Research in the field of parallel computer architectures andparallel algorithms has been very successful in recentyears, and further progress isto be expected. On the otherhand, the question of basic principles of the architectureof universal parallel computers and their realizations isstill wide open. The answer to this question must beregarded as mostimportant for the further development ofparallel computing and especially for user acceptance.The First Heinz Nixdorf Symposium brought together leadingexperts in the field of parallel computing and itsapplications to discuss the state of the art, promisingdirections of research, and future perspectives. It was thefirst in a series of Heinz Nixdorf Symposia,intended tocover varying subjects from the research spectrum of theHeinz Nixdorf Institute of the University of Paderborn.This volume presents the proceedings of the symposium, whichwas held in Paderborn in November 1992. The contributionsare grouped into four parts: parallel computation models andsimulations, existing parallel machines, communication andprogramming paradigms, and parallel algorithms.
A combining mechanism for parallel computers.- A case for the PRAM as a standard programmer's model.- Hashing strategies for simulating shared memory on distributed memory machines.- Better parallel architectures via emulations.- Relations between models of parallel abstract machines.- Frontiers of parallel computing.- Cost effectiveness of data flow machines and vector processors.- The networks of the connection machine CM-5.- Massively parallel computing: Data distribution and communication.- A realizable efficient parallel architecture.- Hardware support for collective communication operations.- Programmable active memories: a performance assessment.- The role of randomness in the design of interconnection networks.- Chaos router: Finally, a practical adaptive router?.- An experimental study of wormhole routing in parallel computers.- Three non conventional paradigms of parallel computation.- Scalable parallel computers and scalable parallel codes: From theory to practice.- A framework for analyzing locality and portability issues in parallel computing.- Optimal implementation of general divide- and-conquer on the hypercube and related networks.- Parallel implementations of graph embeddings.- On sorting by prefix reversals and the diameter of pancake networks.
Series: Lecture Notes in Computer Science
Number Of Pages: 236
Published: 29th June 1993
Publisher: Springer-Verlag Berlin and Heidelberg Gmbh & Co. Kg
Country of Publication: DE
Dimensions (cm): 23.39 x 15.6
Weight (kg): 0.35