lindong28 opened a new pull request, #21576:
URL: https://github.com/apache/flink/pull/21576
## What is the purpose of the change
Improve the performance of Flink programs that reads data from
IteratorSourceReaderBase.
## Brief change log
Updated IteratorSourceReaderBase#pollNext() to push records to ReaderOutput
in a while loop.
## Verifying this change
```
DataStream<Long> stream = env.fromSequence(1, 1000000000L)
.map(x -> x)
.addSink(new DiscardingSink<>());
```
Here are the benchmark results by running the above program with
parallelism=1 and 5 runs per setup.
- Prior to the proposed change, the average execution time is 31.3 sec with
std=2.3 sec.
- After the proposed change, the average execution time is 46.1 sec with
std=5.1 sec.
- The proposed change increases throughput by 47.3%.
## Does this pull request potentially affect one of the following parts:
- Dependencies (does it add or upgrade a dependency): (yes / no)
- The public API, i.e., is any changed class annotated with
`@Public(Evolving)`: (yes / no)
- The serializers: (yes / no / don't know)
- The runtime per-record code paths (performance sensitive): (yes / no /
don't know)
- Anything that affects deployment or recovery: JobManager (and its
components), Checkpointing, Kubernetes/Yarn, ZooKeeper: (yes / no / don't know)
- The S3 file system connector: (yes / no / don't know)
## Documentation
- Does this pull request introduce a new feature? (yes / no)
- If yes, how is the feature documented? (not applicable / docs / JavaDocs
/ not documented)
--
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]
