The Community for Technology Leaders
Green Image
<p><b>Abstract</b>—Dynamic redistribution of arrays is required very often in programs on distributed memory parallel computers. This paper presents efficient algorithms for redistribution between different cyclic(<it>k</it>) distributions, as defined in High Performance Fortran. We first propose special optimized algorithms for a cyclic(<it>x</it>) to cyclic(<it>y</it>) redistribution when <it>x</it> is a multiple of <it>y</it>, or <it>y</it> is a multiple of <it>x</it>. We then propose two algorithms, called the GCD method and the LCM method, for the general cyclic(<it>x</it>) to cyclic(<it>y</it>) redistribution when there is no particular relation between <it>x</it> and <it>y</it>. We have implemented these algorithms on the Intel Touchstone Delta, and find that they perform well for different array sizes and number of processors.</p>
Array redistribution, distributed-memory computers, High Performance Fortran (HPF), data distribution, runtime support.
Alok Choudhary, Rajeev Thakur, J. Ramanujam, "Efficient Algorithms for Array Redistribution", IEEE Transactions on Parallel & Distributed Systems, vol. 7, no. , pp. 587-594, June 1996, doi:10.1109/71.506697
88 ms
(Ver 3.3 (11022016))