Efficient dependence analysis for java arrays

Abstract

This paper studies dependence analysis for Java arrays, emphasizing efficient solutions that avoid a large compile-time overhead. We present a new approach for dependence analysis based on sparse congruence partitioning representations in SSA form. Since arrays in Java are dynamically allocated, our approach takes pointer-induced aliasing of array objects into account in conjunction with analysis of index values. We present experimental results to evaluate the effectiveness of our approach, and outline directions for further improvements.