Efficient program analysis using DJ graphs
Year: 1995
Type: dissertation
Abstract: Program analysis is a process of estimating properties of a program statically. Program analyses have many applications, including compiler optimizations, software maintenance and testing, and program verification. In this dissertation we present a new framework for efficient program analysis. At the heart of our approach is a new program representation called the DJ Graph. Using DJ graphs we present several new algorithms for solving problems encountered in program analysis. The problems that we have solved range from a simple loop identification problem to sophisticated exhaustive and incremental data flow analysis, including the construction of Sparse Evaluation Graphs. The algorithms presented here are simple, more general, and/or more efficient than existing methods for solving similar problems. To study the effectiveness of our algorithms on real programs we implemented many of them, and experimented on a number of FORTRAN procedures taken from standard benchmark suites. Our results indicate that the algorithms presented here perform well in practice.
DOI: 10.82308/44818
Author: Sreedhar, Vugranam C.
Institution: McGill University
Language: English
Subfield: Software
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W1580051730