The A12 is clearly at parity now vs. Intel's existing mobile offerings, probably somewhat ahead given the long delays with 10nm.
The question upthread is whether or not switching architectures makes financial sense, not whether it's a (mild) technical win.
Switching to their own chips cuts Intel out of the loop, but as far as business risk that simply replaces one single source manufacturer with another (TSMC).
It probably saves money per-part, which is good. But then Apple is still drowning in cash and immediate term savings really aren't much of a motivator.
> By 2021 Intel can probably deliver a 1.1-1.2x speed up tops for Macbook airs, while Apple can probably deliver a 2x speed up for 2021 with the A15 and another one for 2023 with the A16.
That's going to need some citation. Moore's law is ending for everyone, not just Intel. TSMC has pulled ahead of Intel (and Samsung has caught up) for sure, but progress is slowing. That kind of scaling just isn't going to happen for anyone.
The question upthread is whether or not switching architectures makes financial sense, not whether it's a (mild) technical win.
Switching to their own chips cuts Intel out of the loop, but as far as business risk that simply replaces one single source manufacturer with another (TSMC).
It probably saves money per-part, which is good. But then Apple is still drowning in cash and immediate term savings really aren't much of a motivator.
> By 2021 Intel can probably deliver a 1.1-1.2x speed up tops for Macbook airs, while Apple can probably deliver a 2x speed up for 2021 with the A15 and another one for 2023 with the A16.
That's going to need some citation. Moore's law is ending for everyone, not just Intel. TSMC has pulled ahead of Intel (and Samsung has caught up) for sure, but progress is slowing. That kind of scaling just isn't going to happen for anyone.