| Back | Map | Glossary | Information |
Third is the HP/Convex cluster, with the number-crunching Convex Exemplar...
Wilmer Zehr; access
All the machines in this pyramid use the same HPUX operating systems. The
Convex Exemplar typifies a distributed shared memory system. The Exemplar, which
consists of eight networked hypernodes (each hypernode consists of eight
super-fast processors), can be easily upgraded to 16 or more hypernodes,
making it a likely candidate for a teraflop machine in the future.
JPEG Image (14.1 KB)
NCSA's Convex Exemplar has:
The Convex Exemplar is complemented by a cluster of Hewlett-Packard workstations. The entire system is part of a joint NCSA and University of Illinois Computer Science Department project aimed at evaluating distributed shared memory computer architectures The project is funded by the National Science Foundation and the Advanced Research Projects Agency (ARPA)'s Information Technology Office (ITO).
Because the Exemplar is so new, scientists are still experimenting with it. But NCSA cosmologists Michael Norman and Greg Bryan, who create computer simulations of the evolution of galaxies and other large structures in the universe, are in the process of moving their cosmology simulation codes to the Exemplar. Cosmological simulations are large and complex, requiring very fast processors capable of communicating with each other very efficiently -- a shared memory problem. The Exemplar, with it fast processors and distributed memory system, may prove to be just the machine for the task.
Return to Processors