Well that's kinda my point. I think I'll be able to find the 9.7 Pro for close to 50% off once the 10.5s hit. And I only use my iPad to browse the web and watch movies so I can easily wait another generation to spend a maxed out iPad at $700.
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.
A10X is the new A8X.
- Thread starter Kendo
- Start date
- Sort by reaction score
Well that's kinda my point. I think I'll be able to find the 9.7 Pro for close to 50% off once the 10.5s hit. And I only use my iPad to browse the web and watch movies so I can easily wait another generation to spend a maxed out iPad at $700.
In that case, have at it. For just $100-200 off, though? Imho, not worth it.
TBH, regarding longevity the iPP 12,9" didn't´t really need any hardware upgrade at all.
18 months in and it still has like 200-400% of the RAM that´s needed for anythings iOS-related..
...and at the current rate this machine gets any relevant software releases you can probably still use it successfully in 2027..
18 months in and it still has like 200-400% of the RAM that´s needed for anythings iOS-related..
...and at the current rate this machine gets any relevant software releases you can probably still use it successfully in 2027..
Regarding your quote to @Kendo in Post #26, I disagree with you and it is worth the purchase. They stated they are only going to use the 9.7 Pro for basic media and watching movies (Which is more than enough power and continued support to meet those tasks). The 9.7 Pro is selling for as little as for $425 already and will continue to drop in price once the new iPads actually are in store/available. Not to mention, price matching and gift card incentives will likely follow suit to liquidate any 9.7 Pro remaining stock.
Ofcourse 120hz vs 60hz helps with multitasking also. 120hz and ProMotion gives more responsive, lower latency with pencil and much smoother overall experience just like Apple said in the keynote. It's going to be very useful when doing multitaskings.
How that doesn't compute?
I think the A10X is like the A8X for many reasons, all positive unlike some seem to think. The A8X was, and still is, a very powerful chip. It was the first Apple SoC with more than two cores and more than 1GB of RAM. From that standpoint alone, the A8X will go down as one of the most future-proof SoCs for iPads. The A9 didn’t beat it in benchmarks, and the A9X just barely did because all three cores in the A8X were ‘High’ power versus the 2 high/2 low ‘quad’ setup in the A9/A9X. The Pro only got close because of a higher clock speed. The Pro 2 took the 3 high power cores of the Air 2, increased the clock speed to that of the A9X, tacked on 3 low power cores and quadrupled the L2 cache from 2MB to 8MB (the A9X only had 3MB). Let me put it this way: iOS 11 beta 1 runs buttery smooth on my 2.5 year old Air 2. If this tradition continues, I see no reason why the Pro 2 won’t last just as long, if not longer. I have doubts about the Pro 1 though.
Tack on the fact that all iPad Pro 2s will have 4GB of RAM, the A10X will go down with the A8X as one of the largest performance jumps and one of the most future proof design. That is, of course, if the leaked benchmarks are true (and I’m willing to bet they are).
[doublepost=1496995288][/doublepost]
Ehh... not quite. It’ll make system animations faster as long as the empty bufferspace in VRAM is large enough and the GPU isn’t over-taxed at the time. The refresh rate is changed by this ‘ProMotion’ companion chip (kinda like the display controller in the 5k iMac I assume) but the bits are pushed by the GPU. If the GPU can’t handle the refresh rate the ProMotion chip is requesting, it will slow down the refresh rate to whatever it can handle at the time, getting as close as it can go what the ProMotion chip requested
This is just how I’m assuming it’s going to work. But I know your refresh rate will only be as fast as your GPU allows.
Tack on the fact that all iPad Pro 2s will have 4GB of RAM, the A10X will go down with the A8X as one of the largest performance jumps and one of the most future proof design. That is, of course, if the leaked benchmarks are true (and I’m willing to bet they are).
