Mapping strategies in message based multiprocessor systems
Machines with distributed memory have the mapping problem — assigning processes to processors. In this paper we define the mapping problem as an optimization problem and discuss the question, how far is an optimum solution from an average or random solution.
The term robustness is introduced and explained in detail with two examples, the SUPRENUM and the Hypercube architecture. For the SUPRENUM architecture we show that a simple mapping strategy (optimal clustering of the processes) gives almost as good results as the optimal mapping. Optimal mapping is more important for the Hypercube architecture.
Mapping strategies are difficult to apply for inhomogeneous networks. For this networks adaptive routing seems promising.
KeywordsCommunication Cost Mapping Strategy Optimal Partitioning Mapping Problem Multiprocessor System
Unable to display preview. Download preview PDF.
- [ArRa80]R. K. Arora, S. P. Rana, Heuristic Algorithms for Process Assignment in Distributed Computing Systems, Information Processing Letters 11 (12/80) 199–203Google Scholar
- [BGM86]P. M. Behr, W. K. Giloi, H. Mühlenbein, SUPRENUM: The German Supercomputer Project — Rationale and Concepts, IEEE Internat. Conference on Parallel Processing (1986)Google Scholar
- [Bokh81]S. H. Bokhari, On the Mapping Problem, IEEE Transaction on Computers C-30 No. 3 (1981) 207–214Google Scholar
- [BrSc86]J. E. Brandenburg, D. E. Scott, Embedding of Communication Trees and Grids into Hypercubes, Intel iPSC User Group No. 1 (1986)Google Scholar
- [CHLE80]W. W. Chu, L. J. Holloway, M. T. Lan, K. Efe, Task Allocation in Distributed Data Processing, Computer (11/80) 57–69Google Scholar
- [DaSe86]W. J. Dally, Ch. L. Seitz, The Torus Routing Chip, Distributed Computing 1 (1986) 187–196Google Scholar
- [EBP86]A. K. Ezzat, R. D. Bergeron, J. L. Pokoski, Task Allocation Heuristics for Distributed Computing Systems, IEEE Int. Conf. on Distributed Systems (1986)Google Scholar
- [HHR86]C. E. Houstis, E. N. Houstis, J. R. Rice, Partitioning PDE Computations: Methods and Performance Evaluation, Report Purdue University (1986)Google Scholar
- [HoMü86]H. C. Hoppe, H. Mühlenbein, Parallel Adaptive Full-Multigrid-Methods on Messagebased Multiprocessors, Parallel Computing 3 (1986) 269–287Google Scholar
- [Hock85]R. W. Hockney, MIMD computing in the USA — 1984, Parallel Computing 2 (1985) 119–136Google Scholar
- [HMSCP86]J. P. Hages, T. Mudge, Q. M. Stow, S. Colley, J. Palmer, A Microprocessor based Hypercube Supercomputer, IEEE Micro, Vol. 6, No. 5 (1986)Google Scholar
- [MBC83]M. A. Marsan, G. Balbo, G. Conte, Comparative performance analysis of single bus multiprocessor architectures, IEEE Trans. Computers, C-31 (1983) 1179–1191Google Scholar
- [MGK87]H. Mühlenbein, M. Gorges-Schleuter, O. Krämer, New Solutions to the Mapping Problem of Parallel Systems — The Evolution Approach, Parallel Computing, to be published (1987)Google Scholar
- [MüKr87]H. Mühlenbein, O. Krämer, Parallel Solutions of the Graph Partitioning Problem, to be publishedGoogle Scholar
- [MKLMS87]H. Mühlenbein, O. Krämer, F. Limburger, M. Mevenkamp, S. Streitz, Design and Rationale for MUPPET — A Programming Environment for Message-Based Multiprocessors, Journal of Parallel and Distr. Proc., to be publishedGoogle Scholar
- [NiSa86]D. Nicol, J. Saltz, Dynamic Remapping of Parallel Computations with Varying Resource Demands, ICASE Report 86–45, Nasa Langley (1986)Google Scholar
- [SAKA86]C. Saito, H. Amano, T. Kudoh, H. Aiso, An adaptable cluster structure of (SM) 2 -II, CONPAR 86, Aachen (9/86) 53–60Google Scholar
- [SaSc85]Y. Saad, M. H. Schultz, Topological Properties of Hypercubes, Research Report RR-389 Yale University (6/85)Google Scholar
- [ScGa85]K. Schwans, C. Gaimon, Automatic Resource allocation for the CM* Multiprocessor, 5th Distr. Comp. Conf. (1985)Google Scholar
- [ScJo85]K. Schwans, A. K. Jones, Specifying Resource Allocation for Parallel Programs on the CM* Multiprocessor, Ohio State University, OSU-CISRC-TR-85-10 (1985)Google Scholar
- [Snyd82]L. Snyder, Introduction to the Configurable, Highly Parallel Computer, Computer 15 (2/82) 47–56Google Scholar
- [Ullm84a]J. D. Ullman, Computational aspects of VLSI, Computer Science Press (1984)Google Scholar
- [Ullm84b]J. D. Ullman, Some thoughts about Supercomputer Organization, Proc. IEEE Compcon, San Francisco (1984) 424–431Google Scholar