Comment by bloaf

The next questions in my head are:

A) Is there any performance penalty on the compiled result due to being optimized in separate pieces

B) If not, why isn't the optimizer smart enough to split them up internally and optimize in parallel?