sunhaibotb commented on a change in pull request #8409: [FLINK-12478] Decompose
monolithic run-loops in StreamTask implementa…
URL: https://github.com/apache/flink/pull/8409#discussion_r284588294
##########
File path:
flink-streaming-java/src/main/java/org/apache/flink/streaming/runtime/tasks/OneInputStreamTask.java
##########
@@ -98,13 +96,8 @@ public void init() throws Exception {
}
@Override
- protected void run() throws Exception {
- // cache processor reference on the stack, to make the code
more JIT friendly
- final StreamInputProcessor<IN> inputProcessor =
this.inputProcessor;
-
- while (running && inputProcessor.processInput()) {
- // all the work happens in the "processInput" method
- }
+ protected boolean performDefaultAction() throws Exception {
Review comment:
I only tested the following two cases, Case 1 is much faster than Case 2.
These two cases are a little different from the change here, and I can not sure
if the change here has a performance regression.
Case 1:
```
=== TwoInputXStreamTask.java ===
@Override
protected void run() throws Exception {
// cache processor reference on the stack, to make the code more JIT
friendly
final StreamTwoInputXProcessor<IN1, IN2> inputProcessor =
this.inputProcessor;
inputProcessor.processInput();
}
=== StreamTwoInputXProcessor.java ===
public boolean processInput() throws Exception {
while (running) {
...
if (element.isRecord()) {
continue;
}
...
}
return true;
}
```
Case 2:
```
=== TwoInputXStreamTask.java ===
@Override
protected void run() throws Exception {
// cache processor reference on the stack, to make the code more JIT
friendly
final StreamTwoInputXProcessor<IN1, IN2> inputProcessor =
this.inputProcessor;
while (running && inputProcessor.processInput()) {
// all the work happens in the "processInput" method
}
}
=== StreamTwoInputXProcessor.java ===
public boolean processInput() throws Exception {
while (true) {
...
if (element.isRecord()) {
return true;
}
...
}
}
```
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
