Apple’s Dual-Sourced A9 SoC ⇥ engadget.com
Another update on a story that broke last week, this time on the confirmation that Apple has been dual-sourcing their A9 processors from Samsung and TSMC. I didn’t write about it because it didn’t seem like that big of a deal: Apple has surely run their tests and found that they perform virtually identically, so it shouldn’t matter what’s in your phone.
But there’s a general assumption that Samsung parts generally perform better than components manufactured by other companies; this probably stems from the crappy LG panels Apple shipped with the first batch of Retina MacBook Pros. Indeed, in the case of the A9, the Samsung-made part uses a 14nm process, while the TSMC part uses a 16nm process. In theory, this should mean that the Samsung edges the TSMC technically and potentially sees greater efficiency.
But a new set of numbers published on Engadget suggests that the vast majority of iPhone 6S models use the TSMC chip, while a slight majority of 6S Plusses use the Samsung chip. Why would Apple put what is supposed to be a more power-hungry chip in the smaller phone? Because the other numbers published with the article suggest that the less-advanced, bigger TSMC processors are actually more power-efficient than the Samsung processors.
In all likelihood, the chips probably perform the same, broadly speaking. I doubt that the claimed two-hour advantage of the TSMC is replicable in real-world circumstances, and I’d be surprised if there were a noticeable difference between the two. It’s unlikely that they perform identically, but they’re probably very, very close.
Update: On the other hand, John Poole has posted a Geekbench chart with two obvious spikes from iPhone 6S users who have taken their battery test. This could be low-power mode, or it could be a hundred other factors. Or it could be a difference in processor foundry. Someone should test this.