Book Image

Java 9 High Performance

By : Mayur Ramgir, Nick Samoylov
Book Image

Java 9 High Performance

By: Mayur Ramgir, Nick Samoylov

Overview of this book

Finally, a book that focuses on the practicalities rather than theory of Java application performance tuning. This book will be your one-stop guide to optimize the performance of your Java applications. We will begin by understanding the new features and APIs of Java 9. You will then be taught the practicalities of Java application performance tuning, how to make the best use of garbage collector, and find out how to optimize code with microbenchmarking. Moving ahead, you will be introduced to multithreading and learning about concurrent programming with Java 9 to build highly concurrent and efficient applications. You will learn how to fine tune your Java code for best results. You will discover techniques on how to benchmark performance and reduce various bottlenecks in your applications. We'll also cover best practices of Java programming that will help you improve the quality of your codebase. By the end of the book, you will be armed with the knowledge to build and deploy efficient, scalable, and concurrent applications in Java.
Table of Contents (11 chapters)

Monitoring threads

There are two ways to monitor threads, programmatically and using the external tools. We have already seen how the result of a worker calculation could be checked. Let's revisit that code. We will also slightly modify our worker implementation:

class MyRunnable03 implements Runnable {
private String name;
private double result;
public String getName(){ return this.name; }
public double getResult() { return this.result; }
public void run() {
this.name = Thread.currentThread().getName();
double result = IntStream.rangeClosed(1, 100)
.flatMap(i -> IntStream.rangeClosed(1, 99999))
.takeWhile(i -> !Thread.currentThread().isInterrupted())
.asDoubleStream().map(Math::sqrt).average().getAsDouble();
if(!Thread.currentThread().isInterrupted()){
this.result = result;
}
}
}

For the worker thread identification, instead of...