A Simulation Approach to Evaluate the Time Cost of Parallel Computations

AbstractIn this paper, the execution time cost of a parallel computation in a shared memory environment is defined as a function of the input, the algorithm, the data structure, the processor speed, the number of processors, the processing power allocation and the communication between processors. The computation structure model is modified to describe the impact of these seven factors on the computation's execution time cost. A simulation technique is developed to analyze and evaluate the computation's execution time. A software tool, TCSS (Time Cost Simulation System) that implements the simulation technique is also described