I watched a good video on Delimited Continuations from Alexis King last night. She explains them very well both semantically and with some examples. I'm glad to see that native support for them has been introduced to Haskell.
In figuring out what approach to take for my full program analysis for algebraic effects, I'm looking into a few things:
One is this paper Functional derivation of a virtual machine for delimited continuations, which I think was recommended to me by Kimball when I was first looking into algebraic effects, and first learning about control flow analyses. I'll probably also look into both the typed directed continuation passing style transformation that was originally implemented for Koka, as well as the more recent Generalized evidence passing for effect handlers: efficient compilation of effect handlers to C.
As far as progress today on figuring out some benchmarks for how to improve Koka's performance, I've added some basic benchmarks, but there is still more research to go, to really show the difference. I wish there were more real-world example programs I could benchmark the changes on. I'm sure I'll be contributing to that due to my interest in Koka. I mostly spent time on the benchmarking scripts, but in doing so I also added / exposed a few methods for dealing with getting file information.