Fast and incremental computation of weak control closure
Control dependence is a fundamental concept used in many program analysis techniques such as program slicing, program debugging, program parallelization, and detecting security leaks. Since the introduction of this concept in the late eighties, numerous definitions of control dependencies and their computation methods have been developed. The later definitions are progressively more generalized covering a wide spectrum of modern programming language constructs. The most generalized concepts are the weak and strong control closure (WCC and SCC) that capture the nontermination (in)sensitive control dependencies of a given program. In this paper, we have developed a novel method to compute WCC incrementally. Any client application of WCC such as program slicing requires computing the WCC repeatedly in a fixpoint computation. An incremental algorithm to compute WCC will improve the performance of the client application significantly. We have provided the proof of correctness and the theoretical worst-case complexity of our algorithms. We have performed an experimental evaluation on well-known benchmarks, and our experiments reveal that we have significantly improved the practical efficiency in computing WCC incrementally. We have obtained an average speedup of 31.03 in all benchmarks and a maximum speedup of 35.29 than the best state-of-the-art algorithm computing WCC.
Wed 7 DecDisplayed time zone: Auckland, Wellington change
10:30 - 12:00 | |||
10:30 30mTalk | Semantic Foundations for Cost Analysis of Pipeline-Optimized ProgramsVirtual SAS Solène Mirliaz ENS Rennes / IRISA / Inria, David Pichardie Meta, Gilles Barthe MPI-SP, Germany / IMDEA Software Institute, Spain, Adrien Koutsos INRIA Paris, Peter Schwabe Max Planck Institute for Security and Privacy | ||
11:00 30mTalk | Principles of Staged Static+Dynamic Partial Analysis SAS | ||
11:30 30mTalk | Fast and incremental computation of weak control closure SAS Abu Naser Masud Malardalen University |