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.
Apple's A7 Processor Truly 'Desktop Class', iOS Apps Don't Take Full Advantage
- Thread starter MacRumors
- Start date
- Sort by reaction score
Source for S5 benchmarks? Here are some for the HTC One M8
http://www.anandtech.com/show/7893/the-htc-one-m8-review/7
The 5c isn't that high up except for SunSpider
I don't think they'd release a laptop to do an iPad's job.
If they do ship a laptop with an ARM processor, it won't be until it's powerful enough to do more than web and email.
For those on the go, like photographers perhaps, the unprecedented battery life will be a huge benefit.
Not really.
ARM only has a power advantage over Intel when completely idling, or performing an exceptionally lightweight task like playing music. Tablets and phones happen to do that extensively, so its beneficial for them.
Laptops however tend to either be active, or completely have the CPU shut down (sleep mode).
Its well known among hardware experts that the performance-per-watt is much worse even in the most efficient ARM implementations. Furthermore, ARM's power efficiency does not scale to higher performance levels well.
For that reason, an ARM based Macbook would typically see worse battery life even in baseline apps like Safari. If an emulation layer was included to maintain x86 compatability; power efficiency and performance would be utterly decimated in those apps.
Last edited:
not in any iDevices.
If you have 8 core processor 4GB ram in a phone what more you will do? Its pointless to have PS or xbox level game on such a small device (even if they could) its just over kill, you wont see any of the details anyway.
But as the screen becomes bigger then the productivity increases and so the demand for better sharper looking games. Then we'll need more ram and processor, eg ipads
I think its more economical for them to mass produce chip to put in all 3 devices, its also better for them build more fine tuned OS, perfectly matching the hardware, hence more optimized and smooth OS in less ram and less GHz
i guess its not always the potential use of something but also whats more economically viable.
64-bit is in similar league, on the phone its pointless, but as whole, software, driver and app development process gets highly streamlined.
Maybe it's because those high end devices are 4.7"+ and have a bigger battery to compensate for the more RAM??
EDIT**Sorry I see you have answered that quote already...I just never understand why you need to prove you're right over and over. You maybe right about the RAM issue as my iPad Air reloads quite frequently too, but it's the way you come across, if someone doesn't agree with you get over it, there opinion is just as relevant as yours.
As someone has previously stated this cannot be proven once and for all so move on..
Last edited:
Again:
1, in all my posts, I avoided the comparison to Android as I know very well that it's far from efficient and can lag / stutter / etc. This has also been explicitly pointed out in all my posts too.
2, "Who cares if iOS takes up more RAM to browse the web when it still runs faster than competitors?" - you ask. This has been answered by me several times in this #. People that hate the background tabs to be reloaded, for example. Which is a major issue, particularly on the, theoretically, as Web browsing-friendly advertised retina iPad. (Not so on the iPhones as, the screen buffers being much smaller, the same Web page is allocated about 3 times less RAM than on a retina iPad.)
----------
Again: as I've also pointed out many times, it's mostly the iPads that suffer from the lack of RAM because of the far higher screen resolution meaning far higher memory usage in Safari and the like.
And, as I've also explained above, iPads have almost an order of magnitude larger / more powerful battery than iPhones, meaning the additional power usage of an additional 1GB of RAM would definitely be negligible. I would have expect Apple to put 2GB RAM in at least the iPads. After all, RAM shortage problems have always been rampant on iPads (not only the iPad 1).
This is simply an engineering choice forwarded by the marketing department.article said:
Shoot the marketers and go into engineering and production!
Rocketman
No one forces you not to love tab reloads in Safari or minimized app kills in the OS. All the wonders of 1GB RAM, particularly on the more memory-hungry iPads. With 2GB of RAM, this wouldn't be that big an issue. And, again, 2GB of RAM certainly wouldn't have resulted in any battery life decrease. After all, Surface Pro 2 users with 8GB of RAM (that's eight times more than in iDevices!) don't complain about unbearable battery life either. Neither does the AnandTech review: http://www.anandtech.com/show/7478/microsoft-surface-pro-2-firmware-update-improves-battery-life
You don't seem to follow professional iOS game reviews written by people comparing iOS ports of desktop / console titles, do you? For example, the above-mentioned XCOM: Enemy unknown (probably the only AAA quality, almost-direct console port on iOS) reviews all pointed out the iOS port had to lower the texture quality because of the significantly lower RAM. That is, the difference in texture quality is certainly visible - as opposed to what you stated.
For example: http://toucharcade.com/2013/06/19/xcom-enemy-unknown-review/ :
" The more I look at the game, the more I'm noticing bits and pieces of asset compression, which makes it look like a "low res" port. "
etc.
They already have some huge margins on all their products. Including another 1GB of RAM at least in the iPad (as it's suffering from the lack of RAM a lot more than the iPhone) would have cost Apple some $5. It's a win for them - and a huge loss for users that have to put up with constant tab / app reloads.
Why would it? It does have the potential. While, due to the complex nature of my apps (and some libraries, for example, ZXing, only supporting 32-bit), I haven't ported my own AppStore apps to 64-bit, I'm absolutely certain there would be a performance boost in, for example, my pathfinder algorithms when properly optimized for 64 bits. The same (64-bit optimized & compiled apps do result in a performance boost) is stated by the AnandTech reviews.
I think you are mixing a few things.
First, yes, desktop CPUs are at a speed where year-over-year incremental improvements are hardly noticeable. That wasn't the case five or ten years ago, but it certainly is now. You have to specifically look to notice the performance difference most of the time.
But, "desktop class CPU" does not equal "desktop class system". This is a phone. The A7 CPU may well be able to compete with Intel's chips in terms of raw performance. But, the rest of the system around it is nowhere near what you will want to be using 3-5 years from now. The memory constraints have been talked to death on this board, but that's not the end of the compromises put in place to get this system to fit into a little slab of aluminum and glass that in turn fits intone the palm of your hand. All those other compromises are seeing rapid advancement, and so year over year the raw overall performance of the device can and will improve.
Putting the A7 into a desktop would be an interesting experiment. I suspect someone is already carrying it out (or has already done it and drawn conclusions based on it). I would expect performance to degrade slightly, but power consumption to improve significantly. The real problem is that this wouldn't be competing with i5s and i7s, but with Intel's low-power lines, where performance might be comparable (or better) and power consumption might be comparable (or better), but who today would look at the i5 and i7 and an Atom-based machine side by side and think, "Yeah, I'll go with the Atom"?
Yes but people are also wanting laptops to be always on, always connected and always up to date. Which is also where ARM would shine, especially considering Apple already has the skill and code to manage such connections on ARM. ARM however doesn't scale to well once you get higher demand uses.
So why can't Apple have the best of both worlds?
Have both processors in one machine, one always on the other on by demand.
So the Surface Pro has 87% of the battery life of an iPad 4 while surfing WiFi (low RAM use) and 57% of the battery life while playing video (heavy RAM use). Although, I don't see any indication that this is the 8GB model in Anand's review (I can assume he got the highest-end model, but it isn't stated that I can see).
Yeah, that sounds about like what I'd expect if extra RAM was consuming a lot more battery. Note that the battery capacities here are roughly equivalent (the Surface is slightly smaller at 42 vs 42.5Wh). Good job, Microsoft! Maybe Apple will up its game next year.
Oh yeah, it did, a week after the Surface 2's firmware update. With 25% less battery (32.4Wh instead of 42.5Wh) the iPad Air gets a smudge more battery life on each of Anand's scenario tests. So, add another 30% to those numbers from the Surface Pro 2 to compare contemporaries.
As Anand points out with the comment about Sony's Vaio for the first time ever finally being comparable in power use while idle to Apple's, Apple highly values power efficiency. Ever-increasing power efficiency has been a core tenet of the Apple hardware team since before the Intel transition (and it in fact drove the Intel transition to happen when it did). I would fully expect them to put heavy emphasis on power efficiency.
On the other hand, if the loss of state in Safari bugs you, please send Apple feedback. Let them know that their customers are unhappy with the effects of having only 1GB of RAM (but if you want your complaint to be taken seriously don't at all mention that the phone has only 1GB of RAM; any complaint which delves into technical specs instantly goes to the bottom of the heap). Letting Apple know what pain you are obviously feeling is the only way Apple will value the benefits of having more RAM higher than the benefits of not having it (and, yes, one of those is an extra buck or so per device, but I don't think that Apple would favor customer complaints over spending that extra buck).
The two devices aren't really comparable, as the Surface has a much more power-consuming x86 CPU. I couldn't come up with a better example of RAM's power usage's being dwarfed by other components than the SP2 simply because non-x86 tablets only go up to 2GB RAM in the test.
But even with that CPU and even with 4 (or 8?) Gbytes of RAM, the battery life of that thing isn't much worse than that of the iPad 4. This clearly shows not even 4 (8?) Gbytes of RAM consume THAT much power. Again, it's the x86 CPU (and probably the, from time to time, active cooling) that consumes the most power.
Taking all this into account, the SP2 has pretty good battery life. Back in time, when, in 2003, I purchased the HP TC1100 (the first Wacom tablet from HP), it only had some 2 hours of battery life.
If the two aren't comparable, why did you compare them?
And yes, the Surface Pro 2 does get battery life close to that of an iPad 4 when not engaging in memory-activity-intensive tasks. When engaging in memory-activity-intensive tasks, it gets about half the battery life of the iPad 4. That is a year older than it.
On its near-contemporary iPad Air, Apple gets significantly better battery efficiency.
But, of course, they aren't comparable. I agree. This comparison of yours has told us nothing at all about how much power RAM uses. If you go back a screen or two you'd see specific citations detailing how much power RAM activity actually does take, measured on real devices (albeit from several years back; we can assume RAM has gotten more efficient per MB since then).
Again, "word on the street" from actual hardware engineers is that RAM does figure significantly into the overall power envelope calculation. That doesn't mean that you always must skimp on RAM, but it is one of the larger "negative" factors.
According to the first "bill of materials" I could find for the iPhone 5s (aftermarket guessing game about how much the individual parts cost), the RAM is pegged at $11, and 16GB Flash at $9.40. http://technology.ihs.com/451425/gr...m-and-manufacturing-cost-ihs-teardown-reveals
From the same site, the Samsung Galaxy S5 pays $28 for both its RAM and flash: http://press.ihs.com/sites/ihs.newshq.businesswire.com/files/2013-03-19_Samsung-S4-Table1.jpg
Assuming the numbers are actually consistent, I'd guess that the difference in just the 1GB vs 2GB part is probably $7.60.
Looking at that Apple BoM again, I see they also costed out the in-box throw-ins at $7. It seems like there are other places Apple could have cut corners to save $8 per device. Heck, they could have gotten about half that back by just going with slower memory rather than LPDDR3.
Does cost go into the calculus? Obviously it does. I just don't see it as the driving factor behind why the iPhone has 1GB of RAM this year.
To answer your "but why is everyone else doing it?" question before you ask it again, "everyone else" sells their phones by the specs as much as anything else. When you are selling by the specs you get perverse economies which make inflating those specs more important than actually delivering the best product possible. See Intel, late-90s/early-2000s.
It's mostly 'bout the Benjamins
More RAM is always a good thing, don't matter what its in. But RAM costs money.
Unfortunately, the iDevices aren't meant for proponents of permaculture. The phones are virtually free to consumers(purchase subsadized by TelCos), and the tablets in low trim go for less than most people spend on clothing in a month. By design, people use the phones for 2 yrs, and iPads for slightly longer.
I suspect that the vast majority of iPads are essentially e-readers, portable media players etc.
If Apple gave buyers the option of BTO with 2, 4, or 8GB with $50, $100 & $200 dollar price increases, I doubt there'd be many takers.
Gaming pushes hardware to its limits more than any other use, and largely drive consumer decisions to upgrade. It seems to that the state of handheld gaming is now approaching if not exceeding the limits of the current specs.
I suspect that the Xmas 2014 iteration of iPad will ship with 4GB, with no option for future-proofing, because it just doesn't make sense on devices that trade for less than a weeks wages.
It is anti-consumer for sure, but it is what it is.
More RAM is always a good thing, don't matter what its in. But RAM costs money.
Unfortunately, the iDevices aren't meant for proponents of permaculture. The phones are virtually free to consumers(purchase subsadized by TelCos), and the tablets in low trim go for less than most people spend on clothing in a month. By design, people use the phones for 2 yrs, and iPads for slightly longer.
I suspect that the vast majority of iPads are essentially e-readers, portable media players etc.
If Apple gave buyers the option of BTO with 2, 4, or 8GB with $50, $100 & $200 dollar price increases, I doubt there'd be many takers.
Gaming pushes hardware to its limits more than any other use, and largely drive consumer decisions to upgrade. It seems to that the state of handheld gaming is now approaching if not exceeding the limits of the current specs.
I suspect that the Xmas 2014 iteration of iPad will ship with 4GB, with no option for future-proofing, because it just doesn't make sense on devices that trade for less than a weeks wages.
It is anti-consumer for sure, but it is what it is.
Stacked die. Any increase in the RAM on top of the A7 might have made the 5s thicker or cut into the battery size, which might have decreased sales more than any potential performance gains from 2GB might have increased them back.
Surely the point at hand isn't how to remove the bottlenecks from the powerful A7 processor, but why such a processor would be in an iPhone if there are no real world applications that warrant that sort of power. We're already through half of this phone's cycle and supposedly nothing is taking advantage of it's power.
It's a nice willy waving point, but until such time as a use for that power comes along, that's all it is. And you're paying for it in that there was little else in the way of selling points for the 5S.
Apple have no doubt done this because there are bigger plans for the chip and they'll reap huge rewards for stuffing it in every device and achieving economies of scale.
It's a nice willy waving point, but until such time as a use for that power comes along, that's all it is. And you're paying for it in that there was little else in the way of selling points for the 5S.
Apple have no doubt done this because there are bigger plans for the chip and they'll reap huge rewards for stuffing it in every device and achieving economies of scale.
Can we get the trolls in this thread to stop using the term fanboy? I'll get a warning for this post while they continue their passive-aggressive sniping at people, and it's really ******** annoying. Grow up guys, ffs, and try to have an adult conversation.
On topic: Err, yeah, nice chip.
On topic: Err, yeah, nice chip.
But, in this case, the question is not that of increasing the RAM size for the sake of increasing without any gain, but to vastly mitigate the problems caused by the current RAM configuration: Safari tab reloading and the like. Which certainly affect a lot of people. This very forum is also full of people complaining about constant tab reloads in all Web browsers (except for the much more memory-friendly, albeit mostly static, Opera Mini, which uses its own renderer.)