Where is the majority of compiling computations done? Is the larger percentage of computation spent moving it from source to bytecode or bytecode to machine code?
I don't know the answer but I guess it will be in the source -> bytecode transformation. The semantics are richer and more complex. The translation from bytecode to machine code should be much more straightforward, by definition.
That was my assumption too, but I guess I'm looking for ballpark percentages on both? It must mostly (maybe 90-99%) be the source to bytecode, since JIT performance is pretty good (see java, python).
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.