Heterogeneous networks of computers are becoming commonplace in high-performance computing. Systems ranging from workstations to supercomputers are linked together by high speed networks. Until recently each computing resource on the network remained a separate unit, but now over 100 institutions worldwide are writing and running truly heterogeneous programs utilizing multiple computer systems to solve applications through the use of a software package called PVM.
PVM stands for Parallel Virtual Machine [13, 11, 3]. PVM is designed from the ground up with heterogeneity and portability as primary goals. As such it is one of the first software systems that allows machines with wildly different architectures and floating point representations to work together on a single computational task.
The Heterogeneous Network Project, being worked on by researchers at Oak Ridge National Laboratory, the University of Tennessee, and Emory University, is involved in the research and development of two software packages specifically designed to facilitate heterogeneous parallel computing. The first package is PVM, which can be used on its own or as a foundation upon which other heterogeneous network software can be built. The second package is called HeNCE, which stands for Heterogeneous Network Computing Environment [2, 1]. HeNCE is being built on top of PVM with the intention of simplifying the task of writing, compiling, running, debugging, and analyzing programs on a heterogeneous network. The goal is to make network computing accessible to scientists and engineers without the need for extensive training in parallel computing and allowing them to use resources best suited for a particular phase of the computation.
What follows is a description of the basic features of these two packages.