rzo1 commented on PR #568: URL: https://github.com/apache/opennlp/pull/568#issuecomment-1855883632
> If I understood the issue and read the benchmark gist results correctly, with the new code JMH benchmark starts faster, then the old code catchs up and plateaus around 1.2K ops/s. But in overall the non interned version still process more ops/s, with the advantage of not having the JVM max memory issue. It is `,`, so we have (if you look at the mean values): - 1,211 ± 0,149 ops/s (old code) vs. - 468,404 ± 140,482 ops/s (new code) Systemarray.copy(...) is the reason for the speed improvement from merely 1 ops/s to 328 op/s (worst case) -- 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]
