Jack J. Dongarra
Oak Ridge National Laboratory and University of Tennessee
G. A. Geist
Oak Ridge National Laboratory
Robert Manchek
University of Tennessee
V. S. Sunderam
Emory University
The Parallel Virtual Machine (PVM), an integrated framework for heterogeneous network computing, lets scientists exploit collections of networked machines when carrying out complex scientific computations. Under PVM, a user-defined grouping of serial, parallel, and vector computers appears as one large distributed-memory machine. Configuring a personal parallel virtual computer involves simply listing the names of the machines in a file that is read when PVM is started. Applications can be written in Fortran 77 or C and parallelized by use of message-passing constructs common to most distributed-memory computers. With the use of messages sent over the network, multiple tasks of an application can cooperate to solve a problem in parallel.
This article discusses components of PVM, including the programs and library of interface routines. It summarizes the characteristics of appropriate applications and discusses the current status and availability of PVM. In addition, the article introduces a recent extension to PVM known as the Heterogeneous Network Computing Environment (HeNCE).