use SnapTreeMap instead of ConcurrentSkipListMap #6719
Closed
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Outdated suggestions cannot be applied.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.
Using jprofiler, I found that the function in ConcurrentSkipListMap was cost more than 30% cpu in our case.
SnapTreeMap is a drop-in replacement for ConcurrentSkipListMap. Form the paper http://ppl.stanford.edu/papers/ppopp207-bronson.pdf:
Experimental evidence shows that our algorithm outperforms a highly tuned concurrent skip list for many access patterns, with an average of 39% higher singlethreaded throughput and 32% higher multi-threaded throughput over a range of contention levels and operation mixes.
After I changing to SnapTreeMap , it's 20%~30% faster than ConcurrentSkipListMap.