Simulations of PRAM on complete optical networks
We study time-processor optimal simulation of EREW PRAM model on a completely connected optical communication parallel computer (OCPC). We propose a new direct algorithm for the realization of h-relation, penalty algorithm, which is very simple and compares favorably with the Geréb-Graus and Tsantilas algorithm. We study also generalizations of the basic OCPC, where each processor has ρ receivers instead of one and where π processors are coupled to a small PRAM module. Our experiments indicate that in this new model the greedy routing algorithm runs without deadlock and fast. Indeed, simulation cost of 2, and even less, is achieved.
Unable to display preview. Download preview PDF.
- 1.R.J. Anderson and G.L. Miller. Optical Communication for Pointer Based Algorithms. Technical Report CRI-88-14, Computer Science Department, University of Southern California, LA, 1988.Google Scholar
- 2.M. Geréb-Graus and T. Tsantilas. Efficient Optical Communication in Parallel Computers. In SPAA'92, 4th Annual Symposium on Parallel Algorithms and Architectures, San Diego, California, pages 41–48, June 1992.Google Scholar
- 3.L.A. Goldberg, Y. Matias, and S. Rao. An Optical Simulation of Shared Memory. In SPAA'94, 6th Annual Symposium on Parallel Algorithms and Architectures, Cape May, New Jersey, pages 257–267, June 1994.Google Scholar
- 4.F. Meyer auf def Heide. Hashing Strategies for Simulating Shared Memory on Distributed Memory Machines. In F. Meyer auf der Heide, B. Monien, and A.L. Rosenberg, editors, Proc. of Parallel Architectures and Their Efficient Use, First Heinz Nixdorf Symposium, LNCS 678, pages 20–29. Springer-Verlag, 1992.Google Scholar
- 5.L.G. Valiant. General Purpose Parallel Architectures. In Algorithms and Complexity, Handbook of Theoretical Computer Science, volume A, pages 943–971, 1990.Google Scholar