Heap Dependence Analysis for Sequential Programs

In this paper we demonstrate a novel intra-procedural technique for detecting heap dependences in sequential programs that use recursive data structures. The novelty of our technique lies in the way we compute, for each statement, abstract heap access paths that approximate the locations accessed by the statement, and the way we convert these paths into equations that can be solved using traditional tests, e.g. GCD test, Banerjee test and Lamport test. The dependence test also uses a field sensitive shape analysis to detect dependences among heap locations arising out of sharing within the data structure. In presence of loops, the technique can be used to discover loop dependences. i.e. the dependence among two different iterations of the same loop. This information can be used by a parallelizing compiler to transform sequential input program for better parallel execution.