Improve SeqDecoder performance by removing overhead #1719
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Outdated suggestions cannot be applied.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.
Hi, I was fiddling around to see if there was some way of making Circe faster and found that the current SeqDecoder has a non-trivial amount of overhead.
I believe this was mostly due to cursor object creation and destruction. But I can't be sure as those don't appear in the profiling flame graph that I used.
To compare, I ran
sbt 'benchmark/jmh:run -f1 io.circe.benchmark.NumberParsingBenchmark.decodeMany*'
before and after the optimization and there is a trend towards making arrays of primitives faster:Just to confirm, I made a JMH test in the style of the previous benchmark class with booleans:
and the effects are quite dramatic: