Originally Posted by VL-Tone
I see a lot of people assuming that the A7 (not A7X) could drive a retina iPad without any problem based on some raw GPU benchmarks that shows the A7 performing better than the A6X.
Actually the GPU isn't an overall win.
Well I'm not so sure about that. I'm not a GPU expert, but I don't think that pushing 4x the number of polygons is the same thing as pushing 4x the number of pixels. The latter might require more parallel processing in the form of a multi-core GPU, such as the ones found in the A6X (or A7X).
The behavior of the processor would vary with apps. Beyond that Apple could easily change the clock rate of the GPU cores in the A7 as opposed to making an A7X. The discussion is frankly a little premature considering nothing has shipped yet. The A7 however brings a lot to the performance table with only a few regressions in the GPU. How much the more dramatic the other GPU improvements are, have already been seen in many apps so it looks like Apple has seen a good balance of improvements. Beyond all of this you have to understand that a Retina in a Mini might not mean a 2X density increase in pixels.
An A7 could perform worse than an A6X if it runs into bottleneck issues because of the large number of pixels it has to push on an iPad.
It could no doubt, but Apple wouldn't put a plan into place that resulted in dramatic reductions in iPad Mini performance. Just see an overall potential increase in performance with A7, and the value of moving to 64 bit as soon as possible.
As an aside there is some mystery with respect to a RAM array embedded in A7, some of us think it is a frame buffer to support the GPUs. Apple could resize that array for other screen sizes. This might not yield an A7X but maybe an A7-0.5X. Of course this is guess work but the point is there are many ways to address A7 processors in a Mini.