We describe a parallel computer system for processing media: audio,vid
eo, and graphics, among others. The system supports medium to coarse g
rain parallelism, using a dataflow model of execution, on a range of m
achine architectures scaling from a single von Neumann or general purp
ose processor (GPP) up to networks of several hundred heterogeneous pr
ocessors. A distributed resource manager, extending or subsuming the f
unctionality of a traditional operating system, is an integral and nec
essary part of the system. While we are building a system for processi
ng a variety of media, in this paper we concentrate on video because i
t provides an extreme case in terms of both data rates and available p
arallelism. (C) 1997 Elsevier Science B.V.