Next-Gen iPhone & iPhone Nano Speculation

[tangey]

Rogue is *possible* for next gen iPad, timing could be tight.
It'll definitely be in the next gen iPhone.

Can't see either of them changing resolution, so the graphics performance improvement should be impressive.

With ipad2/iphone4s, apple deliberately gave them identical on screen performance, it might be their intention to get back to that with the next gen of both. It was impossible to do so on this cycle in any practical way. Or are they just happy for games to aim for 1024x768 on the iPad,

I wonder will they use the A6 in the iPad mini, or will they keep it on 2nd tier performance in the same way the itouch is, likely the latter.

 

[wco81]

Given the expected price of the iPad Mini, you wouldn't expect it to have the highest specs.

 

[zed]

I cant see the ipad 4 being as bigger jump GPU wise sa say ipad 1->2 or 2->3
chief reason, its too heavy ATM they are going to want to cut down the battery size
FWIW the ipad3 is very impressive gfx-wise theres no need to greatly improve it

 

[Lazy8s]

The next iPad's A6X certainly should be a bigger jump in graphics. G6400 and at least some bump in clocks would ensure that result while the 45-to-32nm transition of the app processor from the A5X, more efficient display, and better all-around power management will still enable a thinner and lighter device profile.

Apple wouldn't design their own (3D) GPU. The specialization and expertise needed for that is why they work specifically with ImgTec. Even spending years and humungous resources on it wouldn't ensure competitiveness.

I had been anticipating that Apple's custom ARMv7 could make a debut as early as this new iPhone, and the design reflects their superior grasp of SoC balance as expected. The very-respectable-but-not-outrageous doubling (and more) of GPU performance in the lesser year of their two-year cycle acknowledges the importance of the graphics/highly-parallel processor to the system while also acknowledging the futility of trying to go all-out every year for the performance crown when, invariably, some competition (Qualcomm, for example) will be upgrading on an opposing cadence. The new CPU must obviously be focusing on power efficiency before performance yet still remains competitive, owing to its customized design.

IMG recently commented that they expect Rogue-powered devices on the market by the middle of next year, which gives me the slightest hesitation in expecting it for the next iPad (though I think their wording has enough ambiguity to not throw too much doubt on an A6X debut).

 

[wco81]

I haven't looked too much for games which would tax the graphics yet.

I might be willing to try things like Madden or FIFA if they have online play comparable to consoles.

Looks like EA may not be updating their sports games for iPad this year.

Even with improving graphics there doesn't seem to be a market for 3D games yet.

The most taxing 3D thing that I will probably be doing for awhile is flyover in the maps.

 

[ams]

I absolutely expect to see Rogue debut in the ipad 4 A6X SoC, and I also expect to see the A6X SoC have a significantly larger die size than the A6 SoC (and in turn, the normalized SoC die size of A6X will be larger than A5X). If Rogue is simply not ready until April, May, or June 2013, then that is when the ipad 4 will come out, but I doubt it will be that late.

 

[ltcommander.data]

I absolutely expect to see Rogue debut in the ipad 4 A6X SoC, and I also expect to see the A6X SoC have a significantly larger die size than the A6 SoC (and in turn, the normalized SoC die size of A6X will be larger than A5X). If Rogue is simply not ready until April, May, or June 2013, then that is when the ipad 4 will come out, but I doubt it will be that late.

It may be semantics, but I would think that if changing CPU architectures while leaving the GPU architecture unchanged constitutes a generational leap from A5 to A6, then changing GPU architecture to Rogue while leaving the CPU architecture unchanged would likewise be a generational change from A6 to A7.

 

[ams]

Assuming that CPU performance is fairly similar between iphone 5 and ipad 4, with the main differentiator being the GPU performance, then I would expect Apple to use A6X as the SoC name for ipad 4 (in a similar manner to how they used A5X for ipad 3 after using A5 for iphone 4s).

 

[Exophase]

Who cares what they call the SoC in the next iPad?

I expect the same CPU but with higher clocks and maybe with more cores (not sure why a native triple core CPU seems to be such a taboo concept). But Apple does have more leverage now to make more minor, fluid generational enhancements to their CPUs, now that it's their design. Same with what Qualcomm appears to be doing with Krait.

Rogue would of course make sense, if it's ready.

 

[sebbbi]

So... it seems that there's no brand new GPU in the iPhone 5. Accoding to Anandtech it's PowerVR SGX 543MP3.

