Program analysis for reversible languages

ABSTRACTReversible computing is a technique to “let computations run forwards and backwards” and thus extends the traditional model of computation. As an example, consider a function to compute the factorial of a given number, say 5, yielding 120. Running this program backwards inverses the function by taking 120 as an input and giving 5 as result. Reversible programming languages allow the creation of programs that can be executed backwards as well as forwards and have been a focus of research over the last decade mostly due to the work of Glück, Yokoyama, Mogensen, and many others. In this paper, we report our recent activities to perform program analysis for reversible static-single-assignment form (RSSA) and using them to perform local and global optimizations. This work is based on our compiler translating from the reversible language Janus to RSSA. As far as we know, this is the first compiler from Janus to RSSA, and no results on optimization of reversible intermediate code or programming languages are known to us either. Optimization techniques in “traditional” compilers are always based on the understanding that programs are executed forwards - in reversible languages that assumption is no longer true and program analysis becomes much more difficult. Our first results on applying our analysis methods for common-subexpression-elimination and constant propagation are nevertheless promising and have been implemented successfully.