Design and Evaluation of an HPVM-Based Windows NT Supercomputer

Abstract

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 described 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 software meets the challenges represented by the large number of processors, the peculiarities of the NT operating system, the need for a production-strength job submission facility, and the requirement for mainstream programming interfaces. First, HPVM provides users with a collection of standard APIs like MPI, Shmem, Global Arrays with supercomputer class performance (13 μs minimum latency, 84 MB/s peak bandwidth for MPI), efficiently delivering Myrinet’s hardware performance to application programs. Second, HPVM provides cluster management and scheduling (through integration with Platform Computing’s LSF). Finally, HPVM addresses Windows NT’s remote access problem, providing convenient remote access 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 sample 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.

Entropia: architecture and performance of an enterprise desktop grid system

“…At approximately the same time, the scientific computing community began to move from large expensive supercomputers to relatively inexpensive Unix workstations [45] and, in the late 1990s, to low-cost PC hardware [7,42]. During this time, clusters of inexpensive PCs connected with high-speed interconnects were demonstrated to rival supercomputers for some applications.…”

Design of a performance technology infrastructure to support the construction of responsive software

“…Sweep3D is a complex benchmark for evaluating wavefront application techniques used to evaluate advanced parallel architectures at the Los Alamos Laboratories. The application uses the HPVM-MPI version 1.9 library from University of Illinois [3] for inter-processor communication. The application source code was analyzed by the PACE static performance analysis tools and produced a CHIP 3 S script [2].…”

Solo: A Lightweight Virtual Machine

“…These works make VM suit for the high performance applications better, but they mainly focus on I/O subsystem, especially on the network performance, and ignore the virtualization of the CPU and memory. High level middleware-and language-based VMs have also been studied and implemented for the high performance applications, examples include PVM [33] and the HPVM [34]. However, in this paper, we only address the issue of system-level VM, because it provides more flexibility, and we can even run the language-based VMs on it.…”