Question about CPU's & *iPad Pro

[Will0827]

I like possibly many (if not many then ill still admit to it) get all giddy about geekbench scores for machines without a clue as to what it means in real world use, pulling the obvious aside of a higher number being better, can someone give a sort of leyman's explaination for these scores can be technical if need be. I look at the scores for the the past couple of iPad processors and they double in each iteration and if that continues the A9X will be in line if not slightly above the base i5 MBP 13" from 2012 which is the machine i currently own. If this is true what exactly does it mean that the iPad has that score in the real world. A common thought expressed about the ipad is how it's still a toy and not professional because of iOS. If the reason for this sentiment is because of the apps, and if the apps were limited because of power then i feel going forward that wouldn't be a limitation. There are other things like a proper file management system which seems apple sort of has implemented via the icloud drive app in iOS 9, and then possibly with the ipad pro having a usb c port instead of lightning may allow hardware expandibility. Anyway what the scores mean, will the fact that the A series processors in the ipads are getting just as strong if the scores are the thing to go by, going to encourage developers to make more feature packed apps or make desktop equivalent's of currently well known apps. Thanks MR fam.

 

[Andres Cantu]

I like possibly many (if not many then ill still admit to it) get all giddy about geekbench scores for machines without a clue as to what it means in real world use, pulling the obvious aside of a higher number being better, can someone give a sort of leyman's explaination for these scores can be technical if need be. I look at the scores for the the past couple of iPad processors and they double in each iteration and if that continues the A9X will be in line if not slightly above the base i5 MBP 13" from 2012 which is the machine i currently own. If this is true what exactly does it mean that the iPad has that score in the real world. A common thought expressed about the ipad is how it's still a toy and not professional because of iOS. If the reason for this sentiment is because of the apps, and if the apps were limited because of power then i feel going forward that wouldn't be a limitation. There are other things like a proper file management system which seems apple sort of has implemented via the icloud drive app in iOS 9, and then possibly with the ipad pro having a usb c port instead of lightning may allow hardware expandibility. Anyway what the scores mean, will the fact that the A series processors in the ipads are getting just as strong if the scores are the thing to go by, going to encourage developers to make more feature packed apps or make desktop equivalent's of currently well known apps. Thanks MR fam.

I don't know exactly what the scores represent (or if they are at least directly comparable to those of machines using Intel's architecture vs. the Apple A-series ARM architecture), but I also like to see them grow as fast as they can. That's why I was disappointed with the A8's score, but not the A8X's. I have a feeling that the A9 will have a very similar score to the A8X, as it will probably go tri-core, but it could be a bit higher if it has architectural improvements. If this is the case, then I can see them using an A9 for the iPhone 6s and 6s+, the iPad Air 3, the iPad Mini 4, and an A9X in the iPad Pro. If the A9X is quad-core, then it's Geekbench score could be somewhere around the 5,000's, maybe even reach 6,000! (But not yet over 9,000, sadly)

 

[soupcan]

A9 will probably be A8X on a smaller fab process, so tri-core with 2GB of RAM. Then Apple has a dilemma. If the A9X will be the powerhouse with 4 cores and 3 GB of RAM to support pro apps on the iPad Pro, what will Apple put in the iPad Air? The A8X already eats anything else out there in terms of single- and multi-core performance (for instance, the tri-core A8X is about 5% slower in multi-core performance than the chip in the Galaxy S6 while being clocked 500 MHz slower and the latter being an 8-core while the A8X eats the S6's chip in terms of single-core performance), so Apple can either put the standard A9 in it (like they did with the A7 in the original iPad Air) or make a 3rd line of A9*insert letter here* that's made just for iPad Air. It doesn't need so much performance since no apps can take advantage of both the CPU horsepower and the 9.7" display. It does pave the road though for Apple to introduce their own chip designs in Macs since they already managed to make the A8X a powerhouse. I think folks like Qualcomm (Snapdragon) and Samsung (Exynos) were probably surprised when they saw the A8X performance. Hell, when the A7 was launched and it brought 64-bit to the masses Qualcomm **** their pants because although they had developed some chip designs for use in smartphones Apple beat them by quite a lot of time which took them completely by surprise.

