For your bullet points under the "looking at the facts" section, I think that's a clear indicator - including CPU moves made in the past - that Apple will always factor in all possibilities, and consider anything that works best for their road map. As well as not being pigeon holed by working with a sole chip maker. I think we can expect more hybrid designs - as with what went into the new Macbooks - in coming months and years.

I'm interested to know where that MacBook Nano concept came from. It's a neat mockup, with one major flaw: for something that would fit in the netbook category, Apple would never include a disc drive.

Apple can simply do the following:

1. BUY NVIDIA.

GPUs give an enormous boost in computational power when used for general purpose needs. Apple's Grand Central software in Snow Leopard takes advantage of this.

nVidia has already announced that they would get out of the chipset business.

But nVidia as a subsidiary of Apple can continue to make custom chipsets for Apple.

This would allow nVidia to be a competitor in the general GPU market to AND. - keeping its graphics card business.

But it would take advantage of nVidia's expertise in chipsets to do custom hardware for Apple.

Combine this with the PA Semi team, this would allow Apple to create custom chips and other hardware to differentiate it from the usual Intel crowd yet pull away from them in speed.

Imagine a custom GPU for Apple. This could not be cloned.

Mac OS X would not then be easily used on cloned hardware if its software worked specifically with custom chips.

What gives? Can someone explain the Snow Leopard+PA Semi plus+Papermaster works in conjunction with Intel compatibility. I don't get it. How can PPC architecture and Intel work together on the same device.
I don't believe Apple will put two chips in one machine. I cant tell why you would purchase PA semi's except to say "PORTABLE"
As I understand it the current Intel's can run Windows, but Apples recent moves indicate some extremely low power portable device that could run Snow Leopard.

I do think Steve final move as CEO is to make the iPhone your only computer. You heard it here first. You will be able to use it as a phone. Take it home and doc it to some multi-core cradle and use it as your PC at home with your files in the cloud (AKA Me) all the time.

Quote

Guest wrote:

I do think Steve final move as CEO is to make the iPhone your only computer. You heard it here first. You will be able to use it as a phone. Take it home and doc it to some multi-core cradle and use it as your PC at home with your files in the cloud (AKA Me) all the time.

Yeah, well the Windows 7 videos suggest this is what MS is up to, except they aren't going to need the phone to dock anywhere, you just use your Cloud account on any device you come to (with Windows 7, of course).

I still object to having my data in the Cloud on anyone else's machines. If I can keep it on my own servers, then this could be one of the greatest ideas. So I don't want MS to have it, not Apple in their Mobile Me service either.

I do agree, however, that Apple is going to want you to use their iPhone for everything. That's the direction technology is going. But I think laptops will stick around for quite a while, due to the keyboard issue (which is why Steve claimed to not want to build a Tablet PC, remember?).

The bit you missed was this imaginary PPC design still being able to run Windows OS, or any of that legacy code that still has deep Intel hooks, such as gaming. What would have to be done is an incredible increase in Cocoa programming, which is still a bit of a hurdle ahead, although on better feet than before.

A bit you missed was however, there is a lot of excellent gaming code that is written for PPCs however. No one ever talks about an XBox possibly being able to run Mac code, but isn't there a PPC inside of an Xbox?

There is no way Apple is going to buy Nvidia. It already bought PA Semiconductor. Apple can already design custom based chips. Apple was an active partner in chip design with both IBM and Motorola. Nowadays, people design the architecture, and outsource the building.

Quote

jameskatt wrote:
Apple can simply do the following:

1. BUY NVIDIA.

GPUs give an enormous boost in computational power when used for general purpose needs. Apple's Grand Central software in Snow Leopard takes advantage of this.

nVidia has already announced that they would get out of the chipset business.

But nVidia as a subsidiary of Apple can continue to make custom chipsets for Apple.

This would allow nVidia to be a competitor in the general GPU market to AND. - keeping its graphics card business.

But it would take advantage of nVidia's expertise in chipsets to do custom hardware for Apple.

Combine this with the PA Semi team, this would allow Apple to create custom chips and other hardware to differentiate it from the usual Intel crowd yet pull away from them in speed.

Imagine a custom GPU for Apple. This could not be cloned.

Mac OS X would not then be easily used on cloned hardware if its software worked specifically with custom chips.

I will mention that it will be a bit of sweet irony for Apple if it is able to successfully go after IBM's business. It was not too long ago that IBM used to brag that it made more money off a Mac sale then Apple did, all while treating Apple as a second class citizen.

