For some reason your links are broken.
supportsFeatureSet: returns a BOOL telling you whether the requested feature set is supported.
Apple A8 and A8X
- Thread starter ltcommander.data
- Start date
supportsFeatureSet: returns a BOOL telling you whether the requested feature set is supported.
iMacmatician
Regular
Given the rumors of the iPad "Pro"'s features (stylus, multiple stereo speakers, USB-C port), I think the "Pro" is at least partially aimed at expanding the market of the iPad line.
Well yes, however haven't there also been rumors that the iPad mini might reach its end of life as a category soon? IF its true it means that they're only moving into another direction, which doesn't probably mean more units but probably higher profit margins if the price is high enough.
That's what some people are saying, that the new MacBook is the 12 inch iPad some people wanted.
http://www.macworld.co.uk/review/mac-laptops/new-12-inch-macbook-vs-ipad-air-2-3605589/
http://www.macworld.co.uk/review/mac-laptops/new-12-inch-macbook-vs-ipad-air-2-3605589/
Typically, I'd say that this was unreasonable.The DX10.0/GPU in A8X is under 20nm at 38+mm2. That die area doesn't help because they just mirrored a quad GX6450 but let's ignore that for a second. I'd estimate that it would blow up in die area nearly to 60mm2 under 20nm in theory for an 8 cluster DX11.2 7XT part. Since they're scaling just clusters die area shouldn't scale linearly upwards, but are you sure the bullet for 16 DX11.2 clusters is already "biteable" under 16/14FF? If yes it would mean a ULP SoC GPU roughly as big as the entire Exynos 7420 on estimate
However,
I think this is correct, and the reason why Apple might conceivably be very ambitious with a hypothetical iPad Pro SoC.
Basically, iPads are Apples idea of the future of personal computing devices, with phones typically taking the tasks of communication/entertainment/photography/video.
They want iPads to cannibalise laptops. And they do, last reported quarter Apple sold 21.5 million iPads, as many iPads as they sell macs, in total, over a year. However, tablets have an unfortunate reputation as being devices primarily for surfing the web, and cheap android tablets have reinforced this strongly. (The other day, I was offered an android tablet for €5 if I filled in a survey. I didn't do it.) The only clear exception to this public impression is Microsofts Surface Pro products. Now these don't sell anywhere near Apples volumes, roughly a million were sold the same quarter Apple sold 21.5, but they are clearly pushing into exactly the segment that Apple wants iPads to (also) occupy. I don't think Apple wants to leave that segment uncontested to Microsoft.
Now, having a larger, higher resolution screen allows more information to be displayed. If they have USB-C, it allows connection to high-resolution external monitors via DisplayPort Alt mode. They can expand memory and storage. Together this makes for a more "professional" device, but it will inevitably be compared to the Surface Pro. Can they allow it to fall short performance wise? If I were a strategist at Apple, I would bite the bullet, and ensure that performance was also at a level that both clearly signalled that a new market was being addressed, and didn't fall short of Broadwell/Skylake in Microsofts offerings, die area be damned.
Of course, I am not actually a strategist at Apple.
God knows what they'll do SoC-wise if they actually release such a tablet.iMacmatician
Regular
There have, but recently there have also been photos of a claimed iPad mini 4 shell, showing changes compared to the iPad mini 3.
ltcommander.data
Regular
https://twitter.com/gavkar/status/585808280695349248I'm not sure but it's my understanding that the GT7900 is DX11 only; if yes and there isn't a DX10.0 variant I severely doubt Apple would want to bite the >50% additional transistor count for something they won't have any use for, for quite some time to come. I also wouldn't expect anything as large as a 7900 before 10nm or an equivalent process. Considering alone that it has 32 TMUs it sounds too far fetched for anything 16/14FF for a hypothetical 12-13" device that shouldn't also weigh as much as concrete cement brick.
