NPAC Technical Report SCCS-265

Prospects for Parallel Computing

Geoffrey Fox

Submitted June 01 1992


Abstract

We describe the state of parallel computing in 1992. The technological driving force is the increasing performance of VLSI chips. This leads to nodes which can be joined together in several different architectures to provide powerful machines leading to teraflop performance in 1996 and petaflops 20 years later. This is put in the context of the federal HPCC initiative. We describe the analogies with society and discuss illustrative examples with a chemical flavor. Simple examples of SIMD, MIMD and vector supercomputers are contrasted. On the software side, parallel operating systems and both data parallel and message passing programming environments are briefly reviewed. The types of applications -- Grand Challenges from academia and industry -- expected to benefit from HPCC are noted.


PostScript version of the paper

Hypertext version of the paper