String Enhancements in JDK 9
summary
Benchmarks run with JMH and a GC profiler on JDK 9 Early Access builds (B92 baseline, B93 with compact strings, B104 adding the indy string concat call) show concrete wins: baseline allocates ~1.5KB per concat, compact strings drops that to ~800B, and enabling the inline_size strategy for indy concat cuts allocations ~3x and doubles throughput (over 2x faster) on a real microbenchmark. Compact strings is transparent on upgrade; indy concat requires one recompilation or jlink bytecode rewrite to get the full benefit.
AI-written summary · about
related talks
46mEnough java.lang.String to Hang Ourselves...
48mEnough java.lang.String to Hang Ourselves ...
46mEnough java.lang.String to Hang Ourselves ...
40mWhat Every Java Programmer Should Know About Strings
53mJava Performance Update 2024
41mJava 24, Faster Than Ever
51mRethinking Java String Concatenation
48mJava 9 Performance
