Book Image

Supercharge Your Applications with GraalVM

By : A B Vijay Kumar
Book Image

Supercharge Your Applications with GraalVM

By: A B Vijay Kumar

Overview of this book

GraalVM is a universal virtual machine that allows programmers to compile and run applications written in both JVM and non-JVM languages. It improves the performance and efficiency of applications, making it an ideal companion for cloud-native or microservices-based applications. This book is a hands-on guide, with step-by-step instructions on how to work with GraalVM. Starting with a quick introduction to the GraalVM architecture and how things work under the hood, you'll discover the performance benefits of running your Java applications on GraalVM. You'll then learn how to create native images and understand how AOT (ahead-of-time) can improve application performance significantly. The book covers examples of building polyglot applications that will help you explore the interoperability between languages running on the same VM. You'll also see how you can use the Truffle framework to implement any language of your choice to run optimally on GraalVM. By the end of this book, you'll not only have learned how GraalVM is beneficial in cloud-native and microservices development but also how to leverage its capabilities to create high-performing polyglot applications.
Table of Contents (17 chapters)
1
Section 1: The Evolution of JVM
4
Section 2: Getting Up and Running with GraalVM – Architecture and Implementation
8
Section 3: Polyglot with Graal
13
Section 4: Microservices with Graal

Understanding Graal compiler optimizations

The Graal compiler performs some of the most advanced optimizations on the code just in time. The most critical ones are discussed in the following subsections.

Before getting into this session, please refer to the Understanding the optimizations performed by JIT section of Chapter 2, JIT, HotSpot, and GraalJIT.

Speculative optimization

JIT compilation relies heavily on the runtime profiling of the code. As we have seen, the graphs are optimized based on the HotSpots. HotSpots, as we covered in Chapter 2, JIT, HotSpot, and GraalJIT, are the control flows that the program goes through most frequently. There is no point in trying to optimize the whole code; instead, the JIT compiler tries to optimize the hot control paths/flows. This is based on speculation and assumption. When an assumption is proven wrong during execution, the compiler quickly deoptimizes and waits for another opportunity to optimize based on new HotSpots. We covered...