Jump to content

Caltech Cosmic Cube

From Wikipedia, the free encyclopedia

The Caltech Cosmic Cube was a parallel computer, developed by Charles Seitz and Geoffrey C Fox from 1981 onward.[1] It was the first working hypercube built.[2]

It was an early attempt to capitalise on VLSI to speed up scientific calculations at a reasonable cost. Using commodity hardware and an architecture suited to the specific task (QCD), Fox and Seitz demonstrated that this was indeed possible.

In 1984 a group at Intel including Justin Rattner and Cleve Moler developed the Intel iPSC inspired by the Cosmic Cube.[3] In 1987 several people in the group formed a company called Parasoft to commercialize the message passing interface developed for the Cosmic Cube.[4]

Characteristics

[edit]
  • 64 Intel 8086/87 processors[5]
  • 128kB of memory per processor
  • 6-dimensional hypercube network, i. e. each processor can directly exchange data with six other processors.

References

[edit]
  1. ^ Cosmic Cubism from Engineering & Science, March 1984 http://calteches.library.caltech.edu/3419/1/Cubism.pdf
  2. ^ Anderson, A. John (1994). Foundations of Computer Technology. CRC Press. p. 378. ISBN 978-0412598104.
  3. ^ Cleve Moler (October 28, 2013). "The Intel Hypercube, part 1". Retrieved November 4, 2013.
  4. ^ History of Supercomputing
  5. ^ Birth of the Hypercube
  • The Torus Routing Chip
  • Parallel Computer Archival Documents
  • John Apostolakis, Clive Baillie, Robert W. Clayton, Hong Ding, Jon Flower, Geoffrey C. Fox, Thomas D. Gottschalk, Bradford H. Hager, Herbert B. Keller, Adam K. Kolawa, Steve W. Otto, Toshiro Tanimoto, Eric F. van de Velde, J. Barhen, J. R. Einstein, and C. C. Jorgensen. 1989. "Supercomputer applications of the hypercube"—In Supercomputing systems: architectures, design, and performance, Svetlana P. Kartashev and Steven I. Kartashev (Eds.). Van Nostrand Reinhold Co., New York, NY, USA:1989 Pages 480–577.
[edit]