dcapwell commented on PR #17: URL: https://github.com/apache/cassandra-accord/pull/17#issuecomment-1360658434
In my feedback branch I pushed JMH benchmarks for result.map vs chain.map, I re-structured the output to be easier to read (group by chainSize rather than benchmark). ``` Benchmark (chainSize) Mode Cnt Score Error Units AsyncChainBenchmark.chain 1 avgt 19.267 ns/op AsyncChainBenchmark.result 1 avgt 13.028 ns/op AsyncChainBenchmark.chain 2 avgt 24.885 ns/op AsyncChainBenchmark.result 2 avgt 24.677 ns/op AsyncChainBenchmark.chain 3 avgt 26.092 ns/op AsyncChainBenchmark.result 3 avgt 36.276 ns/op AsyncChainBenchmark.chain 4 avgt 30.516 ns/op AsyncChainBenchmark.result 4 avgt 47.993 ns/op AsyncChainBenchmark.chain 5 avgt 36.291 ns/op AsyncChainBenchmark.result 5 avgt 60.386 ns/op ``` `chain` starts to win after 3 hops I did the same for `flatMap`, but made the assumption that the source of async will produce a `AsyncResult` (which is true in the Cassandra branch) ``` Benchmark (chainSize) Mode Cnt Score Error Units AsyncChainBenchmark.chainFlatMap 1 avgt 33.366 ns/op AsyncChainBenchmark.resultFlatMap 1 avgt 25.701 ns/op AsyncChainBenchmark.chainFlatMap 2 avgt 56.366 ns/op AsyncChainBenchmark.resultFlatMap 2 avgt 51.788 ns/op AsyncChainBenchmark.chainFlatMap 3 avgt 70.098 ns/op AsyncChainBenchmark.resultFlatMap 3 avgt 78.100 ns/op AsyncChainBenchmark.chainFlatMap 4 avgt 96.927 ns/op AsyncChainBenchmark.resultFlatMap 4 avgt 104.856 ns/op AsyncChainBenchmark.chainFlatMap 5 avgt 111.148 ns/op AsyncChainBenchmark.resultFlatMap 5 avgt 133.539 ns/op ``` Given the assumption above, the fact that both models are the same makes sense, as you can not build a chain with `flatMap` (you are producing results with a single hop) -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
