While computer performance has improved dramatically, real productivity in terms of the science accomplished with ever-faster machines has not kept pace. Indeed, scientists are finding it increasingly costly and time consuming to write, port, or rewrite their software to take advantage of the new hardware. While machine performance remains a critical productivity driver for high-performance computing applications, software development time increasingly dominates hardware speed as the primary productivity bottleneck.

Removing or ameliorating productivity bottlenecks in next-generation high-performance computing systems is a key objective of DARPA’s High-Productivity Computing Systems (HPCS) Program.  By ourselves and as part of work for Sun Microsystems, we are exploring studies and tools to understand and improve the productivity of software development done in high performance computing enviroments.

Some of our work is described in:

• Il-Chul Yoon, Alan Sussman and Adam Porter, And Away We Go: Understanding the Complexity of Launching Complex HPC Applications. In Proceedings of the 2nd International Workshop on Software Engineering for High Performance Computing System Applications. St Louis, Mo, May 2005.
  
• Stuart Faulk, Adam Porter, Walter Tichy, Philip Johnson, John Gustafson, Larry Votta. Toward Accurate HPC Productivity Measurement. In Proceedings of the 1st International Workshop on Software Engineering for High Performance Computing System Applications, Edinburgh, Scotland, UK. May 2004. 
• Stuart Faulk, John Gustafson, Philip Johnson, Adam Porter, Walter Tichy and Lawrence Votta. Measuring High Performance Computing Productivity. Journal of High Performance Computing Applications. Dec. 2004.