It seems to me that these calls would likely get reordered using link time optimization. Even then, unless I’m missing something and assuming they have to be in that order because of side effects, it feels like a very unusual case. Do you feel there are more cases less obvious to the eyes, where the lower end belts suffer from this saturation?
It is good to see some code generation getting done though. When can we expect being able to play with the toolchain?
Reply To: Belt saturation in short belts Noit