Partitioning the iteration space can significantly offect the execution time of a loop. In this paper, wi propuse an improvement over previous partitioning methods distributed memory systems, partitioning each loop separately does not guarantee an efficient execution of the code becaouse of across loop data depence. As a results, a global iteration space is formed so that loops in a program are considered whten partitioning eht glbobal space. In addision, a new and general form of expressing data dependence called hyperplane depence is introduced an used in the partitioning. It is dependence shore source and destination are busspaces (of any dimension of the global iteration space.