The Community for Technology Leaders
Green Image
<p>A method for executing nested loops with constant loop-carried dependencies in parallelon message-passing multiprocessor systems to reduce communication overhead is presented. In the partitioning phase, the nested loop is divided into blocks that reduce the interblock communication, without regard to the machine topology. The execution ordering of the iterations is defined by a given time function based on L. Lamport's (1974) hyperplane method. The iterations are then partitioned into blocks so that the execution ordering is not disturbed, and the amount of interblock communication is minimized. In the mapping phase, the partitioned blocks are mapped onto a fixed-size multiprocessor system in such a manner that the blocks that have to exchange data frequently are allocated to the same processor or neighboring processors. A heuristic mapping algorithm for hypercube machines is proposed.</p>
Index Termsdata exchange; mapping; nested loops; constant loop-carried dependencies; parallel;message-passing multiprocessor systems; communication overhead; partitioning; blocks;interblock communication; execution ordering; iterations; time function; hyperplanemethod; fixed-size multiprocessor system; heuristic mapping algorithm; hypercubemachines; multiprogramming; parallel algorithms; parallel programming
J.P. Sheu, T.H. Thai, "Partitioning and Mapping Nested Loops on Multiprocessor Systems", IEEE Transactions on Parallel & Distributed Systems, vol. 2, no. , pp. 430-439, October 1991, doi:10.1109/71.97900
388 ms
(Ver 3.3 (11022016))