Data dependence analysis of assembly code

Determination of data dependences is a task typically performed with high-l evel language source code in today's optimizing and parallelizing compilers . Very little work has been done in the field of data dependence analysis o n assembly language code, but this area will be of growing importance, e.g. , for increasing instruction-level parallelism. A central element of a data dependence analysis in this case is a method for memory reference disambig uation which decides whether two memory operations may access (or definitel y access) the same memory location. In this paper we describe a new approac h for the determination of data dependences in assembly code. Our method is based on a sophisticated algorithm for symbolic value propagation, and it can derive value-based dependences between memory operations instead of jus t address-based dependences. We have integrated our method into the SALTO s ystem for assembly language optimization. Experimental results show that ou r approach greatly improves the precision of the dependence analysis in man y cases.