[doublepost=1496995288][/doublepost]
Ehh... not quite. It’ll make system animations faster as long as the empty bufferspace in VRAM is large enough and the GPU isn’t over-taxed at the time. The refresh rate is changed by this ‘ProMotion’ companion chip (kinda like the display controller in the 5k iMac I assume) but the bits are pushed by the GPU. If the GPU can’t handle the refresh rate the ProMotion chip is requesting, it will slow down the refresh rate to whatever it can handle at the time, getting as close as it can go what the ProMotion chip requested
This is just how I’m assuming it’s going to work. But I know your refresh rate will only be as fast as your GPU allows.
Last edited:
So True. I haven't said opposite either.
So how much more the 120hz takes from the computing power? If you know the math why don't you lighten us and give the exact numbers?
Again, my point has to do with pricing. At significant discount, yes, it's a good option. At just $100-200 off similar configurations, just doesn't seem worth it.
Last edited:
Isn't quite the same as the ipad3.
First, Apple really had no choice but to announce the iPads w/A10X as they did. Had they waited for Sept it would have sent mixed messages since the A11 would have released by then.
The A10X has some staying power (I believe). Perhaps not as much as some would like, but I doubt the Pros get a processor bump in 2018.
Look to 2019 for the next major upgrade.
Yes, it would be have been nice for the Pro2 12.9 to have gotten 6gb RAM. But this is Apple and they don't bump the RAM until it really calls for it.
Last edited:
Yea, this is way off in my opinion. I had the iPad 3, they released it at a time where I think they underestimated how much power the new screen would take to run. The chip in that machine was way under powered. Now the chips are only getting more and more powerful while easily handling any hardware to go along with it.
Because ios needs 6gb of ram. Android is doing quite fine with 4 so ios devices is quite fine with 4
My windows laptop has a haswell Intel i7 quad and my Macbook Pro has broadwell and there fine. My Samsung galaxy tab s2 has the same exynos cpu from the Note 4 which was a year old already when it came out in 2015 and works great still. More power is always better but if it's not even being utilized than its now a huge deal. The A10x will last the lifetime of the product.
Industry best? Don't think so but best optimized since Apple controls it. The snapdragon 835 is better than it and the A10. Octacore cpu and the high performance ones are clocked at 2.5 compared I think the A10x is 3 2.45ghz?
Oh I'd disagree that the A series processors aren't industry best. Just look at the benchmarks, actual real world performance and they beat every other mobile device on the market. Clock speed isn't the only important factor. There is IPC (Instructions Per Clock), Core count, single-core performance, etc, etc. Snapdragon 835 is imo not that great, yes its the best processor for Android devices but on IPC and Single-Core performance its nowhere near the A10 nevermind the A10X. Their Single-Core performance is only reaching the A8X from 2014 with a much smaller node density at 10nm vs 20nm.
You kinda do, by suggesting that 120Hz demands more resources, even in the quote below. It doesn't work that way.
There aren't exact numbers here. It's a bit like asking "what is the 0-60 time of a car?" It depends on a lot of factors and is specific to the car. What you need in order to hit 120fps varies on the app, and what task you are doing at the time. But what 120Hz screens do not ever do is require you to render at 120fps. It just allows you to render at up to 120fps. Much like the 60Hz screens that the iPad has always had don't always mean the iPad is rendering at 60fps.
It all really comes down to a couple key factors though:
1) How much drawing does the CPU need to actually do, per frame?
2) How much overhead is there currently in the GPU to handle the higher fill rate?
The second question is murkier, but in general, it should be feasible to assume the fill rate is there. There are variables there that cannot be easily worked out, because we don't have the source code, but considering it's been possible to hit 60fps with GPUs less than half as powerful as the A10X, GPU fill rate isn't going to be your problem.
So the first question is the more interesting one. There are three scenarios here that demonstrate what will actually happen.
One extreme is certain simple user interactions. Say, dragging an item in a table to a new position, without any scrolling happening. Because all the rows are already drawn, you are just repositioning the existing layers the GPU needs to composite. Here, hitting the higher frame rate is super cheap, since you are drawing 0 pixels. The GPU grunts a bit more, but we've already demonstrated that the GPU has plenty of headroom.
