[GitHub] [flink-benchmarks] dawidwys edited a comment on pull request #5: [FLINK-19884] Add benchmarks for batch-style execution for bounded keyed streams

[GitBox <gi...@apache.org>]

dawidwys edited a comment on pull request #5:
URL: https://github.com/apache/flink-benchmarks/pull/5#issuecomment-720338780


   First of all sorry for a poor quality of the initial contribution. It has way too many problems.
   
   I tried addressing all your comments. Here are the results from an execution on my local machine:
   
   ```
   Result "org.apache.flink.benchmark.SortingBoundedInputBenchmarks.sortedMultiInput":
     1397.457 ±(99.9%) 41.128 ops/ms [Average]
     (min, avg, max) = (1217.022, 1397.457, 1485.721), stdev = 61.558
     CI (99.9%): [1356.330, 1438.585] (assumes normal distribution)
   
   Result "org.apache.flink.benchmark.SortingBoundedInputBenchmarks.sortedOneInput":
     1821.521 ±(99.9%) 29.967 ops/ms [Average]
     (min, avg, max) = (1717.496, 1821.521, 1903.986), stdev = 44.853
     CI (99.9%): [1791.554, 1851.488] (assumes normal distribution)
   
   Result "org.apache.flink.benchmark.SortingBoundedInputBenchmarks.sortedTwoInput":
     1444.770 ±(99.9%) 35.563 ops/ms [Average]
     (min, avg, max) = (1348.234, 1444.770, 1530.831), stdev = 53.230
     CI (99.9%): [1409.206, 1480.333] (assumes normal distribution)
   ```
   
   I also did run a cpu profiler and I do see that the methods that I am interested in benchmarking take the most of the time.
   Some of the hot methods are: (SortingDataInput#emitNext, MultiInputSortingDataInput#emitNext, SortingThread#go, ...)


----------------------------------------------------------------
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:
users@infra.apache.org