DXML: A High-performance Scientific Subroutine Library

Mathematical subroutine libraries for science and engineering applications are an important tool in high-performance computing. By identifying and optimizing frequently used, numerically intensive operations, these libraries help in reducing the cost of computation, enhancing portability, and improving productivity. The Digital eXtended Math Library is a set of public domain and Digital proprietary software that has been optimized for high performance on Alpha systems. In this paper, DXML and the issues related to library software technology are described. Specific examples illustrate how algorithms can be optimized to take advantage of the architecture of Alpha systems. Modern algorithms that effectively exploit the memory hierarchy enable DXML routines to provide substantial improvements in performance.