As for the Geekbench scores: the score that people refer to is the overall performance score which is made up by several sub-scores determining CPU performance, memory speeds and various other things inside the CPU (the GPU part of the chip is left alone). It makes the CPU do various things including but not limited to image compression and decompression, security encryption and decryption, file compression and decompression, and tests the memory on its speed by copying certain files and measuring how fast it does that (and then some floating point stuff). It does that both for single core and multi core, and then takes the scores it gets from that and does some math on it and boom, out comes your final score which is the score you see on so many comparison websites. It's an average performance number and a higher score means that one CPU is faster than another one, overall. However, there are always certain things one CPU does better than another CPU despite the former having a lower score than the other.

 

[Will0827]

A9 will probably be A8X on a smaller fab process, so tri-core with 2GB of RAM. Then Apple has a dilemma. If the A9X will be the powerhouse with 4 cores and 3 GB of RAM to support pro apps on the iPad Pro, what will Apple put in the iPad Air? The A8X already eats anything else out there in terms of single- and multi-core performance (for instance, the tri-core A8X is about 5% slower in multi-core performance than the chip in the Galaxy S6 while being clocked 500 MHz slower and the latter being an 8-core while the A8X eats the S6's chip in terms of single-core performance), so Apple can either put the standard A9 in it (like they did with the A7 in the original iPad Air) or make a 3rd line of A9*insert letter here* that's made just for iPad Air. It doesn't need so much performance since no apps can take advantage of both the CPU horsepower and the 9.7" display. It does pave the road though for Apple to introduce their own chip designs in Macs since they already managed to make the A8X a powerhouse. I think folks like Qualcomm (Snapdragon) and Samsung (Exynos) were probably surprised when they saw the A8X performance. Hell, when the A7 was launched and it brought 64-bit to the masses Qualcomm **** their pants because although they had developed some chip designs for use in smartphones Apple beat them by quite a lot of time which took them completely by surprise.

I can see apple using the newly released macbook as the first macbook to run on their A series processors, it remains to be seen the performance of the skylake core m processors but it seems the AX processors could be viable atleast for the Macbook sooner rather than later. If this were to be the case, how much work would developers have to do on their applications to support the would be arm processors. Im assuming it would be a bit more than just recompiling.

 

[soupcan]

I can see apple using the newly released macbook as the first macbook to run on their A series processors, it remains to be seen the performance of the skylake core m processors but it seems the AX processors could be viable atleast for the Macbook sooner rather than later. If this were to be the case, how much work would developers have to do on their applications to support the would be arm processors. Im assuming it would be a bit more than just recompiling.

If ever decides to put an A-series chip in a Mac it will be an x86-x64 based chip, not an ARM design. Apple can license an x86-x64 architecture without too much hassle (that's why we have both Intel and AMD on the CPU platform and not just Intel). I mean, look at Windows RT. It never left the ground because nobody was interested in doing multiple OS's on different architectures. Apple knows that. Apple would have to create 2 different platforms for the same OS and that just isn't Apple. OS X is for the x86-x64 architecture and iOS is for the ARM architecture. Apple wants to keep things simple because as soon as they make an OS X version for the ARM design they'll have lots people complaining why one app doesn't work on their Mac while it does on another one "because OS X looks the same on both machines", failing to realize the underlying technology is totally different and not compatible in any way, shape or form.

 

[Will0827]

If ever decides to put an A-series chip in a Mac it will be an x86-x64 based chip, not an ARM design. Apple can license an x86-x64 architecture without too much hassle (that's why we have both Intel and AMD on the CPU platform and not just Intel). I mean, look at Windows RT. It never left the ground because nobody was interested in doing multiple OS's on different architectures. Apple knows that. Apple would have to create 2 different platforms for the same OS and that just isn't Apple. OS X is for the x86-x64 architecture and iOS is for the ARM architecture. Apple wants to keep things simple because as soon as they make an OS X version for the ARM design they'll have lots people complaining why one app doesn't work on their Mac while it does on another one "because OS X looks the same on both machines", failing to realize the underlying technology is totally different and not compatible in any way, shape or form.

So essentially apple will be making it's own x86 based chip. Something that i sort of had a grasp on understanding is that the other companies that make mobile processors are liscening the Arm architecture, and making their own on processors. So apple would do the same thing except it will be based on x86. Naturally i would ask why hasn't this been done before and i would assume the answer would be waiting the relationship already established with intel, that's if i even have the right understanding of how this scenario would play out.

