|
Also, if they'd wanted A57, they could have just implemented it; it was available in the right timeframe. No it wasn't. A57 wasn't actually done when Apple outed Cyclone. Apple clearly ran with early aspects of A57 (your disingenuous misreading about ARMv8 being ignored). At exactly the same timeframe, nvidia, Qualcomm, and others had as many details of A57 and ARMv8, and none of them ran with it. Indeed, nvidia and Qualcomm are just getting to 64-bit parts, doing a pretty good job with their existing architectures. Apple wanted to be first, and they pounded every bit out of 64-bits in their marketing. But here, again, we have to pretend that reality isn't as it actually is. Also, it's clear that they DON'T try all that hard; they could have used the six-core variant of the PowerVR used here, at the cost of battery life This is absurdity theater. Apple has chosen higher bin/tier PowerVR parts than her peers for virtually every single generation. But because they didn't choose even higher we are to believe that they don't try hard? Do you actually look at the benchmarks, or just continue on with a ridiculous narrative based upon preconceived notions? |
In April 2013, ARM announced that the first production A57 chip had been fabbed and tested at an unspecified prior date: http://www.arm.com/about/newsroom/arm-and-cadence-partner-to...
> At exactly the same timeframe, nvidia, Qualcomm, and others had as many details of A57 and ARMv8, and none of them ran with it.
Google is showing me lots of documents published by ARM on ARMv8 from 2011. By Oct. 2012, Samsung had licensed the A53 and A57 core designs: http://www.techhive.com/article/2013298/arm-introduces-64bit... . They didn't have 'many details'; they had the design itself.
> Apple clearly ran with early aspects of A57
Which ones? I mean, Cyclone was wider than A9, but that's hardly a shock. My impression is that A57 isn't, in any case, really suitable for phones except in a BIG.little configuration.