This talk discuss advantages and limits of each GC (G1, Serial, Parallel, Z & Shenandoah) to choose the one that fits according to the type of applications We discuss in detail how to tune GC depending on the case: Which GC to use for a batch oriented application? How to tune GC to maximise time to process data and optimize resources?
Which GC for user request/response applications more latency sensitive? What is the difference between Z and Shenandoah? How to size to take best advantage of them?
G1: What are the issues (RememberSet, Humongous, Young gen resize, …)? How to tune it correctly depending on your resources?