Optimal partitioning of random programs across two processors

B. Indurkhya et al. (1986) concluded that the optimal partitioning of a homogeneous random program over a homogeneous distributed system either assigns all modules to a single processor or distributes the modules as evenly as possible among all processors. Their analysis rests heavily on the approximation that equates the expected maximum of a set of independent random variables with the set's maximum expectation. The author strengthens this result by providing an approximation-free proof of this result for two processors under general conditions on the module execution time distribution. It is found that additional rigor leads to a different characterization of the optimality points. The author also shows that under a rigorous analysis one is led to different conclusions in the general P-processor case than those reached using B. Indurkhya et al.'s approximation.<
>