liyafan82 commented on issue #8397: [FLINK-11421][Table SQL/Runtime]Add compilation options to allow comp… URL: https://github.com/apache/flink/pull/8397#issuecomment-493873602 > @liyafan82 Thanks your local benchmark. > I think it is important to try to reproduce the results in a local machine. As you say, cluster environment is very complicate. In that case, the test data may not tell where optimizations exist and where regression occurs. > The compiler optimizer is easy to reproduce on a single machine. Everything is mockable. > > Can you add some other patterns? Like vector computation benchmark? Hi @JingsongLi , I agree with you that the local performance benchmarks can prevent performance regressions. However, performance benchmark can be difficult for Java. Since local and job environments differ greatly, JIT may apply different optimizations in different times. To reproduce the results in job environment, the local environment must apply exactly the same optimizations enabled in the job environment, and the optimizations must be applied in the same timing. Benchmark for vector computation can be a good suggestion. However, vector related code has not been merged into Flink, so how can we introduce the benchmarks based on Flink?
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
