pnowojski commented on code in PR #22852:
URL: https://github.com/apache/flink/pull/22852#discussion_r1246464651
##########
flink-runtime/src/test/java/org/apache/flink/runtime/source/coordinator/SourceCoordinatorAlignmentTest.java:
##########
@@ -131,6 +138,101 @@ void testWatermarkAlignmentWithTwoGroups() throws
Exception {
}
}
+ @Test
+ @Timeout(5)
+ void testWatermarkAggregatorBenchmark() {
Review Comment:
I think it would be valuable to have something like that. Otherwise we
wouldn't spot a performance regression in this area.
Option 1. should be very easy to implement, but I dunno if we can make this
optimisation actually visible in a microbenchmark on a single machine (so
probably max parallelism of sth between 16 to 128)
Option 2. would be more work, but you could check how the `Scheduler`
benchmarks have been implemented.
<img width="818" alt="Screenshot 2023-06-29 at 12 53 44"
src="https://github.com/apache/flink/assets/8957547/5c677291-ae2f-4eba-83fc-66be64f1e11f";>
I have never looked in the code, but I wouldn't be surprised that it's a
very similar construct that you would need for Option 2.
--
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]