One theoretical solution if Apple truly wants the large Pad to differentiate (if it even exists after all as a project....) would be to take a GT7600 and just mirror it for 12 clusters. My gut feeling however so far tells me that if a GT7800 clocked at say 600MHz will deliver =/> performance than the X1 GPU, that they will NOT aim for anything higher than that. IMHO a potential =/>12" user will be more happy with a slightly smaller GPU but N% more battery lifetime instead AND of course X sustained performance
I noticed some talk on Twitter among developers that they are now often texture bound. There may actually be a need to go above the 16 TMUs that a GT7800 offers although 32 TMUs may be too much.
On the iPhone side being an expected "s" speed refresh, I would think they'd be pushing a 2x+ GPU speed improvement, especially since the A8's 50% improvement was largely eaten by the resolution increase so performance per pixel hasn't improved for a while. As such, I could see an A9 with a dual-core new CPU micro architecture (bigger changes than A8 vs A7 although not likely as aggressive as A7 vs A6) with a GT7600 with different clock speeds for the iPhone 6s and 6s Plus. And for the A9X, similar to what you propose, a triple-core new CPU micro architecture and 2xGT7600 with different clock speeds for the iPad Air and iPad Pro. The iPad Mini seems too popular to discontinue, but they may well want to de-emphasize it and bump up profit margins by keeping on a previous generation SoC, in this case an A8 for the iPad Mini 4.
On the other end of the scale and on the eve of Apple Watch availability, any speculation on the CPU and GPU in the S1? It's been rumoured to be on Samsung 28 nm, but there doesn't seem to have been much talk on what's inside. A re-use of existing CPU and GPU designs like Cortex A9, Swift, Cyclone, SGX535, SGX543, a previously unused by Apple but stock design like Cortex A7, Cortex A53, GX5300, G6110 or something new and custom? A dual Cortex A53 with a G6110 and 512 MB of RAM would provide good performance for the future native apps that are said to be coming (unlike the current streamed from iPhone apps), but perhaps is too power hungry?
iMacmatician
Regular
Does anyone expect Intel-style boost to appear on Apple's SoCs anytime soon?
I wouldn't be surprised at 2x GT7600 for the A9X especially if one or more of the 2015 iPads will have 3x Retina. Apple has a number options, assuming a "Pro" will be released this year and the Air 3 and "Pro" are 264 PPI (2x) or 396-401 PPI (3x):
I wouldn't be surprised at 2x GT7600 for the A9X especially if one or more of the 2015 iPads will have 3x Retina. Apple has a number options, assuming a "Pro" will be released this year and the Air 3 and "Pro" are 264 PPI (2x) or 396-401 PPI (3x):
- Release the iPad Air and iPad "Pro" with 2x and use clock speeds to take care of the resolution differences. I think this option is sensible but it won't explain why 3x iPad assets were found last year.
- Release the iPad Air with 3x and the iPad "Pro" with 2x. This option results in similar resolutions, but I tend to think that it is unlikely since previous Retina transitions have started with the high-end models in each lineup first (9.7" iPad, 15" MacBook Pro, 27" iMac).
- Release the iPad Air with 2x and the iPad "Pro" with 3x, and use different SoCs for the two models since the latter has 4x the resolution of the former. May be unlikely for reasons previously mentioned.
- Release the iPad Air with 2x and the iPad "Pro" with 3x, and use the same SoC for the two models. Apple has previously used the A7 in the iPhone 5S and the iPad Air, which have a similar resolution difference, so that may happen here. But if Apple is going the higher-end route with the "Pro" (as in Entropy's post above), then I don't think it makes sense for the iPad "Pro" to perform a lot worse than the iPad Air 3 at their respective native resolutions.
- Release the iPad Air and iPad "Pro" with 3x and use clock speeds to take care of the resolution differences.
For Series5XT it's 32 FLOPs for 2 TMUs and at =/>Series6 it's 64 FLOPs for 2 TMUs. The latter ratio (let's call it 32 FLOPs/TMU) is the same though for all architectures like Rogue, Midgaard, Maxwell amongst others. If memory serves well that ratio is even the same in the desktop today, it's just that there a GPU has a crapload of more (and dedicated) bandwidth to use.