Benchmarking HPJava: Prospects for Performance

Han-Ku Lee, Bryan Carpenter, Geoffrey Fox, Sang Boem Lim
{hkl, dbc, gcf, slim}@grids.ucs.indiana.edu

Pervasive Technology Labs
Indiana University
Bloomington, Indiana 47401-3730

Computational Science and Information Technology,
Florida State University,
Tallahassee, Florida 32306-4120

Abstract:

We briefly review the authors' HPJava programming environment, and compare and contrast with systems like HPF. Because the underlying programming language is Java, and because the HPJava programming model relies centrally on object-oriented run-time descriptors for distributed arrays, the achievable performance has been somewhat uncertain. Individual node performance is the most critical issue. We argue with simple benchmarks that we can in fact hope to achieve performance in a similar ballpark to more traditional HPC languages¹.