Optimization algorithm for parallel demand—driven computation of functional programs
Satoshi Ono, Naohisa Takahashi · Systems and Computers in Japan · 1988
Abstract This paper proposes a new optimization algorithm for the demand‐driven computation suited to the parallel processing of the functional program. The optimization is performed for three points: the preevaluation of requisite parameters; the efficiency of demand propagation; and the analysis of the partialvalue dependent attribute set. To reduce the computation required in the optimization, a hierarchical approach is employed, which is composed of the interfunctional analysis to process the whole program in parallel, and the intrafunctional analysis for each function definition as a separate phase. Using the proposed method, the successive function in the discrete domain can be calculated efficiently based on the fixed‐point computation rule, by providing a certain extension of the data‐flow machine executing in parallel the data‐driven computation.