Guide to Java for High Performance Computing

Poor performance on numerical codes slow adoption of Java within the technical computing community. This guidebook describes a prototype array library and a research prototype compiler that supports standard Java and delivers near-Fortran performance on numerically intensive codes.

---

Poor peformance on numerical codes continues to slow adoption of Java within the technical computing community. A new guidebook describes a prototype array library and a research prototype compiler that supports standard Java and delivers near-Fortran performance on numerically intensive codes. In addition, the reference discuss addresses implementing an efficient Java package for true multidimensional arrays; compiler techniques for generating efficient access to these arrays; and compiler optimizations that create safe, exception-free regions of code that can be aggressively optimized. These techniques work together synergistically to make Java an efficient language for technical computing. In a set of four benchmarks, the authors achieve between 50 per cent and 90 per cent of the performance of highly optimized Fortran code. This represents a several-fold improvement compared to what can be achieved by the next best Java environment. Consequently, it proves that Java is a competitive programming language for developing high-performance numerical applications.