Quote

Guest wrote:

I do think Steve final move as CEO is to make the iPhone your only computer. You heard it here first. You will be able to use it as a phone. Take it home and doc it to some multi-core cradle and use it as your PC at home with your files in the cloud (AKA Me) all the time.

I've been thinking along your idea as well. Also with ultra-thin low-power OLEDs coming up, plus with further miniaturization of components (iPhone system-on-a-membrane anyone?), I envision that the future iPhone brick would be mostly battery with a thin layer of electronics and screen on the top. Imagine the battery life on that sucker.

...than a laptop.

PA Semi was first and foremost a RISC house, a architecture shared by both ARM and PPC. I suspect the ARM SoC angle is much more reasonable than the PPC angle.

Nobody's mentioned, but one side effect of such a design could well be that all the hackintoshes out there would cease to be able to run the latest OS. No more hackintoshes or OSX86 project. I wonder if this isn't part of the rationale?

-Jon

Quote

Terrin wrote:
I will mention that it will be a bit of sweet irony for Apple if it is able to successfully go after IBM's business. It was not too long ago that IBM used to brag that it made more money off a Mac sale then Apple did, all while treating Apple as a second class citizen.

Yes, I was a bit surprised when the X-Box 360 came with 2, 3 (?) G5 CPUs for that kind of price.
CPUs that Apple spent a lot of money designing.

If I imagine being Steve Jobs at that time, I`d put my Golf clubs into the trunk of my Audi and go for IBM s parking lot.

I don't think Apple wants to get back to the horrors of PowerPC and designing their own chipsets, when they can get all of that for free from Intel or any other x86 ecosystem brand. The same goes for GPGPU: OpenCL gives them a generic programming layer which could take advantage of anybody's GPUs, processor cores' unoccupied resources, Intel's Larrabee (if it comes to fruition: Larrabee would be a natural for desktop supercomputing), etc. Plus it's not as if multiprocessing was such a general solution to everything: current Mac Pros have eight cores, and that will scale to sixteen in a few years. Have we seen any killer app taking advantage of that yet? Not really.

Whatever Apple can do at the desktop/laptop processor level, Intel's economies of scale and process prowess will always win, as we saw during Apple's last PPC stages. I can't see Apple doing its own desktop/laptop GPU.

What I can see is Apple evolving their iPhone/iPod Touch technology. That MacBook Nano pic makes a big mistake, in my opinion: such a thing wouldn't be Mac OS X-based but iPhone OS X-based. Iit won't need x86 compatibility, so it will be ARM-powered (PPC at most, but I don't think it is the best fit, really), which gives Apple a real chance to compete with anything Intel can throw to the market in areas such as low power design (a pure RISC design will always be lower power than a RISC-with-a-CISC-emulator), high customizationality (does that word even exist? :-ı) and better leftfieldism.

Apple doesn't need to SCARE their Mac users AWAY from another era of incompatible technologies in a market that doesn't really need it (in a short term way of speaking: we all know the current ways of personal computing are painfully abysmal, OS X included). Actually, its customers are voting with their wallets for Apple to keep being x86/Windows compatible. Where Apple has a real chance of success at being boldly different is out of the traditional personal computer arena: iPhones, "iPads", and any platform that doesn't require the user to fiddle and adjust the working environment as heavily as a Windows or Mac OS X desktop OS does.

Quote

Guest wrote:
Yes, I was a bit surprised when the X-Box 360 came with 2, 3 (?) G5 CPUs for that kind of price.
CPUs that Apple spent a lot of money designing.

Actually, they were not G5-type cores at all, but an IBM custom design. Plus the G5 was itself derived from some IBM's POWER series. Apple hadn't much say in all this.

Quote

Anonymous wrote:
Yeah, well the Windows 7 videos suggest this is what MS is up to, except they aren't going to need the phone to dock anywhere, you just use your Cloud account on any device you come to (with Windows 7, of course).

I do agree, however, that Apple is going to want you to use their iPhone for everything. That's the direction technology is going. But I think laptops will stick around for quite a while, due to the keyboard issue (which is why Steve claimed to not want to build a Tablet PC, remember?).

Yes, the purported cloud computing features in the upcoming Windows OS is something that really sounds interesting to me. Of course the feature has been announced, but I don't think anyone has detailed how it will work, what security features will be in place, and whether or not there will be a subscription fee. And we have yet to find out how well it will work... MobileMe didn't exactly hit the ground running.

