This book explains how the Z Garbage Collector was designed, why it was designed this way, the common pitfalls to avoid when using ZGC in JDK 25 and how to get useful information from the garbage collector.
With a unique angle right between academia and industry, it is useful for computer science students and Java developers alike.
This book explains how the Z Garbage Collector was designed, why it was designed this way, the common pitfalls to avoid when using ZGC in JDK 25 and how to get useful information from the garbage collector.
Java is one of the most widely used programming languages in the world. It is a memory-safe language that completely relies on a good garbage collection (GC) algorithm to work well. The Z Garbage Collector of OpenJDK is arguably the most sophisticated production GC algorithm for any managed language runtime. The book goes into detail and covers core algorithm concepts of ZGC: coloured pointers, GC barriers, concurrent marking, concurrent relocation, concurrent stack scanning, concurrent root scanning, concurrent class unloading, concurrent generational collection, as well as showing how all of this works. It explains how the JVM evolved certain features to be able to accommodate ZGC, and covers the importance of each of the various GC heuristics and how they function.
With a unique angle right between academia and industry, this book is useful for computer science students and Java developers alike.
1. Introduction
2. ZGC Design Overview
3. ZGC Memory Layout
4. ZGC
Generations
5. ZGC Barriers
6. ZGC Collection Phases
7. ZGC Root Scanning
8.
ZGC Class Unloading
9. ZGC Collection Triggers
10. ZGC Thread Sizing
Heuristics
11. ZGC Heap Sizing Policies
12. Safepoint Polls
13. Thread-Local
Handshakes
14. Compiled Method Entry Barriers
15. Compiled Barriers
16.
Virtual Threads
17. Deploying ZGC
18. Measuring Latency
19. Measuring
Throughput
20. Measuring Memory Footprint
21. Observability
22. ZGC Heap
Sizing
Erik Österlund did his BSc, MSc and PhD at the Linnaeus University, Sweden, where his interests and thesis work were all focused on garbage collection. At a memory management conference in Beijing 2012, Erik met Jesper Wilhelmsson from an Oracle development office in Stockholm. They were both surprised to find two Swedes in Beijing, interested in memory management. One thing led to another, and a research internship was arranged. There, Erik was talking about how a concurrent GC algorithm was needed. In 2016, Erik joined Oracle and started working on ZGC, and has been one of the core ZGC developers ever since.
