Adaptive quadrature algorithms for the ILLIAC IV
- 28 Downloads
Two classes of adaptive quadrature algorithms for use on the ILLIAC IV are described, one designed for high processor utilization, the other for high processing speed. The results of simulation tests comparing algorithms from these two classes are summarized and commented on. The results indicate that a more truly adaptive algorithm without consistently high processor utilization is faster than one designed specifically for high utilization. For a better algorithm, the speedup over sequential processing ranges from 1.1 to 46.9 (average = 15.5) for 14 examples; this is considerably less than the maximum possible speedup of 64 but is still substantial.
Key wordsNumerical integration adaptive quadrature parallel processing ILLIAC IV
Unable to display preview. Download preview PDF.
- 1.G. H. Barnes, “The ILLIAC IV computer,”IEEE Trans. Comput. 17:746–757 (1968).Google Scholar
- 2.W. J. Bouknight et al., “The ILLIAC IV system,”Proc. IEEE 60:369–388 (1972).Google Scholar
- 3.Burroughs Corporation, “ILLIAC IV Systems Characteristics and Programming Manual,” Defense, Space and Special Systems Group, IL4-PM1, revised May 16, 1972.Google Scholar
- 4.R. L. Davis, “The ILLIAC IV processing element,”IEEE Trans. Comput. 18:800–816 (1969).Google Scholar
- 5.D. J. Kuck, “ILLIAC IV software and application programming,”IEEE Trans. Comput. 17:758–770 (1968).Google Scholar
- 6.D. H. Lawrie et al., “GLYPNIR: A programming language for ILLIAC IV,”Comm. Assoc. Comput. Mach. 18:157–164 (1975).Google Scholar
- 7.J. M. Lemme, “Speedup in Parallel Algorithms for Adaptive Quadrature,” Ph.D. Thesis, Purdue University (1976).Google Scholar
- 8.J. R. Rice, “A metalgorithm for adaptive quadrature,”J. Assoc. Comput. Mach. 22:61–82 (1975).Google Scholar