Seems more plausible. We'll see, if Apple does use the A5 in a Mac I'm sure they'll explain their reasoning well enough.
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 Testing an ARM (A5) Powered MacBook Air?
- Thread starter MacRumors
- Start date
- Sort by reaction score
Seems more plausible. We'll see, if Apple does use the A5 in a Mac I'm sure they'll explain their reasoning well enough.
That is because current ARM based products are designed for low power not performance. My comment was that ARM has far better POTENTIAL performance than X86.
Darwin (which is the very low level core system of Mac OS X) runs on Intel Macs as well as on ARM iPod Touch/iPhone/iPad.
THe Darwin Mach Kernel is a key to make that possible.
So, the difference between Mac OS X and iOS is Cocoa on the Mac and Cocoa Touch on the iOS devices.
So, from that point it is possible to run Mac OS X on any iOS device (despite the occasional need to recompile some specially optimized parts / components for that very cpu).
A MacBook Air as a computing device for the most common tasks would not need more than an ARM CPU. But users with stronger needs would turn to the MacBook Pro's then anyway - as they already do.
But an ARM CPU gives tremendous battery life.
However, even if think this is interesting I still do believe, that Apple would prefer to have an ultra low energy consuming macbook air, which implies that they would need to change to a low energy display as well. (But, wheren't there already rumors about apple being interested in color OLED?)
If you don't play racing games or anything else with high speed moving graphics, then a macbook air with an OLED and an ARM CPU might be for you: read, write, draw - just what you'd do with a netbook.
But to be honest, where should such arm/oled macbook air be priced? Shouldn't be cheaper than an iPad but must be cheaper than the low end MacBook Pro. That sounds expensive then.
Funny how you guy don't like it when Apple Head and purist speak the truth about Intel and desire something that isn't a dying dog, but once talk of an ARM Macbook Air comes up they seem to be all over it. To you guys I say eat a pie.
Now, on to the best news I have heard all week. If they can bring this out by next year, maybe just maybe we can make people forget about the dark history in time that will be known as the big Intel mistake.
Now, on to the best news I have heard all week. If they can bring this out by next year, maybe just maybe we can make people forget about the dark history in time that will be known as the big Intel mistake.
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)
Not to mention Apple will be implementing Thunderbolt on all sorts of devices. Many of those devices will have ARM or other processors embedded in them. Beyond that Apple would likely want to make sure third party developers have the chips available to them to make low cost accessories.
In this regard I would not be surprised to find a Thinderbolt port embedded right in the next "A" series processor from Apple. The ability to put Thunderbolt right into the SoC would be huge for Apple.
Bear said:
Not to mention Apple will be implementing Thunderbolt on all sorts of devices. Many of those devices will have ARM or other processors embedded in them. Beyond that Apple would likely want to make sure third party developers have the chips available to them to make low cost accessories.
In this regard I would not be surprised to find a Thinderbolt port embedded right in the next "A" series processor from Apple. The ability to put Thunderbolt right into the SoC would be huge for Apple.
Could you explain to me how Intel was a mistake. I'm not attacking you, I honestly would like to hear your opinion.
Perhaps but we're talking about a current ARM processor being used in an MBA and the implications of that, not some pie in the sky processor that doesn't exist.
Intel is not about to give up on the X86 instruction set, and given how they're still banging out faster and faster processors I don't see the need either.
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)
Nice! This is another item I can reference when trying to combat the "no apps use OpenCL" BS. To put it mildly OpenCL is getting a lot of traction in industry.
The problem is many people have unrealistic expectations for GPU exploitation. GPUs are only useful when the task at hand works well sighing the structure of a GPU processor.
Mr. Retrofire said:
Nice! This is another item I can reference when trying to combat the "no apps use OpenCL" BS. To put it mildly OpenCL is getting a lot of traction in industry.
The problem is many people have unrealistic expectations for GPU exploitation. GPUs are only useful when the task at hand works well sighing the structure of a GPU processor.
Mind if I answer that one? Apple going to X86 was not a mistake, X86 currently has the best bang for the buck because a great deal of money is being dumped into making x86 fast. That is not to say X86 is not very flawed.
X86 is CISC based architecture with large numbers of slow executing instructions. X86 gets around this by translating all the CISC instructions into faster executing RISC instructions. This takes up big portions of the core. The translated instructions are also not as efficient as code that was originally written with the native instruction set.
Because ARM does not need the circuitry to translate CISC instructions to RISC, they have more room for cache as well as registers and instruction reordering and speculative execution (Just like the Texas court system). ARM also has the potential to run at much faster clock speeds than legacy X86.
It is perfectly reasonable to expect 64 core ARM chips running at a faster clock than X86 in the near future. It is also perfectly reasonable to expect that those chips will be designed and built by Intel.
duh?
This is pretty obvious.
It's be pretty cool it they could get two in there. Especially when the dual/quad comes out. A6? is A5 dual core? Can't remember.
This is pretty obvious.
It's be pretty cool it they could get two in there. Especially when the dual/quad comes out. A6? is A5 dual core? Can't remember.
This could be the first Mac that only works with applications from the app-store, the Mac that isn't a truck.
I appreciate the response, so I assume ARM is less speedy then x86 due to its lack of attention to development? It sounds entirely more efficient design wise. I know ARM is known for it's battery life capabilities, will speeding that up compromise it's battery potential?
Great addition about the Texas court system, couldn't stop laughing.
The first step to building an OS based around the next generation of ARM is to get it running on the current generation. As to Intel, if they could get 16 times the performance or more for less cost, they would be morons not to move to a better architecture.
