COMPUTING APPLICATIONS WINDOWS NT SUPERCOMPUTER DESIGN AND EVALUATION OF AN HPVM-BASED WINDOWS NT SUPERCOMPUTER

We describe the design and evaluation of a 192-processor Windows NT cluster for high performance computing based on the High Performance Virtual Machine (HPVM) communication suite. While other clusters have been de-scribed in the literature, building a 58 GFlop/s NT cluster to be used as a general-purpose production machine for NCSA required solving new problems. The HPVM soft-ware meets the challenges represented by the large number of processors, the peculiarities of the NT operat-ing system, the need for a production-strength job submis-sion facility, and the requirement for mainstream program-ming interfaces. First, HPVM provides users with a collec-tion of standard APIs like MPI, Shmem, Global Arrays with supercomputer class performance (13 m s minimum la-tency, 84 MB/s peak bandwidth for MPI), efficiently deliver-ing Myrinet’s hardware performance to application pro-grams. Second, HPVM provides cluster management and scheduling (through integration with Platform Comput-ing’s LSF). Finally, HPVM addresses Windows NT’s re-mote access problem, providing convenient remote ac-cess and job control (through a graphical Java-applet front-end). Given the production nature of the cluster, the performance characterization is largely based on a sam-ple of the NCSA scientific applications the machine will be running. The side-by-side comparison with other present-generation NCSA supercomputers shows the cluster to be within a factor of 2 to 4 of the SGI Origin 2000 and Cray T3E performance at a fraction of the cost. The inherent scalability of the cluster design produces a comparable or better speedup than the Origin 2000 despite a limitation in the HPVM flow control mechanism.