POINTER OPTIMIZATION USING SSA BASED INTERMEDIATE REPRESENTATION FOR OPTIMIZING COMPILERS

Abstract

In optimizing compilers, most of the optimizing algorithms are applied on intermediate representation. Optimizing algorithms such as constant propagation can be applied on intermediate code using data flow analysis which is expensive regarding compilation time. Also for pointer variables, these algorithms cannot be applied directly on pointers as pointers are difficult to analyze. In this dissertation entitled "POINTER OPTIMIZATION USING SSA BASED INTERMEDIATE REPRESENTATION FOR OPTIMIZING COMPILERS", new approach for applying optimization algorithms such as constant propagation on pointer variables has been proposed. To avoid data flow analysis, static single assignment (SSA) form is used for intermediate representation and for pointers, alias classes are used. So using SSA and each alias class as a single variable, constant propagation algorithm can be applied on pointer variables very efficiently reducing the number of instructions in a program.