 

[soupcan]

So essentially apple will be making it's own x86 based chip. Something that i sort of had a grasp on understanding is that the other companies that make mobile processors are liscening the Arm architecture, and making their own on processors. So apple would do the same thing except it will be based on x86. Naturally i would ask why hasn't this been done before and i would assume the answer would be waiting the relationship already established with intel, that's if i even have the right understanding of how this scenario would play out.

Because in the past Intel could keep up with its schedule, something they haven't been able to do for a while now. If Intel was still on schedule we would've been on Skylake right now, not Broadwell. This means Apple couldn't roll out the products they wanted to roll out (for instance, a redesigned MacBook Pro).

The experience they got from building the A-series and the great public response they got from it means Apple is ready to make it into the market of x86 chips. That way they have even more control over the hardware so that their designs can be even more powerful/power efficient and they can roll out their products whenever they feel like it. Couple of problems though. One, they need to have a stable design that is bug free. Two, well, rather an easy one, they have to find a place to fabricate the thousands if not millions of chips for them. Three, and this is the hardest one, Apple has to convince the market that their chip is better than the competition's, and simply saying "this is twice as fast" isn't doing it anymore.

I'm waiting for Apple to shake up the industry again, like they did with the A7 and A8X.

 

[Andres Cantu]

Because in the past Intel could keep up with its schedule, something they haven't been able to do for a while now. If Intel was still on schedule we would've been on Skylake right now, not Broadwell. This means Apple couldn't roll out the products they wanted to roll out (for instance, a redesigned MacBook Pro).

The experience they got from building the A-series and the great public response they got from it means Apple is ready to make it into the market of x86 chips. That way they have even more control over the hardware so that their designs can be even more powerful/power efficient and they can roll out their products whenever they feel like it. Couple of problems though. One, they need to have a stable design that is bug free. Two, well, rather an easy one, they have to find a place to fabricate the thousands if not millions of chips for them. Three, and this is the hardest one, Apple has to convince the market that their chip is better than the competition's, and simply saying "this is twice as fast" isn't doing it anymore.

I'm waiting for Apple to shake up the industry again, like they did with the A7 and A8X.

Yep, I'm waiting for that too. I'm sure it will happen, especially if it can match the S6's score with "only" three cores. Also, I feel like they will use the A9 chip on the iPhone 6s, iPhone 6s+, and iPad Air 3 (maybe iPad Mini 4). Historically, they've never used an AX chip on an iPad Mini, so it would make sense to do what they did for the A7 again in the A9 and take advantage of economies of scale. They could make the A9 in the iPad Air 3 run at a higher frequency like the iPad Air's A7. I feel like they would reserve an A9X chip for the iPad Pro. But I'm sure that the iPad Pro will have 2GB of RAM (albeit LPDDR4), they'll save the 4GB of RAM for a future model (not saying it's the ideal case, just what I think they will do).

 

[XTheLancerX]

Yep, I'm waiting for that too. I'm sure it will happen, especially if it can match the S6's score with "only" three cores. Also, I feel like they will use the A9 chip on the iPhone 6s, iPhone 6s+, and iPad Air 3 (maybe iPad Mini 4). Historically, they've never used an AX chip on an iPad Mini, so it would make sense to do what they did for the A7 again in the A9 and take advantage of economies of scale. They could make the A9 in the iPad Air 3 run at a higher frequency like the iPad Air's A7. I feel like they would reserve an A9X chip for the iPad Pro. But I'm sure that the iPad Pro will have 2GB of RAM (albeit LPDDR4), they'll save the 4GB of RAM for a future model (not saying it's the ideal case, just what I think they will do).

Yeah then they will barely increase the power of the A9 from the A8X, but the iPad Air 3 will move to an @3x display and become the next dud iPad. Odd numbers are just bad for Apple apparently. iPad 1 didn't have enough RAM, iPad 3 didn't have enough GPU power for that retina display, iPad Air is a combination of the iPad 1 and iPad 3.

