Performance Improvements in JDK 24

a year ago

Performance improvements in JDK 24 compared to JDK 23.
Core Libraries: Enhancements in Foreign Function & Memory API (FFM) bulk operations for better performance.
String concatenation strategy improved by exploiting hidden classes, reducing runtime overhead.
SHA3 algorithm performance improved by up to 27% by reducing conversions between data types.
ClassFile API finalized in JDK 24, improving bytecode generation and reducing startup overhead.
Runtime: JEP 491 improves virtual thread scalability by reducing pinning during synchronization.
Secondary super cache (SSC) optimization for better type query performance in multithreaded applications.
String::indexOf performance improved by 1.3x using AVX2 SIMD instructions on x64 platforms.
G1 garbage collector barriers simplified with late expansion, reducing C2 compilation overhead.
JEP 483 introduces ahead-of-time (AOT) class loading and linking for improved startup and footprint.
Experimental JEP 450 reduces object header size to 64 bits, potentially lowering memory usage by 10-20%.
RISC-V platform improvements including string comparison and intrinsic implementations.
Community involvement encouraged for further performance enhancements and feedback.

Hasty Briefsbeta