I am just wondering if there's any possibility that the forthcoming iOS versions will enable OpenGL ES 3.0 support for iPhone 5. According to the PowerVR guys in the forums their chips technically support MRT rendering, so that requirement could be already met by the hardware. Current iPhone also already supports limited occlusion query by extensions. Volume (3d) textures, transform feedback and instancing are not supported, and neither are the new compressed texture formats. I doubt all of these requirements are supported by existing PowerVR hardware... so it unfortunately seems that we have to wait for iPhone 6 to have full OpenGL ES 3.0 support

Android phones with ES 3.0 support are going to be available later this year.

 

[Lazy8s]

Android phones which would support it might be available, but GL ES 3 content itself won't be ready in any appreciable fashion until Apple launches the next iPad with support for it in the first half of 2013, I bet.

And yeah, SGX will not quite reach GL ES 3 compliance.

Anandtech's guess of 267 MHz for the iPhone 5's GPU is most certainly wrong. Tangey appears to have it right; it's either 300 or 325 MHz (slight possibility of being even higher). Looking back in this thread, my guess of "either an MP3 @ 250 MHz or an MP4 @ 200 MHz" several months ago was pretty close, but Apple's ambition pleasantly surprised me here. They've still left plenty of room for a huge graphics upgrade in the new iPad along with new ways/applications to take advanage of it as being one of it's major selling points.

 

[Ailuros]

IF A6 is truly just roughly 95mm2@32nm, it leaves quite a bit of headroom for the next tablet SoC early 2013. Even more so if they go for 28nm.

 

[Lazy8s]

Definitely. I also think the move to their own CPU design was made specifically with the goal of efficiency -- so that they could still scale up in the future without having to play the same gigahertz game of some of their competitors -- in order to leave more headroom, again, for pushing harder on the GPU in the future.

 

[Ailuros]

Definitely. I also think the move to their own CPU design was made specifically with the goal of efficiency -- so that they could still scale up in the future without having to play the same gigahertz game of some of their competitors -- in order to leave more headroom, again, for pushing harder on the GPU in the future.

By the way unless I'm reading out stuff wrong at the Kishonti database, the iPad3 results seem to be 2.5.0 only.

 

[Diamond.G]

This is a fairly elementary question, but I can't seem to find a good answer. How important is Fill Rate in a TBDR arch? I noted in a MacRumors thread that the A5X's MP4 is CPU limited as shown by the A6's MP3. Someone chimed in and said with was due to memory bandwidth being so high in the A6.

 

[french toast]

Very smart of apple...improve memory subsystem/cache, use wider execution and keep clocks low, thus leaving precious tdp for advanced graphics acceleration.

Ailuros has a point, if apple moved to 28nm, rogue, lpddr3....that is going to be one advanced power efficient chip.

Such a pity I don't/won't purchase apple products...if only Nokia designed those apple chips and stuck them in a lumia!!

 

[GpuMonkey]

Apple wouldn't design their own (3D) GPU. The specialization and expertise needed for that is why they work specifically with ImgTec. Even spending years and humungous resources on it wouldn't ensure competitiveness.
/QUOTE]
"We are currently building a GPU team."

That is a direct quote from an Apple recruiting email received a month ago.


Does not necessarily mean they are designing a GPU, of course.

 

[Lazy8s]

They'll want a GPU team for customizing implementations, perhaps developing a power management or power saving companion core, but central IP will still be PowerVR by all indication. Apple has invested in a strategic partnership with IMG to ensure the roadmap accounts for their needs.

 

[anexanhume]

They'll want a GPU team for customizing implementations, perhaps developing a power management or power saving companion core, but central IP will still be PowerVR by all indication. Apple has invested in a strategic partnership with IMG to ensure the roadmap accounts for their needs.

How much customization does their license buy them? I'm also curious how much custom work ImgTec does for their customers for their cores. For example, the 543MP4+ in the Vita. I'm assuming the + refers to the now known 1024 bit memory interface. I'm very curious to see how (if) they implement Rogue on the A6X and if they go 28nm, which means TSMC before H2 2013.

 

[Lazy8s]

Samsung's foundry side has been delivering for Apple and should be able to transition them to 28nm relatively smoothly, and both companies are mature enough to keep their mutually beneficial business interests separate from their unrelated patent/industry-politics warfare.

I can imagine Apple wanting the flexibility of multiple suppliers, yet I don't think they have a pressing need for it right now.