An algorithm used to eliminate intra-iteration data dependencies

Data dependencies among tasks are inevitable and have an adverse effect on the parallelism of an application. Therefore, this paper addresses the problem of setting appropriate delays on each edge of a task graph representing an application to eliminate intra-iteration data dependencies so as to produce a prologue and epilogue as small as possible. Based on the principle of the retiming technique, we propose an algorithm called NNP (number of nodes along a path) to determine the minimum retiming value of each node according to the number of nodes along the path with the maximum number of nodes from its immediate successor nodes to leaf nodes. Computation time can be decreased by the above transformation due to eliminating intra-iteration data dependencies. We experiment with random task graphs on 2, 3 and 4 processor cores, respectively. The experimental results show that the algorithm can obtain substantial scheduling length reduction.