For papers, does it make sense to put the abstract in the description? Perhaps not; it’s a few paragraphs. Anyway, the part that jumped out at me was:
Furthermore, we present a novel program transformation to eliminate the overhead caused by higher-order functions. The main component of this transformation is lambda mangling: an important transformation primitive in Thorin. We demonstrate that lambda mangling subsumes many classic program transformations like tail-recursion elimination, loop unrolling or (partial) inlining.
If this and their performance claims are both true, I imagine a lot of compiler authors are going to be bemoaning how very difficult it is to switch to a better IR. The paper is short, but deep, and requires more time than I can give it while at work. I’ll try to add further commentary tonight, and I’d be fascinated to hear everyone else’s thoughts.