Maximizing performance with GraalVM

on YouTube ↗59m

" summary

GraalVM shows concrete wins: using JMH for measurements, escape analysis removes allocations so boxed-int and array allocations vanish in practice, and some HotSpot vs Graal comparisons show up to ~30x slower on HotSpot for complex Stream-heavy microbenchmarks. Building a native-image produces small standalone binaries (~7–16 MB in demos), cuts Micronaut startup from ~1.8s to ~20ms and RSS from ~300 MB to ~28 MB, and compiles javac runs from ~600 ms to ~49 ms. The native-image snapshotting works but requires explicit reflection/dynamic-loading config.

AI-written summary · about

" chapters (11)

1. 0:00Introduction to performance
2. 1:03Performance metrics explained
3. 4:34Overview of GraalVM
4. 7:45Escape analysis demo
5. 13:28Benchmarking abstractions
6. 19:10Stream API optimizations
7. 23:44Native image technology
8. 30:53Native image demo
9. 37:25Memory usage analysis
10. 43:46Polyglot capabilities
11. 50:07Production readiness and Q&A

" related talks

• 45mMaximizing performance with GraalVMOleg Šelajev · Øredev 2019
• 52mMaximizing Performance With GraalVMOleg Šelajev · Devnexus 2020
• 56mMaximizing Performance with GraalVMOleg Šelajev · Developer Summit 2020
• 20mMaximizing Java Application Performance with GraalVMOleg Šelajev · GOTO Chicago 2020
• 42mJava in the Cloud with GraalVMAlina Yurenko · GOTO Aarhus 2023
• 03h 07mEverything you need to know about GraalVMThomas Wuerthinger, Oleg Šelaje · Devoxx Belgium 2019
• 44mJava In The Cloud with GraalVMAlina Yurenko · GOTO Amsterdam 2023
• 34mFast and Efficient Java Microservices with GraalVMAlina Yurenko · Devnexus 2021