Hacker News Reader

Comment by 6ren | Hacker News Reader

How can the iPad 5 have x2 the graphics power over the iPad 4, when the iPad 5 has the same A7 as the iPhone 5S - rather than an A7X, as they've done previously.

Have they just stopped using the A7X style name? Have they just x2 the GPU frequency? (seems implausible, but it's Rogue, which is apparently extremely efficient, and maybe could handle x2 frequency?).

edited above for clarity

BTW: The anandtech iPhone 5S review (http://www.anandtech.com/show/7335/the-iphone-5s-review/7) noted that the A7's rogue G6430 is more powerful than was utilized in the iPhone 5S:

  At 200MHz that would give the A7 twice the peak theoretical performance of the GPU
  in the iPhone 5. And from what I’ve heard, the G6430 is clocked much higher than
  that.

If 200MHz is correct, I bet the iPad 5 just doubled the GPU frequency to 400MHz.

View on HN

The iPad4 doesn't pack an A7.

http://en.wikipedia.org/wiki/Apple_A6X

True, but I don't follow how that's connected to my comment... maybe I should explain the background of my comment.

  The iPhone 5S (A7) has x1 the graphics power as the iPad 4.
  The iPad 5    (A7) has x2 the graphics power of the iPad 4.

So... how to get double the graphics power out of the same SoC (A7)? That was the puzzle (likely ans: x2 frequency)

You may be incorrectly assuming all A7 SoCs are identical. The Apple TV has an A5, but it's only single core, unlike it's dual core phone and tablet brethren. Maybe they just dropped the X moniker.

If you have a look at my original comment, I suggested that: "Have they just stopped using the A7X style name?"

oh, right. I think that's the answer.

You could be right, but have a read of the review I linked above, search for the text I quoted, and read around that. To me, it's compelling that the iPhone 5S's GPU is severely underclocked, leaving headroom for reuse in the iPad 5. And of course higher-clocking is usual in iPads, with their larger battery.

BONUS: for next year, the G6630 gives 50% performance increase + the expected node shrink allows a further ~30% overclock. 1.5*1.3 = 1.95 ~ 2.0.

For the following year, they could go for G6630MP2 (the 6 clusters can be multicored as a whole). Next (three years from now), they could go G6630MP3 plus another node shrink. By that time, there could be 8/10 clusters available G6830, G6A30 etc, and an entirely new architecture in the works, though probably not yet available.

It's possible they wanted to be misleading like that. So the A7 in iPhone 5S could have a certain level of performance, while the "A7" in iPad Air has its GPU clocked twice as higher, but the same name.

Reply on news.ycombinator.com