Issue No. 03 - July (1990 vol. 1)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/71.80155
<p>An accurate and computationally efficient method for predicting the performance of a class of parallel computations running on concurrent systems is described. A parallel computation is modeled as a task system with precedence relationships expressed as a series-parallel directed acyclic graph. Resources in a concurrent system are modeled as service centers in a queuing network model. Using these two models as inputs, the method outputs predictions of expected execution time of the parallel computation and the concurrent system utilization. The method is validated against both detailed simulation and actual execution on a commercial multiprocessor. Using 100 test cases, the average error of the prediction when compared to simulation statistics is 1.7%, with a standard deviation of 1.5%; the maximum error is about 10%.</p>
Index Termsperformance prediction; parallel computations; concurrent systems; task system; series-parallel directed acyclic graph; service centers; queuing network model; simulation; commercial multiprocessor; directed graphs; parallel processing; performance evaluation; queueing theory
S. Lundstrom and V. Mak, "Predicting Performance of Parallel Computations," in IEEE Transactions on Parallel & Distributed Systems, vol. 1, no. , pp. 257-270, 1990.