Summary
We consider parallel computers (PC's) with fixed communication network and bounded degree. We deal with the following question: How efficiently can one PC, a so-called universal PC, simulate each PC with n processors? This question is asked in [1] where a universal PC with O(n) processors and time loss O(log(n)) is constructed. We improve this result in two ways by construction two universal PC's which many users can efficiently work with at the same time. The first has the same number of processors and the same time loss as that one above. The second has O(O 1 + ɛ) processors for an arbitrary ɛ>0 but only time loss O(loglog(n)). Finally we define three types of simulations the most general of which includes all known simulations. We prove non-linear time-processor tradeoffs for universal PC's associated with the above types.
Similar content being viewed by others
References
Aho, A.V., Hopcroft, J.E., Ullman, J.D.: The Design and Analysis of Computer Algorithms. Reading, MA: Addison-Wesley, 1974
Galil, Z., Paul, W.J.: A Theory of Complexity of Parallel Computation. Proc. 13th Ann. ACM Symp. Theory Comput Milwaukee, pp. 247–262, 1981
Paul, W.J.: Komplexitätstheorie. Stuttgart: Teubner Verlag, 1978
Preparata, F.P., Vuillemin, J.: The Cube-Connected Cycles: A Versatile Network for Parallel Computation. Proc. 20th Ann. IEEE Symp. Found. Comput. Sci., Puerto Rico, pp. 140–147, 1979
Waksman, A.: A Permutation Network. J. ACM 15, 159–163 (1968)
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
auf der Heide, F.M. Efficiency of universal parallel computers. Acta Informatica 19, 269–296 (1983). https://doi.org/10.1007/BF00265559
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF00265559