Basic HTML version of Foils prepared 21 October 1995

Foil 16 Point to Point Networks (Store and Forward) -- I

From New CPS615 Foils-- B 28 August 95 CPS615 Basic Simulation Track for Computational Science -- Fall Semester 95. by Geoffrey C. Fox
1 Processors directly connected to only certain other processors, and must go multiple hops to get to additional processors. Also called store-and-forward.
2 Usually distributed memory
3 Examples:
  • ring
  • mesh, torus
  • hypercube -- as in original Caltech/JPL machines
  • binary tree
4 Usually NUMA, nonblocking, scalable, upgradable
5 Processor connectivity modeled by a graph in which nodes represent connections between processors. Much theoretical work here but now obsolete (irrelevant) as even if this hardware used, pipelining masks transmission time which ends up not showing term proportional to distance travelled
in Table To:

© on Tue Oct 7 1997