richardstartin commented on pull request #8397: URL: https://github.com/apache/pinot/pull/8397#issuecomment-1077589112
I've reverted to JODA because I noticed that `DateTime` objects appear to scalarise in loops like this, where java.time results in allocation. I only tested for `year` but the difference is big enough to stay on JODA: ``` Benchmark (_size) Mode Cnt Score Error Units BenchmarkTimeAPI.convertJavaTime 1024 avgt 5 35.190 ± 1.019 us/op BenchmarkTimeAPI.convertJavaTime:·gc.alloc.rate 1024 avgt 5 3842.276 ± 112.354 MB/sec BenchmarkTimeAPI.convertJavaTime:·gc.alloc.rate.norm 1024 avgt 5 212992.015 ± 0.002 B/op BenchmarkTimeAPI.convertJavaTime:·gc.churn.G1_Eden_Space 1024 avgt 5 3845.330 ± 353.367 MB/sec BenchmarkTimeAPI.convertJavaTime:·gc.churn.G1_Eden_Space.norm 1024 avgt 5 213156.517 ± 17414.389 B/op BenchmarkTimeAPI.convertJavaTime:·gc.churn.G1_Old_Gen 1024 avgt 5 0.003 ± 0.009 MB/sec BenchmarkTimeAPI.convertJavaTime:·gc.churn.G1_Old_Gen.norm 1024 avgt 5 0.190 ± 0.510 B/op BenchmarkTimeAPI.convertJavaTime:·gc.count 1024 avgt 5 91.000 counts BenchmarkTimeAPI.convertJavaTime:·gc.time 1024 avgt 5 54.000 ms BenchmarkTimeAPI.convertJodaTime 1024 avgt 5 6.541 ± 0.302 us/op BenchmarkTimeAPI.convertJodaTime:·gc.alloc.rate 1024 avgt 5 ≈ 10⁻⁴ MB/sec BenchmarkTimeAPI.convertJodaTime:·gc.alloc.rate.norm 1024 avgt 5 0.003 ± 0.001 B/op BenchmarkTimeAPI.convertJodaTime:·gc.count 1024 avgt 5 ≈ 0 counts ``` -- 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]
