Running reg2mem at the end sounds like a great idea to get going quickly and see if the rest of the project is feasible. If the unnecessary reloading is a performance drag, but there are no other blocking issues, then writing a new "spill" pass would be justified. I did not realize that the common codegen was not used, but that makes sense given that "real" register allocation is not needed. These are exactly the sort of pointers I was looking for. Thanks!
-- You received this message because you are subscribed to the Google Groups "emscripten-discuss" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/d/optout.
