Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Is the new A7 chip the breakthrough that retina macbooks are waiting for?

S

surjavarman

macrumors 6502a
Original poster
Nov 24, 2007
645
2
Is the new A7 chip the breakthrough in computing industry?

I guess with the 64bit processor and instruction set they can finally port OS X to the A7 chip. They could finally make a macbook pros even smaller and thinner with retina screens without sacrificing battery life.
 
surjavarman said:
Is the new A7 chip the breakthrough in computing industry?

I guess with the 64bit processor and instruction set they can finally port OS X to the A7 chip. They could finally make a macbook pros even smaller and thinner with retina screens without sacrificing battery life.
Click to expand...

Too bad none of the programs I use on a daily basis for work, e.g. C compilers, Lightroom, CAD software, Openoffice, Matlab, none of those actually run on an A7 processor. So yeah I'm sure it's great if you only need to access twitter.
 
and what about every existing program for osx that is compiled for x86? Moving to different CPU architecture is not a simple matter and would be a huge job for software developers.
 
surjavarman said:
Is the new A7 chip the breakthrough in computing industry?

I guess with the 64bit processor and instruction set they can finally port OS X to the A7 chip. They could finally make a macbook pros even smaller and thinner with retina screens without sacrificing battery life.
Click to expand...

I fear it will be the other way around, they'll port iOS to the Mac ...
 
I don't care about a computer thinner than an Ethernet port.

Although I would like Apple to abandon the crappy x86 architecture and make nice machines again, it's not practical.

And the A7 does not have enough power.
 
OS X is already ported to ARM, its called iOS. If your software does not include any assembly code and its all C/Cocoa, building it under a different architecture is as easy as invoking the compiler - unless you do something very weird in the code. That said, the A7 is still very far from the performance levels of Intel CPUs.
 
surjavarman said:
Is the new A7 chip the breakthrough in computing industry?

I guess with the 64bit processor and instruction set they can finally port OS X to the A7 chip. They could finally make a macbook pros even smaller and thinner with retina screens without sacrificing battery life.
Click to expand...
It would be easier to just use Intels new BayTrail or its 14nm successor Arimont.
The A7 chip won't beat those Atoms and it is quite a bit away from an ULV Haswell. That would be some step back in performance.
You can just recompile some code others simply is in frameworks that don't exist in iOS and need some ARM version. Everybody lauds Nvidia's drivers but Intel is known for having the best compilers. That and code optimization for x86 would make a switch to ARM probably quite bad in performance.
iOS is different. It mostly runs one application and no compositing GUI. Not desktop like multi tasking and Windows 8 with Atoms shows how poorly adequate tablet SoCs perform when faced with a full fledged desktop OS.
iOS vs OSX is the difference between Windows 8 modern UI and Windows 8 Desktop.

Those Atoms would also be really cheap compared to Haswell and Intel has the fab advantage which is going to grow now with 14nm around the corner. If Apple wanted to save power, they would use those for OSX. Considering how much of the power the display needs and Intel's Core consistently lowering power consumption the only benefit would be price. Lowering price doesn't seem to be an Apple priority.
 
cube said:
I don't care about a computer thinner than an Ethernet port.

Although I would like Apple to abandon the crappy x86 architecture and make nice machines again, it's not practical.

And the A7 does not have enough power.
Click to expand...

There are no real competitors to Intel if you want desktopish performance and low power consumption in a notebook. We'll see what ARM brings us in the next 5-10 years hopefully :D
 
In short, no. The A7/A7X don't support a lot of the core technologies of the MBP/MBA, such as Thunderbolt (Which is an Intel technology).

ARM has no place as an Intel replacement in laptops, IMO.
 
Purant said:
But they had good reason then and the transition was NOT pleasant.
Click to expand...

I was once involved in a project to migrate an entire OS from one CPU family to another, and it definitely was not pleasant. It was most nasty down in the lowest reaches of the kernel, where code depended on specific hardware oddities. Some applications just needed recompile, but others were a pain.
 
old-wiz said:
I was once involved in a project to migrate an entire OS from one CPU family to another, and it definitely was not pleasant. It was most nasty down in the lowest reaches of the kernel, where code depended on specific hardware oddities. Some applications just needed recompile, but others were a pain.
Click to expand...

*shudder*
 
If we just step back and look at why they went 64-bit in the A7 in the iPhone 5S I think it is pretty clear to all of us that it gives no benefit.

