Revision of JMH Tests

[benjchristensen]

This updates all perf testing to use only JMH. I tried to get all of them to follow a common pattern, and to comply with examples at https://github.com/nitsanw/jmh-samples/blob/master/src/main/java/org/openjdk/jmh/samples/ by @nitsanw. (@nitsanw if you're willing to take a look at this and confirm or correct them, I'd appreciate that).

I am doing this in 0.19 so that work on 0.20 has a common set of perf tests to use for comparison.

[cloudbees-pull-request-builder]

RxJava-pull-requests #1315 SUCCESS
This pull request looks good

[nitsanw]

Sorry, not sure when I'll be able to review these changes.
In the next few days/weeks perhaps…

[benjchristensen]

@nitsanw No problem, I know you're very busy. If you ever have the chance and interest, it would be good to know we're following your examples right. In the meantime though I'm pushing forward based on my current understandings of how to use JMH. I appreciate your responses and involvement in our project!

[nitsanw]

My examples are very dated by now, you should look at the latest samples on the JMH project.

On 25 Jun 2014, at 18:34, Ben Christensen [email protected] wrote:

@nitsanw No problem, I know you're very busy. If you ever have the chance and interest, it would be good to know we're following your examples right. In the meantime though I'm pushing forward based on my current understandings of how to use JMH. I appreciate your responses and involvement in our project!

—
Reply to this email directly or view it on GitHub.

[benjchristensen]

I ended up finding those after looking at yours, and went through them as well. I assume you're referring to these? http://hg.openjdk.java.net/code-tools/jmh/file/tip/jmh-samples/src/main/java/org/openjdk/jmh/samples/

[nitsanw]

Yes. :)

On 25 Jun 2014, at 21:22, Ben Christensen [email protected] wrote:

I ended up finding those after looking at yours, and went through them as well. I assume you're referring to these? http://hg.openjdk.java.net/code-tools/jmh/file/tip/jmh-samples/src/main/java/org/openjdk/jmh/samples/

—
Reply to this email directly or view it on GitHub.

[shipilev]

Hi there :) Not sure if that's intentional or not, but poking around a few benchmarks there, it seems you can make them even cleaner: instead of using the explicit Input nested class, you might as well use the enclosing class holding the @Benchmark methods as @State class. You seem to be having @State(Thread) already on the enclosing classes, only to delegate everything to nested class afterwards...

That is, this sample works:
http://hg.openjdk.java.net/code-tools/jmh/file/0.9/jmh-samples/src/main/java/org/openjdk/jmh/samples/JMHSample_04_DefaultState.java