jmh VS async-profiler

JMH for benchmarks

I would recommend using JMH instead.

Reflection is probably very roughly between 10 and 1000 times slower. Why don't you measure it yourself using JMH?

Aleksey Shipilev is another such leader, whom is especially knowledgeable about the internals of the JVM. His writings are invaluable. He is (was) the lead of the Java microbenchmark framework (JMH} which is how one would write small performance experiments in Java, and learn what really makes a difference or now.

In the last episode we implemented a parallelMap operation using streams, raw threads, a threadpool with futures, and coroutines. At first glance the raw threads was quickest, followed by futures, coroutines and then streams. In this, part 56 of an exploration of where a Test Driven Development implementation of the Gilded Rose stock control system might take us in Kotlin, we investigate the performance of the different functions further, in particular digging down into why coroutines seem to be slow and finding a way to speed them up. We also find a way to use a particular ForkJoinPool to run the streams code, making it as fast as the others (bar the raw threads). Frankly we only use very rough benchmarks here, with no statistical testing except 'it looks like'. That's OK for gross differences, but is highly suspect when deciding which of two similarly performant approaches is faster. For that check out JMH and you could watch my video from KotlinConf 2017

A few months ago, we started to use JMH in our project to test and find performance issues. The tool provides multiple modes and profilers, and we found this useful for our purposes.

Using the Async Profiler we were able to build flamegraphs for the first and second queries to picture the differences in path length of the two transactions execution.

> The other two languages I’ve used mostly in recent decades are Java and Ruby and the profiler situation is for both those languages is kind of shitty. I had to pay real money to get the Java profiler I used at AWS and while it worked, it was klunky, not fun to use.

These days, async profiler (https://github.com/jvm-profiling-tools/async-profiler) is much better than the Go tooling for performance. It is a joy to use and features a top-like view for the hottest methods. It works for locks, allocations and CPU time. It also integrates with JMH.

Also, running it under a profiler (I recommend async-profiler[1]) should give you a good idea of where the slowdown occurs which might help you pin it down further.

[1] https://github.com/jvm-profiling-tools/async-profiler

Profile it. async-profiler is really great. Alternatively you can check out VisualVM/JProfiler/YourKit

async-profiler is really great at analyzing allocations, give it a shot!

JFR to Flame Graph Converter

Async Profiler converter tool does support JFR to Flame Graph, JFR to FlameScope, collapsed stacks to Flame Graph -https://github.com/jvm-profiling-tools/async-profiler#download