pnowojski commented on a change in pull request #31:
URL: https://github.com/apache/flink-benchmarks/pull/31#discussion_r710969942
##########
File path: src/main/java/org/apache/flink/benchmark/MultipleInputBenchmark.java
##########
@@ -36,18 +50,20 @@
import org.openjdk.jmh.runner.options.OptionsBuilder;
import org.openjdk.jmh.runner.options.VerboseMode;
-public class MultipleInputBenchmark extends BenchmarkBase {
+import java.util.concurrent.CompletableFuture;
+public class MultipleInputBenchmark extends BenchmarkBase {
public static final int RECORDS_PER_INVOCATION =
TwoInputBenchmark.RECORDS_PER_INVOCATION;
public static final int ONE_IDLE_RECORDS_PER_INVOCATION =
TwoInputBenchmark.ONE_IDLE_RECORDS_PER_INVOCATION;
+ public static final int CHAINED_IDLE_RECORDS_PER_INVOCATION = 3000;
Review comment:
With throughput so low, you are not benchmarking any performance
critical code path. I'm not sure what's causing this problems, maybe indeed
collecting the results, but whatever it is, it's the performance bottleneck,
which your benchmark is benchmarking. And if collecting results via HTTP
request is that bottleneck, that's definitely not what we want to benchmark.
On the other hand, something might be still very wrong with this code, and
the performance regression might be still out there. Without properly working
benchmark how do you know you have fixed the performance regression?
--
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]
