Abstract
The problem of realizing an idealized parallel architecture on a (possibly fault-laden) physical architecture is studied. Our formulation performs the mapping in the light of the algorithm that one wants to implement on the idealized architecture. A version of the mapping algorithm suggested by the DIOGENES methodology for designing fault-tolerant VLSI processor arrays is settled definitely. Two quality metrics for mappings are considered, the first embodying an idealized notion of average delay, which relates to power consumption, and the second being the length of the longest run of wire. For the average-delay measure, four algorithms that optimally assign the m vertices of the embedded graph to the n fault-free processors that have been fabricated are presented. The most general algorithm makes no assumptions about the structure of the array or the physical format of the processors; it runs in time O(m · (n - m)2). The other algorithms assume that the processors are laid out in such a way that interprocessor distances obey the triangle equality; they run in times ranging from time O(max{m, n - m} · log min {m, n - m}) for certain array structures, including linear arrays, to time O(max{m, n - m}) for a narrow class of array structures, including pyramid arrays. For the max-wire-run cost measure, it is shown that the problem of finding cost-optimal vertex-to-processor assignments is NP-complete. However, an algorithm is presented that yields, in time O(m · (n - m)2), vertex-to-processor assignments that are within a factor of 3 of optimal (they are optimal when the input graph-embedding is outplanar). This algorithm can easily be converted to one that yields, in time O(m · (n - m)3), vertex-to-processor assignments that are within a factor of 2 of optimal. Finally, an algorithm that yields optimal assignments when the interprocessor distances obey the triangle equality is presented; this algorithm operates in time O(m · (n - m) · log(m · (n - m)) · log M), where M is the largest interprocessor distance.
- 1 AHO, A. V., HOPCROFT, J. E., AND ULLMAN, J. D. The Design and Analysis of Computer Algorithms. Addison-Wesley, Reading, Mass., 1974. Google Scholar
- 2 ALON, N., AND CHUNG, F. R.K. Explicit constructions of linear-sized fault-tolerant networks. Typescript, Massachusetts Institute of Technology, Cambridge, Mass., 1985.Google Scholar
- 3 BHATT, S. N., CHUNG, F. R. K., LEIGHTON, F. T., AND ROSENBERG, A.L. Optimal simulations of tree machines. In Proceedings of the 27th IEEE Symposium on Foundations of Computer Science. IEEE, New York, 1986, pp. 274-282.Google Scholar
- 4 CHUNG, F. R. K., LEIGHTON, F. T., AND ROSENBERG, A.L. DIOGENESmA methodology for designing fault-tolerant processor arrays. In Proceedings of the 13th International Conference on Fault-Tolerant Computing. IEEE, New York, 1983, pp. 26-32.Google Scholar
- 5 CHUNG, F. R. K., LEIGHTON, F. T., AND ROSENBERG, A.L. Embedding graphs in books: A layout problem with applications to VLSI design. SIAM ~ Algebraic and Discrete Methods 8 (1987), 33-58. Google Scholar
- 6 GAREY, M. R., AND JOHNSON, D.S. Computers and Intractability. Freeman, San Francisco, Calif., 1979.Google Scholar
- 7 GREENBERG, O. S., HEATH, L. S., AND ROSENBERG, A.L. Optimal embeddings of FFT graphs in hypercubes. Tech. Rep. 88-23. Univ. of Massachusetts, Amherst, Mass., 1988. Google Scholar
- 8 GREENE, J. W., AND EL GAMAL, A. Configuration of VLSI arrays in the presence of defects. J. ACM 31, 4 (Oct. 1984), 694-717. Google Scholar
- 9 HASTAD, J., LEIGHTON, F. T., AND NEWMAN, M. Reconfiguring a hypercube in the presence of faults. Typescript, Massachusetts Institute of Technology, Cambridge, Mass., 1986.Google Scholar
- 10 HEATH, L. S., ROSENBERG, A. L., AND SMITH, B.T. The DIOGENES design methodology: From embedding to layout. Tech. Rep. 87-17. Univ. of Massachusetts, Amherst, Mass., 1987. Google Scholar
- 11 LEE, S.-Y., AND AGGARWAL, J. K. A mapping strategy for parallel processing. IEEE Trans. Comput. C-36 (1987), 433-442. Google Scholar
- 12 LEIGHTON, F. T., AND LEISERSON, C.E. Wafer-scale integration of systolic arrays. IEEE Trans. Comput. C-34 (1985), 448-461.Google Scholar
- 13 ROSENaERG, A.L. The Diogenes approach to testable fault-tolerant arrays of processors. IEEE Trans. Comput. C-32 (1983), 902-910.Google Scholar
- 14 ROSENBERG, A.L. On designing fault-tolerant VLSI processor arrays. In Advances in Computing Research, vol. 2. F. P. Preparata, Ed. JAI Press, Greenwich, Conn., 1984, pp. 181-204.Google Scholar
- 15 TARJAN, R. E., AND VAN LEEUWEN, .{. Worst-case analysis of set union algorithms. J. ACM 31, 2 (Apr. 1984), 245-281. Google Scholar
Index Terms
- The physical mapping problem for parallel architectures
Recommendations
Parallel Algorithms for the Longest Common Subsequence Problem
A subsequence of a given string is any string obtained by deleting none or some symbolsfrom the given string. A longest common subsequence (LCS) of two strings is a commonsubsequence of both that is as long as any other common subsequences. The problem ...
Comments