The key difference between that and MobileMe, I think, is that Windows is claiming your account will follow you everywhere, where Apple's approach is to have MM mostly tied to hardware that you own (or have a user account on).

I do not agree that iPhone is the center of Apple's future strategy. For one, many of us do not want to be tied to a computer device of minimal use that requires a steep monthly subscription fee. Also, while mobile computing is the big trend these days, not all of us spend our time blogging or playing around on Twitter. Many folks still use computers in an office setting, with a need for the use of hardware such as printers and scanners. That's not the most glamorous thing - or even an area that's driving market growth, but computers are first and foremost a productivity tool and that is what will drive future trends.

Quote

I still object to having my data in the Cloud on anyone else's machines. If I can keep it on my own servers, then this could be one of the greatest ideas. So I don't want MS to have it, not Apple in their Mobile Me service either.

Ditto that. While pervasive cloud computing could become the next generational leap, I'm concerned that the big players are packaging it as a way to draw a continual revenue stream. Guess I'm old school that way: I don't want to keep paying the piper; I want to buy my Mac and software (on disk, of course), then be on my merry way.

Quote

Terrin wrote:
It was not too long ago that IBM used to brag that it made more money off a Mac sale then Apple did, all while treating Apple as a second class citizen.

I never heard that. I just remember a few years ago, when the rumors that Apple was going to switch to Intel were hitting pretty hard, the execs over at Big Blue were scrambling to get Apple on the phone and they weren't returning their calls.

Can you point me to a reference, so I can read up on this?

The largest lesson from the success of the ipod seems to be that a product which is given a chance to integrate into households and can live as a service with tremendous additional add ons.
I had a buddy who didn't want an Ipod, but was essentially dragged into an ipod because of the number of accessories which were available for car, for home, for carrying between car and home.
Apple as a company is interested in having an Apple ecology of Consumer Electronics. The Apple TV is just the first reconnaissance into that territory. That golden land requires low power chips, requires integration at a level that the Atom is not well suited at this time. A recently revealed patent from Apple showed an integration idea vis a vis phone and other devices a customer may carry, see http://arstechnica.com/news.ars/post/20081024-apple-patent-would-let-your-shoes-talk-to-your-watch-iphone.html.
Think of such an integration agenda at home, alarm clocks, TV's, Monitors, Game Consoles; devices able to take info from one database (calendar) and display in areas other than the input account. PA semi is part of that solution, Iphone is part of that solution. The success of netbooks will force apple to show their hand, they really cannot do a sub $700 Mac unless there is a value add which cannot be commoditzed.
The days of speed demo's is almost over. What home owners want is the efficiencies of sync between devices, they will buy two, or more.

John Martellaro and I feel exactly the same way about the future silicon that will eventually drive ALL of Apple's product lines. The new architecture will be RISC without a doubt, and will have an instruction set similar to IBM POWER. The new ISA will be uniquely Apple's and may turn out to be a superior design in both parallel processing capability and breakthrough power consumption.

Steve Jobs several years ago had a falling out with IBM over not delivering the chips that he wanted. He is now in the process of getting all of the pieces and people into his house to get the silicon he has always wanted for his machines to run on, even if he has to steal away IBM's elite to get it done.

nVidia is NOT EXITING THE CHIPSET MARKET.

It's extremely old news that some idiotic analyst made it up and nVidia flatly denied it.

Also, in my opinion, PPC is not going into Macs. Steve Jobs has already said PA Semi will design custom chips for Apple's iPhone/iPod Touch line.

So...this analysis was interesting, but I don't agree with the conclusions.

In response to people asking about Windows/x86 compatibility . . . I don't expect Apple to pull it, but what you can see are the following pieces :

1) A binary architecture that already supports multiple CPU and 32/64-bit.
2) An O/S that works with 3 very different CPU architectures
3) Work being put into a new C-compiler project (clang) that outputs to LLVM (low-level virtual machine) as well as direct x86
4) The general acceptance that future speed increases are about multi-core, rather than clockspeed.
5) OpenCL support (running processes on non-x86 GPU)
6) GrandCentral (whatever that may turn out to be)
7) Increasing amount of development in higher (non-C) level languages.

My guess for the long-term is a platform which is CPU/GPU agnostic, but contains at least one x86 chip.