Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Hoping for the A11X processor to be the following
- Thread starter gobikerider
- Start date
- Sort by reaction score
Yeah the graphics processor will probably be laptop level at this point 
A10X already is the A11X is going to be ridiculous, not gaming laptop crazy but more then enough for photography, video rendering, and even cad and other 3D modeling things.Yeah the graphics processor will probably be laptop level at this point
Yeah definitely. Maybe Apple can finally put an operating system that actually suits that kinda power on it
(iOS 11 is a step in the right direction, but still not a powerhouse OS). Even so, the Pro is a very useful device because of its power.
Last edited:
I honestly feel like iOS 11 is a tease for iOS 12 when Apple will introduce Final Cut and Xcode for iOSYeah definitely. Maybe Apple can finally put an operating system that actually suits that kinda power on it
I believe Final Cut still heavily relies on GPU's, as well as multiple cores, and the iPhone's GPU will never get even close to a dedicated one. You just can't compete with a 150W+ part in such a small package.
A base model MacBook Pro seems to run Final Cut perfectly, and the current iPad Pro is performing right there with it. I think you may be exaggerating a tiny bit...
A10X already more than enough for me, but of course more power good
* 7nm fabrication
* 3 High Power Cores
* 5 Low Power Cores
* 8 Core 2nd generation Apple GPU
Predicted Geekebench scores
Single Core: 4,500
MultiCore: 12, 000
Compute: 45,000
Present A11 scores
Single Core 4500ish
Multi core 10k
Predicted for A11X
Single core 6000
Multi core 15000
I doubt that, the A11X would need a big clock speed increase to get 6000 single core speed, honestly I think we are reaching the upper limits of a single core in a proccesor, I mean look at other mobile chips. 4500-4900. Your multicore estimate could come to fruition depending on improved core perf and the ability of 8 core running together. I'm not sure though, I don't see 7nm processor being that insane. I'm keeping the proccesor estimates conservative and banking on an awesome new GPU from Apple's wizards.
I have a Mac Pro with an upgraded CPU and GPU. I often watch the activity of both the multicore Xeon and GPU, and Strelok is right. Both are taking up a LOT of power and producing a lot of heat to get a lot of work done quickly.
Although my iPad Pro 12.9” is wonderful, (I love it!!), and it does great with 4K video in LumaFusion, and the CPU performs great with Cubasis and Notion iOS, the wattage and massive torque of high core count Xeons and a separate high-performance GPU just can’t be matched by a SoC. It goes against the laws of physics. More power in, more power out.
Yes, a SoC will beat an i5, but not a workstation Xeon any time soon.
Even the 12 inch MacBook can run final cut. Albeit not fantastic 4K exporting
they've only just moved to 10nm fabrication, so I don't see 7 within the next year. It's harder than used to be to get to the next step down and still get sufficient yields after binning. Might be a big improvement in GPU if they're designing it from scratch specifically to be impressive...
My most wanted is an Apple designed GPU. Think that would be more groundbreaking then any of the above, would pretty much means Apple has cut its ties with imagination.
I’m thinking Geekbench:-
Single core: 4300
Multi core: 16000
Compute: 74000
Why?
A 7nm process will allow more transistors/watt so an extra performance core or even two is a given.
The A11 halved the GPU cluster count/power whilst increasing compute performance to 5,000/cluster - that’s nuts, Vega is sub-4,000! So even back to the 12 clusters of the A10X would give 60K but process shrink could go to 16 clusters to a theoretical max of 80K.
The 2018 iPad Pro will out perform the top spec 2017 15” MacBook Pro.
But wait, there’s more. If Apple can crack the 16-cluster PowerVR limit they could increase the power budget of the A11X2 to accommodate a 32, 48 or even 64 cluster design giving GTX1070 - GTX1080ti/Vega 64 + 50% and push for 4+4 CPU within a 50W package - that’s desktop performance in an ARM MacBook Pro!
[doublepost=1513247895][/doublepost]
If you’re running Adobe Apps you have no software efficiency so your argument would stand.
Single core: 4300
Multi core: 16000
Compute: 74000
Why?
A 7nm process will allow more transistors/watt so an extra performance core or even two is a given.
The A11 halved the GPU cluster count/power whilst increasing compute performance to 5,000/cluster - that’s nuts, Vega is sub-4,000! So even back to the 12 clusters of the A10X would give 60K but process shrink could go to 16 clusters to a theoretical max of 80K.
The 2018 iPad Pro will out perform the top spec 2017 15” MacBook Pro.
But wait, there’s more. If Apple can crack the 16-cluster PowerVR limit they could increase the power budget of the A11X2 to accommodate a 32, 48 or even 64 cluster design giving GTX1070 - GTX1080ti/Vega 64 + 50% and push for 4+4 CPU within a 50W package - that’s desktop performance in an ARM MacBook Pro!
[doublepost=1513247895][/doublepost]
Your laws of physics ignore the laws of software efficiency and the laws of shared CPU/GPU memory (how Serif get an A10X to outperform a Core-i7).
If you’re running Adobe Apps you have no software efficiency so your argument would stand.
More power is always good. My iPad Pro is my only computer besides my iPhone X so I always like to see progression here. Looking forward to the X-style revision, A11X, and iOS 12. A little worried it will be more so focused on the iPhone, but we will see.
The performance of the iPad Pro can only compete with entry level MacBooks in short bursts. This is why synthetic benchmarks don't tell the whole story. Video editing requires lots of number crushing over long periods of time. The iPad would soon throttle and slow considerably, but the MacBook will happily crunch away for hours on end.
I’ve never seen issues with iPad Pro throttling it’s chip. I have seen my iPhone do it but not iPad.
At a basic level yes, but the memory subsystem is usually also beefed up to handle increased bandwidth from CPU and GPU. They also usually have larger L1/L2 cache.
7nm is very likely a given if history repeats itself. 7nm would further allow for a larger battery and less heat. A new gpu from Apple with the thermal headroom of iPad is going to take the iPad to new levels of performance and graphic capability for professionals.
A11X is guaranteed to be 7nm.
Apple now uses the AX processors as pipe-cleaners to test a new node before using it for iPhone.
Keep in mind TSMC's 7nm is roughly equivalent to Intel's 10nm, so this is not some groundbreaking technology.
Apple used to be 12-18 months behind in manufacturing process. Since 2015, Apple caught up to the leading edge process, which was 14/16nm for A9.
New process tech is relatively risky. By using AX processors as pipe-cleaners, Apple reduces risk when using the new node for iPhone. We saw it with the A10X - an old architecture - fabbed using the new 10nm process. The more mature 10nm process is now used to fab A11 iPhones.
Ah yes the transistor density that Intel have been really struggling to get out in volume and have repeatedly pushed back, and indeed are still having teething problems with... of course...
I think that’s about to change. The A11X synthetics should match quad core i7s (only 15K). Whilst video rendering will push the thermals, Affinity Photo clearly demonstrates the Ax advantage a rung down the ladder where architectural efficiency breaks the wattage = workload myth.
Last edited: