Code generation for complex subscripts in data-parallel programs

Data parallel languages like High Performance Fortran, demand efficient compile and run-time techniques for tasks such as address generation. Array references with arbitrary affine subscripts can make the task of compilers for such languages highly involved. This paper deals with the efficient address generation in programs with array references having two types of commonly encountered affine references, namely coupled subscripts and subscripts containing multiple induction variables (MIVs). Methods discussed in the paper utilize the repetitive pattern of the memory accesses. In the case of MIV, we address this issue by presenting runtime techniques which enumerate the set of addresses in lexicographic order. Our approach to the problem incorporates a general approach of computing in O(k) time, the start element on a processor for a given global start element. Several methods are proposed and evaluated here for generating the access sequences for MIV based on problem parameters. With coupled subscripts, we present two construction techniques, namely searching and hashing which minimize the time needed to construct the tables. Extensive experiments were conducted and the results were then compared with [8] to indicate the efficiency of our approach.