Better Defunctionalization through Lambda Set Specialization

Higher-order functions pose a challenge for both static program analyses and optimizing compilers. To simplify the analysis compilation of languages with higher-order functions, rich body prior work has proposed variety defunctionalization techniques, which can eliminate from by transforming to semantically-equivalent first-order representation. Several modern take this step further, specializing respect on they operate, in turn allowing compilers generate more efficient code. However, existing techniques restrict how function values may be used, forcing implementations fall back costly dynamic alternatives. We propose lambda set specialization (LSS), first technique imposes no restrictions used. formulate LSS terms polymorphic type system tracks flow through program, use recast their arguments as form monomorphization. show that our admits simple tractable inference algorithm, give formalization fully-mechanized proof Isabelle/HOL assistant showing soundness completeness algorithm system. benefits LSS, we evaluate its impact run time performance code generated MLton compiler Standard ML, OCaml compiler, new Morphic functional programming language. find pre-processing achieves speedups up 6.85x under MLton, 3.45x OCaml, 78.93x Morphic.