I really am super anxious about the next iPad because I am looking to make that *my* next iPad. I've suffered enough with the under powered iPad mini 2, I don't want to go through it all over again with the iPad Air 3. Please Apple, focus on performance, not spec wars. We don't need a higher resolution on our regular iPads, only on the supposed "Pro" version which will be able to handle it...

 

[Andres Cantu]

Yeah then they will barely increase the power of the A9 from the A8X, but the iPad Air 3 will move to an @3x display and become the next dud iPad. Odd numbers are just bad for Apple apparently. iPad 1 didn't have enough RAM, iPad 3 didn't have enough GPU power for that retina display, iPad Air is a combination of the iPad 1 and iPad 3.

I really am super anxious about the next iPad because I am looking to make that *my* next iPad. I've suffered enough with the under powered iPad mini 2, I don't want to go through it all over again with the iPad Air 3. Please Apple, focus on performance, not spec wars. We don't need a higher resolution on our regular iPads, only on the supposed "Pro" version which will be able to handle it...

I agree with you on previous iPads, but not the iPad Air 3 moving to a 3x display. With the rumored resolution for the iPad Pro resulting in a PPI of 264 like all the 9.7" Retina iPads, I do not see them changing the resolution (as its already Retina). In all honesty, I have no clue what the iPad Air 3 will bring. The iPad Air 2 improved on pretty much all the specs, and I don't see them including force-touch until next year (much like they waited on Touch ID).

On a different note, would you still buy the iPad Air 3 if your iPad Mini 2's performance is increased in iOS 9?

 

[XTheLancerX]

I agree with you on previous iPads, but not the iPad Air 3 moving to a 3x display. With the rumored resolution for the iPad Pro resulting in a PPI of 264 like all the 9.7" Retina iPads, I do not see them changing the resolution (as its already Retina). In all honesty, I have no clue what the iPad Air 3 will bring. The iPad Air 2 improved on pretty much all the specs, and I don't see them including force-touch until next year (much like they waited on Touch ID).

On a different note, would you still buy the iPad Air 3 if your iPad Mini 2's performance is increased in iOS 9?

I guess I'm being pessimistic for the @3x thing because if it isn't the case (pretty likely according to you and others lately) I will be much happier with the iPads announcement.

Quite honestly, I don't know what they could do with the Air 3 either.

Also, that is quite a hard question actually. It's too early to tell whether iOS 9 will "fix" my mini 2. It seems like it will help, but not make a world's difference in the end. I'm pretty pessimistic about this as well because I've seen my iPad stutter all the way through it's very first OS, and take a turn for the worse after its first update.

If iOS 9 does make everything pretty much perfectly smooth, though, I will be much less likely to upgrade. However, a family member sort of wants a tablet. They don't want the biggest or the best, just something larger than their phone that works. I think my iPad would be perfect for their use case, and it would be a fine opportunity for me to upgrade. We already essentially share the thing, and they love when they get to use it. When I point out to anybody I know that my iPad stutters and lags, they look at me like I'm crazy, so it's a non-issue for them apparently. I've emphasized that I kinda feel bad just giving them a hand-me-down but they insist that they would be perfectly fine with my iPad so that is why I'm looking at the Air 3.

 

[Andres Cantu]

I wouldn't worry about the iPad Air 3. Even if there is a resolution increase, the fact that it will have an even better GPU would make it run fine (since the one currently in the iPad Air 2 is already overkill).

I can see why you would want a better iPad. Even though most people probably don't notice the lag or don't care, some people like you and I do. Maybe the iPad Mini 4 could be an option for you if it has an A8X (or, more likely an A9), 2GB of RAM, and fixes the color gamut problem.

 

[itsOver9000]

If the A9X is quad-core, then it's Geekbench score could be somewhere around the 5,000's, maybe even reach 6,000! (But not yet over 9,000, sadly)

Soon..

 

[XTheLancerX]

I wouldn't worry about the iPad Air 3. Even if there is a resolution increase, the fact that it will have an even better GPU would make it run fine (since the one currently in the iPad Air 2 is already overkill).

I can see why you would want a better iPad. Even though most people probably don't notice the lag or don't care, some people like you and I do. Maybe the iPad Mini 4 could be an option for you if it has an A8X (or, more likely an A9), 2GB of RAM, and fixes the color gamut problem.

