PathScale customers now include NASA, the U.S. Air Force, the U.S. Navy, four of the leading Department of Energy (DOE) National Labs, the European Center for Medium Range Weather Forecasting (ECMWF), three of the largest HPC centres in Germany, one of the largest automotive manufacturers in Europe and the Victorian Partnership for Advanced Computing (VPAC). Some of the leading universities purchasing the PathScale compilers include the University of Utah, New York University, Penn State University, University of Georgia, Cambridge University, Warsaw University, University of Zurich and University of Chicago.

"Everyone on our campus who is developing applications for Linux clusters and is truly concerned about performance is using PathScale", stated Martin Cuma, scientific applications programmer at the University of Utah. "We recently implemented a 1000-CPU PathScale-optimized AMD Opteron-based cluster that runs our scientific applications two to three times faster than was possible with our previous supercomputer system."

PathScale's compiler technology has been contributing to the rapid growth in demand for AMD64-based HPC systems, largely due to the industry-leading price/performance of AMD Opteron systems tuned with PathScale compilers.

"Our computing centre supports about 500 researchers running a mix of applications such as molecular dynamics, structure analysis, fluid dynamics and gravitational physics", stated Vijay Agarwala, Director of High-Performance

Computing and Visualization, Information Technology Services, Penn State University. "Reduced time to discovery, reduced time to results and optimal use of computing resources are very important. The PathScale compiler helps us maintain a cutting edge and get the best possible performance from our Opteron cluster system at all times."

PathScale's accelerating compiler suite revenues and traction within the HPC user community has established a strong foundation for the additional HPC tools and technology offerings that the company plans to announce at the upcoming SuperComputing 2004 exhibition in Pittsburgh in November.