Efficient simulations among several models of parallel computers (extended abstract)
We consider parallel computers (PC's) with fixed communication network and deal with the question, how fast can we simulate PC's with n processors whose communication network has unbounded degree (unfair PC's) by PC's where this degree is bounded (fair PC's). An important class of unfair PC's is thatone of PC's with predictable communication (pred.com.). In such a PC each processor can compute in 0(t) steps the sequence of processors it wants to communicate with during the next t steps. A famous example of such PC's are cubes initialized with ascend-and descend programs as introduced by Preparata and Vuillemin in . They could simulate such unfair PC's with pred.com. with constant time loss using only as many processors as the cube has. We generalize this result by presenting a fair PC which can simulate each unfair PC with pred.com., n processors, and 0(log(n)) storage locations per processor with constant time loss using 0(n1+ɛ) processors for an arbitrary ɛ>0.
KeywordsInitialization Time Parallel Computer Main Root Time Loss Efficient Simulation
Unable to display preview. Download preview PDF.
- M.Ajtai, J.Komlos, E.Szemeredi: An 0(nlog(n)) Sorting Network, Proc. of the 15′th Annual ACM Symposium on Theory of Computing (1983), Boston, USA, 1–9.Google Scholar
- K. Batcher: Sorting Networks and their Applications, AFIPS Spring Joint Comp. Conf. 32 (1968), 307–314.Google Scholar
- Z. Galil, W.J. Paul: A General Purpose Parallel Computer, Journal of the ACM 30(2) (1983), 360–387.Google Scholar
- F. Meyer auf der Heide: Efficiency of Universal Parallel Computers, Acta Informatica 19 (1983) 269–296.Google Scholar
- F.P. Preparata, J. Vuillemin: The Cube-Connected Cycles: A Versatile Network for Parallel Computation, Communications of the ACM 24 (1981), 300–310.Google Scholar
- J.H.Reif, L.G.Valiant: A Logarithmic Time Sort for Linear Size Networks, 15′th Annual ACM Symposium on Theory of Computing (1983), Boston, USA, 10–16.Google Scholar
- E.Upfal: Efficient Schemes for Parallel Communication, Proc. of the ACM Symposium on Principles of Distributed Computing (1982), Ottawa, Canada.Google Scholar
- L.G.Valiant, G.J.Brebner: Universal Schemes for Parallel Communication, Proc. of the 13′th Annual ACM Symposium on Theory of Computing (1981), Milwaukee, USA, 263–267.Google Scholar
- A. Waksman: A Permutation Network, Journal of the ACM, 15(1) (1968), 159–163.Google Scholar
- F.Meyer auf der Heide: Efficient Simulations among Several Models of Parallel Computers, Interner Bericht des Fachbereichs Informatik der J.W.Goethe-Universität, Frankfurt, 2/83, to appear in SIAM J.on Comp.Google Scholar