4 ms·
Maybe it’s because Apple/ARM can ditch all legacy x86 stuff so the chips only contain instructions relevant for today, saving space for more transistors to do t
by tobyhinloopen 6y ago
Maybe it’s because Apple/ARM can ditch all legacy x86 stuff so the chips only contain instructions relevant for today, saving space for more transistors to do these relevant instructions or just to keep the chip smaller and more efficient
- hvidgaard 6y agoUnderneath the x86/x64 ISA is a really different architecture. So to reach the backwards compatibility it needs a x86/x64 decoder and that is not a significant portion of the silicon as far as I know. Those instructions might be somewhat inefficient, but this is not really an issue when comparing, any ARM chip emulating will have to tackle the same problem. If Apples chip does not need this backwards compatibility, they essentially have to do it in software when emulating. They do this with a combination of recompilation at install time and some emulation at runtime. I suspect the compilation step is one of the main contributors to this boost in speed and efficiency. Windows could do the same, optimize the executable to better fit the underlying chip.