Yes it can handle larger integers but it's not like we were hitting a mathematical wall on the A6 32-bit chip. And the phone isn't sporting 4GB or more memory. Most likely 1-2GB of RAM.

So what is the real reason for them doing this? The only thing I can think of is that this is their first step in a broader plan to eventually transition Macs to ARM. Now I know this sounds crazy today but this isn't a today kind of plan this is them thinking ahead, 3-4 maybe 5 years from now.

And even if their plan isn't to go ARM on their Macintosh lines it does provide a huge amount of leverage over Intel. We all know that Intel is trying very hard to make low power chips for portable devices, it isn't a secret it's all they've been going on about for the past 7 years since Core Duo. But talking about it and actually delivering chips in ARM power envelopes are two separate things and Intel still has a way to go.

Now from my personal perspective I don't care what chip architecture the Macs use. X86, ARM, PowerPC. All I want is it to be fast and the product (if it's a portable) to have great battery life. If Apple feels confident enough in their rate of chip innovation to bring ARM to their Mac product line then I say go for it but I do have my doubts, Intel has been at this game a lot longer and I think they know a lot more about this stuff than Apple will ever know and I most certainly don't want a future where we get 2-3 generations of Macs with ARM processors which kick Intels ass and then a decade of Intel obliterating ARM and us being stuck on ARM like what happened with PowerPC and X86 before the switch.

In any event this is all really far off I think we will see Broadwell, Skylake and Cannonlake in our Macs yet. And hey Intel still has a great line up of technology yet to come with 10nm in 2016.
 
surjavarman said:
Is the new A7 chip the breakthrough in computing industry?

I guess with the 64bit processor and instruction set they can finally port OS X to the A7 chip. They could finally make a macbook pros even smaller and thinner with retina screens without sacrificing battery life.
Click to expand...

No, and don't hold your breath.
 
Azathoth said:
Too bad none of the programs I use on a daily basis for work, e.g. C compilers, Lightroom, CAD software, Openoffice, Matlab, none of those actually run on an A7 processor. So yeah I'm sure it's great if you only need to access twitter.
Click to expand...

OpenOffice runs on ARM. As does Clang and GCC. Matlab has some support for ARM (look for Matlab Coder).

old-wiz said:
I was once involved in a project to migrate an entire OS from one CPU family to another, and it definitely was not pleasant. It was most nasty down in the lowest reaches of the kernel, where code depended on specific hardware oddities. Some applications just needed recompile, but others were a pain.
Click to expand...

Sorry, but what does porting an OS has to do with porting an application? Its a completely different thing. The whole OS X programming experience is designed around circumventing hardware oddities, because you have the rich API to take care of that. Virtually every App Store application should be trivial to port (ad by trivial I mean cross-compiling it). For instance, software tools I write are all developed with 64bit OS X in mind but they compile for linux, windows (both 64 and 32) and ARM linux without a single complaint.
 
There is another angle on this, IMO.

Think beyond consumer products, and look at the potential implications for Apple in their data center spaces. This could be an excellent chip for scale-out in their huge data centers where they have to deliver all of the content of their various electronic storefronts.
 
Quu said:
We all know that Intel is trying very hard to make low power chips for portable devices, it isn't a secret it's all they've been going on about for the past 7 years since Core Duo. But talking about it and actually delivering chips in ARM power envelopes are two separate things and Intel still has a way to go.
Click to expand...
Why do they still have a way to go.
Look at the BayTrail which beats ARM on power efficiency. Better dynamic range and if at all only the A7 cores on the Qualcomm chips can compete in the lowest power states.
Baytrail manages Kabini performance (on the CPU front) while sipping power like an ARM SoC.
Broadwell also allows TDP classes that allow fanless tablets. With average power of about 4-5W and peak TDP of around 8W. And Broadwell is a lot faster too and bigger and more expensive but still.
Baytrail already competes with any ARM SoC in power efficiency.

As sad as the situation is for consumers and the competition, Intel has caught up to ARM already. The only light at the end of the tunnel is the fact that successive fab processes will be harder and harder so once Intel runs up against the wall and cannot keep the pace the ARM front should catch up on that front. Until that happens Intel has earned a lot and probably the best R&D money can buy. The x86 decode also is less and less of a power efficiency problem and at 22nm we are already where it doesn't matter compared to ARM instructions.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back