In this paper, we synthesize nested For-loops with partitions on the innermost loop. First, we present a nonlinear transformation algorithm to exploit the parallelism of the For-loops. By the mapping of nonlinear transformation, iterations of For-loops can be executed in a parallel form. The proposed algorithm is useful in exploiting the parallelism of For-loops with one or more partitions on the innermost loop. Then, we also design algorithms to partition and map the nested For-loops onto the fixed size systolic arrays. Based on the time and space mapping schemes, all the iterations of For-loops can be correctly executed on the array processors in a parallel form.
關聯:
IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS
