In Microsoft Build for OpenJDK, an experimental compiler optimization is added that promotes scalar replacement of objects involved in merges of object allocations, improving garbage collection and performance.
OpenJDK, Microsoft’s open source Java distribution, has added an experimental feature to improve escape analysis compiler optimizations.
The October 2022 PSU release, issued by Microsoft on October 24, seeks to improve escape analysis results by increasing the number of opportunities for scalar replacement.
Compiler optimisation involves escape analysis, an algorithm for determining the dynamic scope of pointers, or where a pointer can be accessed within the program. Scalar replacement enables the Java just-in-time compiler to allocate the members, or scalars, of an object on the stack or in registers instead of allocating the full object on the heap. Garbage collection is less stressed.
It is the goal of the new escape analysis enhancement to simplify object allocation merges in order to facilitate scalar replacement of the objects involved in the merge. A simpler version of this enhancement was included in Java Development Kit (JDK) 11, while a version exploiting more opportunities was included in Java Development Kit (JDK) 17.
You can find updates for OpenJDK 11 and OpenJDK 17 here. Downloads of the Microsoft Build of OpenJDK, which is available for Windows, Linux, and macOS, can be found at learn.microsoft.com. As of this release, developers can inspect source code in newly established Microsoft repositories on GitHub, at jdk17u and jdk11u, to observe the exact changes made to the builds.
Microsoft has updated its container images to include the Microsoft CBL-Mariner 2.0 Linux distribution as well as OpenJDK 8 images based on the CBL-Mariner with Eclipse Temurin binaries. Azul Zulu for Azure Builds of OpenJDK are no longer supported or updated by Microsoft.
Kubernetes is highlighted in Oracle Cloud Native Environment 1.5.7
There are a number of improvements in Oracle Cloud Native Environment 1.5.7, including faster deployments…
Python Releases Python 3.11.0 With Big Performance
Improved performance and more understandable error messages are included in the latest release…