Who cares about speed
summary
Argues that speed includes developer feedback loops, startup time and memory churn not just CPU cycles and shows concrete wins: replacing Lombok/equals with precompute hashCode yields ~20% faster hash-collection throughput; a tiny custom autoconfiguration filter cuts Spring Boot cold-start from ~1.1–1.2s to ~0.9–1.0s; profiling with GCViewer exposes 1,700 minor GCs and ~1.1s wasted in GC while sending 55 GB, and a small Netty allocator fix reduces lifetime allocations to ~200 MB; recommends JMH, JCStress and the 80/20 rule.
AI-written summary · about
chapters (20)
- 0:00Introduction and background
- 2:38History of computing speed
- 6:18Distractions from performance focus
- 10:20Legacy code misconceptions
- 14:16Prioritizing speed in development
- 16:03Identifying inefficient code
- 21:11Loop performance in Java
- 21:35Optimizing Spring Boot startup
- 26:59HashCode and equals performance
- 32:47Garbage collection inefficiencies
- 35:42Quick wins in optimization
- 37:26Writing simple maintainable code
- 38:53The role of testing
- 40:23Collaboration and feedback
- 42:32Efficient meetings and planning
- 45:03Pair programming benefits
- 47:11Feature toggles and bloat
- 48:20Reusing code and methodologies
- 49:01Adapting development methodologies
- 51:09Q&A and closing remarks
related talks
48mLean Spring Boot Applications for the Cloud
06mSimple Tweaks to Get the Most Out of your JVM
39mHow to Reduce the Footprint of Your Spring Boot Applications
11mPatrick Baumgartner at JNation 2024
40mLudicrous Speed: Designing For Performance on the JVM
42mPractical Optimizations
42mSimple tweaks to get the most out of your JVM
39mMemory Efficient Java