Ahead of the OpenJDK Java 24 release expected in the coming hours today, Oracle already released “GraalVM for JDK 24 Community 24.0” edition.
This high-performance JDK distribution that also supports spinning Java apps into their own standalone binaries and supporting other programming languages continues advancing. GraalVM for JDK 24 picks up more features to enhance not only the Java support but also for Ruby, Python, JavaScript, and Node.js.
GraalVM for JDK 24 brings a new generation of machine learning powered profile inference: the Graal Neural Network Static Profiler. With the Graal Neural Network “GNN” code, the developers observed a 7.9% peak performance improvement across a wide range of micro-services benchmarks. The Graal Neural Network profiler is enabled with the “-O3” optimization level when building Graal Native Image binaries.
GraalVM also improves its experimental Vector API support for Native Image Ahead Of Time (AOT) compilation, introduces the “SkipFlow” experimental extension for Native Image static analysis, better Software Bill of Materials (SBOM) integration, support for Java module system-based service loading, experimental Bytecode DSL as a new framework for implementing bytecode interpreters within the context of the Truffle framework, and other changes.
There are also enhancements to the GraalPy Python code, GraalWasm for WebAssembly, and new ECMScript features for GraalJS JavaScript.
Downloads and more details on today’s GraalVM for JDK 24 Community 24.0.0 release via GitHub.