The Fundamentals of Garbage Collector Performance
summary
Frames garbage-collection choices as tradeoffs: throughput, latency, footprint and a fourth capacity (CPU available to the app). Uses generational GC details — eden, two survivor spaces, promotion-age/tenuring and old gen — and concrete knobs: increase eden (-Xmn or adjust young size) to double time between minors, tune SurvivorRatio and max tenuring to avoid premature promotions, or raise G1 pause target / InitiatingHeapOccupancyPercent. Notes concurrent collectors cut pauses but cost CPU/memory, and warns against soft references as caches.
AI-written summary · about
chapters (13)
- 0:00Agenda
- 0:26The 203 Principle
- 3:17Generational GCs
- 5:43Weak Generational Hypothesis
- 10:30Frequency
- 16:35Strategies
- 20:33Old Generation
- 28:17Reducing Concurrent Collections
- 29:42Duration of Old Generation Collections
- 31:03Soft References
- 32:33Old Generation Collectors
- 35:41Reducing Old Generation Collectors
- 38:19Summary
related talks
51mJava Garbage Collection - The Basics
52mGarbage collection for the humans
50mTrash Talk - Exploring the JVM memory management
16mGarbage Collectors: Battle Royale
46mTrash Talk - Exploring the Memory Management in the JVM
01mThe evolution of Garbage Collection Algorithms in Java
50mTrash Talk - Exploring the JVM memory system
25mGarbage Collection Primer