically the same: