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.
iPad computational power
- Thread starter MacBH928
- Start date
- Sort by reaction score
sammich
macrumors 601
Take this list:
http://www.primatelabs.ca/blog/2011/10/iphone4s-benchmarks/
And compare it to these lists:
http://www.primatelabs.ca/geekbench/pc-benchmarks/
http://www.primatelabs.ca/geekbench/mac-benchmarks/
http://www.primatelabs.ca/blog/2011/10/iphone4s-benchmarks/
And compare it to these lists:
http://www.primatelabs.ca/geekbench/pc-benchmarks/
http://www.primatelabs.ca/geekbench/mac-benchmarks/
ChristianVirtual
macrumors 601
Ok, the numbers looks awful; but I'm still impressed with apps like infinity blade or x-plane. Amazing what the ipad2 can deliver here in real life and some good programmer.
sammich
macrumors 601
I can see why you could say they're 'awful' but you have to remember that these are those are CPU benchmarks, not GPU (the GPU in the A5 made a bigger leap than the CPU did going from the A4).
You have to consider what these processors are designed for. You can't compare these miniature, low power envelope CPUs with their desktop parts. Phone CPUs typically use <1W of power when in use, while laptop/desktops use on the order of >10W.
I still consider it amazing that my iPhone 4 is basically on par with my 12" 867mHz Powerbook.
ChristianVirtual
macrumors 601
You are right; I fully understand that the CPU are are total different animals.
I missed to make the word awful more supporting my intention that we should even not compare just the benkmark results with regular desktops numbers.
And that the real life experience is the relevant one telling us: the iPad / iPhone has enough power for its purpose (and the purpose is not rendering 3D models or calculate weather forecast).
I suppose you could compare the over-all power (so both graphical and computing) to those big desktop personal computers in the early 2000s.
You must not forget that the A5 and A4 processors are very small and extremely low power chips compared to these computers. Also, they don't require cooling like a notebook.
--------------------------
So if you really want to compare it to desktops: in terms of power, you can compare it with the parts available since the early 2000s. In terms of size, energy usage, cooling and the performance it delivers you could compare it to the distant-future. 🙂
You must not forget that the A5 and A4 processors are very small and extremely low power chips compared to these computers. Also, they don't require cooling like a notebook.
Opening Safari on the iPhone already used about 1.9 W. Just opening. Notebooks and desktops use way more than ">10 W" (I realize '>' means anything above 10 W). Just a small comparison: one of Intel's cheaper mobile i7 chips, like the i7-2670QM, can use up to 45 W alone. That's without the display, GPU or anything else (like WiFi and bluetooth). Desktop PCs use even more.
--------------------------
So if you really want to compare it to desktops: in terms of power, you can compare it with the parts available since the early 2000s. In terms of size, energy usage, cooling and the performance it delivers you could compare it to the distant-future. 🙂
anthonymoody
macrumors 68040
Ipad2 is more or less equivalent to a G4 PowerBook.
http://ipadalone.com/2011/03/15/ipad-2-as-powerful-as-a-g4-powerbook/
http://ipadalone.com/2011/03/15/ipad-2-as-powerful-as-a-g4-powerbook/
This may be true on paper, but I still use a G4 PowerBook and I can tell you that the iPad runs far more smoothly and quickly. There's absolutely no contest. And my PowerBook has more RAM than my iPad.
But your powerbook is probably running more intensive software. for example, we nee to make the Software running on your powerbook , run on the ipad then we can make a comparison.
OR we can make the software running on the ipad, run on your powerbook and see if it is just as smooth as the ipad or not.
You can hardly compare different processor architectures in terms of processing power. ARM and x86 have completely different instruction sets. ARM is a RISC processor, which makes common instructions A LOT faster. This, combined with the fact that iPad relies a lot on its graphical power, makes it feel a lot snappier than the powerbook that's been mentioned. As long as we don't go encoding video or other things that need advanced instruction sets on our iPads, the RISC architecture will almost always have the advantage.
anthonymoody
macrumors 68040
Of course guys. At the end of the day it's all about the total user experience. It's what's made iOS devices so insanely successful. And frankly they've changed our expectations of how stuff should work.
What I will say about 'power' though is that regardless of clock speed, or other testing and measurement criteria, I used to spend waaaaaaaaaaaay more time - cumulatively - "waiting around" while using my Mac (starting up and shutting down, opening and closing apps, opening and closing files, occasionally intense recalculations of spreadsheets, etc.) than I ever have on my iPads. The overall usage experience is far better (and don't even get me started about how it feels when I sit down to a windows machine...yikes).
What I will say about 'power' though is that regardless of clock speed, or other testing and measurement criteria, I used to spend waaaaaaaaaaaay more time - cumulatively - "waiting around" while using my Mac (starting up and shutting down, opening and closing apps, opening and closing files, occasionally intense recalculations of spreadsheets, etc.) than I ever have on my iPads. The overall usage experience is far better (and don't even get me started about how it feels when I sit down to a windows machine...yikes).
What if we use ARM processor for common instructions and x86 for video encoding on a laptop with SSD?
Will that be the fastest combination ever?
If it was possible: yes. It is however (to my knowledge) impossible.
1: no desktop OS is meant to run on ARM (except the forthcoming W8, but that'll only support metro, not the full desktop).
2: it's impossible to run 2 completely different processor architectures as competing CPU's.
The A5 chip contains a video encoder too - which probably makes it as fast as the intel chip.
Remember that the A5 is a *system* on a chip (SoC), where the the intel chip is just a CPU. The A5 would be comparable to a CPU, motherboard, graphics card, sound card, video encoding card, camera control board, and probably more. In a tiny package that uses trivial amounts of power. They're very hard to compare!
The other thing is the software. Unless you're running the *same* software on both, any comparison is pretty meaningless. As an example, take 2 PCs, a new low-end netbook, and a windows 95 box from 1995. The netbook is probably 100x faster according to the benchmarks. But load MS Word on them - word 6 on the old box, and 2010 on the netbook. I bet the '95 PC is faster to use.
Same with the iPad and a desktop. The software is designed for it, so it naturally runs well. If you ran desktop software on it you'd find it horribly slow.
Remember that the A5 is a *system* on a chip (SoC), where the the intel chip is just a CPU. The A5 would be comparable to a CPU, motherboard, graphics card, sound card, video encoding card, camera control board, and probably more. In a tiny package that uses trivial amounts of power. They're very hard to compare!
The other thing is the software. Unless you're running the *same* software on both, any comparison is pretty meaningless. As an example, take 2 PCs, a new low-end netbook, and a windows 95 box from 1995. The netbook is probably 100x faster according to the benchmarks. But load MS Word on them - word 6 on the old box, and 2010 on the netbook. I bet the '95 PC is faster to use.
Same with the iPad and a desktop. The software is designed for it, so it naturally runs well. If you ran desktop software on it you'd find it horribly slow.
Proof please? It would be downright ridiculous to put encoding specific hardware on the A5.
I guess you'll find plenty of proof in the developer docs. It's definitely there - I write video and photo apps and use it quite extensively 🙂
I'd say it makes complete sense to put it in there too. The iPad + iPhone have cameras that record HD video - the iPhone 4S records 1080p even. They use H264 video because of the smaller file size, but that format is pretty heavy to decode and encode. The CPU in these devices is pretty weak, in fact the A5 CPUs have pretty much zero chance of encoding 1080p h264 video at 30fps. The only way to do it is to have a hardware encoder.
Menel
Suspended
It's quite typical. No specific specs, but you can see the TI OMAP has it. Which uses the same CortexA9 and a similar though lesser PowerVR GPU.
http://www.ti.com/general/docs/wtbu...ateId=6123&navigationId=12843&contentId=53243
marzer
macrumors 65816
Yes it's completely possible. Just because no one has actually implemented such a configuration doesn't make it impossible. Not much different than off loading graphical work to a GPU.
I was going to post the same thing. I have a 1GHz iBook G4 and however they might compare in raw computational power, my iPad 2 is superior for just about any task imaginable. Bring the GPU into the equation and the iPad is running laps around any G4 notebook.
anthonymoody
macrumors 68040
Exactly. It's all about the overall user experience. This is something apple figured out long ago, and which spec fanboys continue to fail to understand.