SymPas: Symbolic Program Slicing

Program slicing is a technique for simplifying programs by focusing on selected aspects of their behavior. Current mainstream static methods operate dependence graph PDG (program graph) or SDG (system graph), but these friendly representations may be bit expensive some users. In this paper we attempt to study light-weight approach program slicing, called Symbolic Slicing (SymPas), which works as dataflow analysis LLVM (low-level virtual machine). our SymPas approach, slices are stored in symbolic forms, not procedures being re-analyzed (cf. procedure summaries). Instead re-analyzing multiple times find its each callling context, calculate single slice can instantiated at call sites avoiding re-analysis; implemented with perform intermediate representation (IR). For comparison, systematically adapt IFDS (interprocedural finite distributive subset) and the SDG-based method (SDG-IFDS) statically IR programs. Evaluated open-source benchmark programs, backward shows factor-of-6 reduction time cost factor-of-4 space cost, compared SDG-IFDS, thus more efficient. addition, result that after studying from 66 ranging up 336 800 instructions size, highly size-scalable.