The other extreme would be something like a 120fps animated GIF. You have to render each frame on the CPU (no hardware assist here), and you are potentially redrawing the entire screen. At 60fps, the CPU just drops every other frame, and draws 60 frames every second. At 120fps, the CPU is going to try to draw 120 frames every second, doubling the amount of CPU you need to accomplish the task. But these sorts of things are why we use hardware decoders for video, because this sort of work can get expensive quick.
A common real-world scenario though is scrolling. This one is interesting, because you can measure scrolling in terms of "pixels per second". The faster you scroll, the more expensive it can be. A slow scroll produces very little new drawing each frame, but a super-fast scroll you can wind up trying to redraw the entire screen every frame. But what's interesting here is that more frames doesn't by itself incur more drawing. Instead, you split the drawing into smaller pieces (each frame is moving half the distance, in half the time). Where this gets interesting is during those super-fast scrolls.
If I can scroll at N pixels per second on a 60Hz display, before it starts "skipping" lines of pixels between frames, then a 120Hz display will be able to accomplish 2N pixels per second before the same "skipping" of lines occurs. But the catch is that to perfectly draw the faster scroll, it does take more CPU. So any scroll under N pixels per second will use the same CPU power on both older and newer processors. Any scroll between N and 2N pixels per second on the new iPad will require more CPU at 120fps. If it can't though, you will start to see things like checkerboarding/etc again, or slower framerates.
Now, if you want to talk worst case, that's probably a good place to start. Take an app that manages to max out the A9X CPU during certain tasks, is single-threaded, but can hit 60fps during those tasks. That app will still render at ~80fps or better on the A10X, since the new screen allows it to. The CPU may still be maxed, but you are getting something for your trouble. Apps that aren't able to hit 60fps (and they exist today) will still be closer to 60fps on the A10X than on the A9X. There is effectively nothing to lose, and all to gain by going to 120Hz. Worst case, an app will "only" have a 30% higher frame rate, instead of 100%.
The short answer though: the A10X will produce better latency and smoother animations, when paired with the 120Hz display, than the A9X paired with a 60Hz display. Period. It doesn't matter what app, they will benefit from it. The only variable is how much.
Edit: Punctuation.
Last edited:
When I asked exact numbers, I really meant percentages (%) in theory, on the same task, with the same variables only difference is 120hz vs. 60hz. So the real question is, does the A10X really have enough headroom to allways do 120hz when available, even with lets say running 3 apps or tasks simultaneously?
Thank you very much to explain to the so called self appointed pundits on this forum.
120 hz display doesn’t mean the App has to compulsorily run at 120 fps. It can run at 60 fps or even 30 fps too just to save battery as per Apple’s dynamic display refresh rate technology.
I would call a buy on this 10.5” just as I called one on Air 2 about 2.5 years ago. Almost every specification has doubled over it. This one is the longevity model just like Air 2. Buy buy and buy.
Please mark my words and message me if this 10.5” model becomes unusable within 2 years. Worst comes to worst, Apple will make all iOS transition animations run at 60 fps on it. Apple is known to bog down the UI with unnecessary effects and transparency. If single transparency does not make it slow, Apple will add transparency over transparency just to artificially gimp it’s performance in the future. But that won’t stop me from buying it.
My own reasons for buying 10.5” model
1. 120Hz refresh rate
2. LTE-A 3CA up to 450 Mbps speeds
3. 256 GB storage is mid tier now. Fits my requirements.
4. Better display colors HDR, True-Tone.
5. 4 GB RAM in smaller size. 12.5” is too big for keeping it on chest and night reading.
Wishful thinking : Support for USB 3 transfer speeds over USB-C to lightning cable. That we’ll know once people start reviewing it themselves.
Last edited:
The answer is in the above post you just quoted. It is 100% dependent on the task being performed, and how much CPU it uses. So let's just use ranges (assuming single-threaded apps):
A9X + 60Hz Display: Baseline
A9X + 120Hz Display: 0-100% frame rate improvement.
A10X + 60Hz Display: 0-30% frame rate improvement.
A10X + 120Hz Display: 30-100% frame rate improvement.
Does this make it any clearer?