Pick your poison, optimise ScalaC, wait for Dotty, start from scratch? Scala compiler performance is an issue on large projects. Whilst Dotty is faster, it is also a significant amount of time away. Should we just wait, can we and should we improve ScalaC? Can we do better if we rewrite the compiler? We have learned a lot from the performance of DotC, from the optimisation work going on right now and from the exploratory work in Kentucky Mule. This talk show the current attacks on performance, the potential paths and explores both how we are doing and how far we can go.
Kentucky Mule or Kamikaze – Attacking Scala compiler performance