Yeah. I'm worrying myself way too much. With the Air 2's amazing octa-core GPU and triple-core CPU, I think it could even handle an @3x display now. Any improvements added to the chipset after that would make it even better. But you're right, they most likely won't increase the display resolution anyway. That is probably reserved for the Pro.

As much as I love the size of my mini, I feel like I'm ready to move to the full-size iPad. I feel like the mini will always be slagged by Apple as the low-end iPad even though it shouldn't be. I don't feel like I'm ready to accept more generations of iPad mini with inferior technology. And I do feel a little "cramped" lately, the extra screen real-estate is quite welcome.

 

[Andres Cantu]

Yeah. I'm worrying myself way too much. With the Air 2's amazing octa-core GPU and triple-core CPU, I think it could even handle an @3x display now. Any improvements added to the chipset after that would make it even better. But you're right, they most likely won't increase the display resolution anyway. That is probably reserved for the Pro.

As much as I love the size of my mini, I feel like I'm ready to move to the full-size iPad. I feel like the mini will always be slagged by Apple as the low-end iPad even though it shouldn't be. I don't feel like I'm ready to accept more generations of iPad mini with inferior technology. And I do feel a little "cramped" lately, the extra screen real-estate is quite welcome.

Yes, it does seem like Apple will from now on keep the iPad Mini a generation behind in order to sell more iPad Airs. They've done this with many other products before. Still, it's not a problem if you do prefer the larger size.

 

[Andres Cantu]

Soon..

This is hilarious! I can totally see competitors being the "Vegitas" with "scouters" analyzing the future A-chips, and then the A-chips pulling a "Kaioken" with an unanticipated fourth-core, finally reaching over 9,000 in Geekbench. For now, the A9X can be like Goku while the Snapdragon 810 is like an "overheated" Nappa. Even their Geekbench scores match-up with their power levels!

 

[Matheew944]

Yep, I'm waiting for that too. I'm sure it will happen, especially if it can match the S6's score with "only" three cores. Also, I feel like they will use the A9 chip on the iPhone 6s, iPhone 6s+, and iPad Air 3 (maybe iPad Mini 4). Historically, they've never used an AX chip on an iPad Mini, so it would make sense to do what they did for the A7 again in the A9 and take advantage of economies of scale. They could make the A9 in the iPad Air 3 run at a higher frequency like the iPad Air's A7. I feel like they would reserve an A9X chip for the iPad Pro. But I'm sure that the iPad Pro will have 2GB of RAM (albeit LPDDR4), they'll save the 4GB of RAM for a future model (not saying it's the ideal case, just what I think they will do).

4GB of RAM for Ipad Pro 2!!

 

[Andres Cantu]

4GB of RAM for Ipad Pro 2!!

Seems likely.

 

[soupcan]

Seems likely.

It is the only logical thing to do. Anything less than that and there is no point in having a "pro" model designed to do more things. It'll do so much more with the larger screen. Photoshop or After Effects on a tablet? You're going to be running out of RAM really fast with just 2 or 3GB since you also have other apps open in the background.

Apple has to be careful of what they put in it, spec wise. One misstep like not enough RAM and boom, product fails.

 

[Andres Cantu]

It is the only logical thing to do. Anything less than that and there is no point in having a "pro" model designed to do more things. It'll do so much more with the larger screen. Photoshop or After Effects on a tablet? You're going to be running out of RAM really fast with just 2 or 3GB since you also have other apps open in the background.

Apple has to be careful of what they put in it, spec wise. One misstep like not enough RAM and boom, product fails.

I agree. This is their chance to finally introduce a product with proper specs and not take a wait-and-see approach, much like they did with the original iPad.

 

[cynics]

"Power" comes down to usage.

Easy example would be the iPad Air 2's h264 decoder. The Air 2 leaves a lot to be desired. It will only play about 5 minutes of my DVD encodes before the video freezes.

You might say, "well you set your encode level too high". Which technically is true but to put it in perspective my launch day Motorola Xoom (Feb 2011) is easily capable of decoding the same videos the Air 2 struggles and eventually freezes with.

Which as usual illustrates the uselessness of benchmarks vs real world usage. The Air 2 could have a 10,000,000 geekbench score but that is just a relatively irrelevant number.