Linked by Thom Holwerda on Mon 5th Nov 2012 23:40 UTC
Apple Another Apple-to-switch-Macs-to-ARM post. "Apple engineers have grown confident that the chip designs used for its mobile devices will one day be powerful enough to run its desktops and laptops, said three people with knowledge of the work, who asked to remain anonymous because the plans are confidential. Apple began using Intel chips for Macs in 2005." No idea when Apple will make the switch, but they will do it. I'm thinking 5-10 year timeframe.
Permalink for comment 541109
To read all comments associated with this story, please click here.
Not a slam dunk...
by Panajev on Tue 6th Nov 2012 08:42 UTC
Member since:

When Apple switched from PowerPC to Intel x86 cores, they did it because out of pragmatism. Since they cannot rely on first party apps alone, they need the performance of third party apps (which are seldom optimized down to the metal for all their target platforms) so they need to have at least comparable performance to PC's sold at that time.

Doing so with their own CPU designs, with the PowerPC chips they were developing together with IBM, would have cost quite a lot of money and the cost would have spread over far less processors than Intel ships regularly each quarter. Going to Intel was, long term, a very good move for Apple and a pragmatic one to boot.

If they can keep their iOS based devices selling as well as they do now or better, the possibility to switch from x86 to ARM would be there IMHO only if they could do it with a scale up strategy: invest in micro-architecture enhancements, use few cores on iOS devices, and scale up to very large number of cores on Desktops and Laptops. Still, they would have to solve the problem of keeping up single thread performance with Intel and help developers achieve at least similar performance to Desktop CPU designs by Intel without jumping through hoops.

This is not so much a HW challenge, but a very delicate balance of SW and HW efforts to make it happen. It might be why apple has been investing lots of resources on the LLVM project and initiatives such as GCD, but much more needs to be done.

Reply Score: 3