If you ramp up clock speed, add more cores or more cache, you will increase power consumption. That is not to say, high performance RISC will not use less power than a CISC chip of the same performance.
Your initial post said he was most likely using disk intensive applications, which is not the case and not why he saw an improvement. The reason he saw an improvement was the read/write speeds between a regular HDD vs an SSD. It's pretty simple. Just about every program accesses the HDD when you launch it, and many continue to access it after the application has launched. This is where he's seeing the improvement. It has nothing to do with a program having heavy disk usage, moderate disk usage, or even light disk usage.
Wirelessly posted (Mozilla/5.0 (iPhone; U; CPU iPhone OS 4_3_3 like Mac OS X; en-us) AppleWebKit/533.17.9 (KHTML, like Gecko) Version/5.0.2 Mobile/8J2 Safari/6533.18.5)
Just because this prototype was built on an AIR it does not mean that is where the final eager platform will be. Going to ARM on an AIR would simply result in to many lost customers.
However let's imagine that Apple was to target the low end education market aggressively. An ARM based laptop would allow them to place a laptop with extremely long battery life into a glass room very cheaply. The end product might very well be a MacBook though I think it is more likely to be an iOS device. IOS devices are great for the k-6 market as the administrative costs would be low and the machines can be easily locked down. The trouble with iOS of course being the touch screen interface
In this market performance wouldn't be an issue. Even then a laptop likely could support a hotter running processor though a fanless machine if preferrable. If Apple can do this for $400 a crack there will be lots of interest.
In any event I agree with the many here, AIR needs more powerful hardware not less powerful.
Just because this prototype was built on an AIR it does not mean that is where the final eager platform will be. Going to ARM on an AIR would simply result in to many lost customers.
However let's imagine that Apple was to target the low end education market aggressively. An ARM based laptop would allow them to place a laptop with extremely long battery life into a glass room very cheaply. The end product might very well be a MacBook though I think it is more likely to be an iOS device. IOS devices are great for the k-6 market as the administrative costs would be low and the machines can be easily locked down. The trouble with iOS of course being the touch screen interface
In this market performance wouldn't be an issue. Even then a laptop likely could support a hotter running processor though a fanless machine if preferrable. If Apple can do this for $400 a crack there will be lots of interest.
In any event I agree with the many here, AIR needs more powerful hardware not less powerful.
I could sorta see some kind of hybrid where the core OS and all Apple supplied software runs on the A5 so you could be editing your Keynote presentation, surf the web, write your E-Mails etc. on the power saving chips. For photoshop, video/audio editing, or games, it could switch on the Intel CPU (and a real GPU other than the one on the G5) and run under some sort of compatibility mode.
Sort of how Intel Macs can run PPC programs using rosetta. Let's face it, most consumers only run rudimentary applications on their shiny Apple laptops and would love it if they had double the runtime on the same battery.
Sort of how Intel Macs can run PPC programs using rosetta. Let's face it, most consumers only run rudimentary applications on their shiny Apple laptops and would love it if they had double the runtime on the same battery.
Yes it does. If your program is only sustaining 10 MB/sec of read/write I/O, then a normal HDD is plenty and you won't see a difference with a SSD.
Again, you're reading way too much into this. I simply pointed out that he saw the improvement because in his case, he had a workflow based around disk I/O, which does not apply to everyone.
You're going off on a tangeant again btw. We're saying the same thing, you're just applying the "SSD is best!" in a more general sense and I'm simply realistically looking at SSD advantages and not claiming it's a miracle cure.
Aren't there companies selling big servers with like 256 ARM CPUs?
I think I've read stuff that said ARM is *far* more powerful than x86 per watt.
So if a typical x86 CPU is 10 times more powerful than an ARM CPU, you could put ten ARM CPUs in a small box, get the same performance as the x86 CPU, yet still use *less* power than the x86 CPU.
Maybe the A5-based MacBook Air is like a 16-core ARM or something.
Also, is A4/A5 the name for ARM+PowerVR specifically, or just ARM+Graphics? Couldn't Apple load these with ARM + NVidia GPUs? That would help a lot.
7 pages so not enough time to read through all 7 but even if it was posted before, it warrants posting again:
Apple made test intel machines before they moved, and KNEW they would be moving, to intel. Apple isn't stupid and hedging their bets. This might see the light of day and may not. Just because they made it doesn't mean it will come to pass.
Apple made test intel machines before they moved, and KNEW they would be moving, to intel. Apple isn't stupid and hedging their bets. This might see the light of day and may not. Just because they made it doesn't mean it will come to pass.
Yes it does. If your program is only sustaining 10 MB/sec of read/write I/O, then a normal HDD is plenty and you won't see a difference with a SSD.
Again, you're reading way too much into this. I simply pointed out that he saw the improvement because in his case, he had a workflow based around disk I/O, which does not apply to everyone.
Every application you launch uses the disk... anything that's not already in memory uses the disk. Anyone who uses Photoshop regularly is going to notice it opening in 3 seconds rather than 10 seconds. I get the feeling you don't understand what disk heavy is, because these are not disk heavy applications. Anything you launch is at least partially loaded into RAM, that's accessing the disk right there. Therefore anything that accesses the disk for more than 50MB is going to be noticeably more responsive. Any feature within the program that you click that has to access the disk will be faster, page filing is used quite frequently since not everything can be loaded into RAM... this another feature where the SSD gain an advantage. Boot times increase. I could go on and on all day, but it's come to my attention that you just like to breeze over facts I'm throwing at you because you're someone who just can't admit to being wrong. So I'm not going to bother anymore.