I’m genuinely curious as to what you’re asking. My background is in drivers (GPU), compilers, and I’m (currently) embedded in a HW group—I’m having a difficult

I’m genuinely curious as to what you’re asking. My background is in drivers (GPU), compilers, and I’m (currently) embedded in a HW group—I’m having a difficult time imagining why ARM would be at a disadvantage for compilation. I suppose older parts may not have had efficient byte-aligned memory fetch; or, perhaps you’re thinking of really old/low-power designs that aren’t out-of-order or superscalar? Either way, the big parts have all the bells-and-whistles x86-64 does; maybe with a bit of an instruction fetch disadvantage (THUMB doesn’t quite make up for the lack of a high quality variable-length CISC).