Every year, Apple integrates more ancillary functionality into its silicon, all in service of the needs of the OS. When Apple finally integrates its nascent cellular hardware into its processors, does anyone think they'll unravel that to go back Intel, which sold all of their cellular capability to... Apple? From what I've seen of recent designs, less than 50% of Apple silicon is the CPU. Intel would have to beat Apple's efforts in CPU, GPU, AI, Encode/Decode and 5G/Wi-fi/60GHz RF.
Intel's only hope of winning Apple business is as a foundry, earning foundry margins.
The only way this could ever happen is if Intel dropped x86 and switched to a modern architecture. But I don't think that's going to happen for a long time. They invented Itanium for exactly that reason and refused to add 64-bit support to x86 to try and make people switch, but it didn't catch on. Instead AMD bodged 64 bit onto x86 so Intel followed, and here we are.
We know Intel hit the performance wall some time ago, so it remains to be seen whether Apple can keep up with its leaps in performance each year. Hopefully the lack of focus on performance at this year's iPhone keynote was just a bump - though a worrying number of people known to be incredible silicon engineers have left Apple in the past few years.
Agreed - as multiple others have posted Intel's big mistake has been trying to ride a dying horse. There's nothing that says they can't do both, though. If they were to develop an ARM chip, or even an ARM chip that allowed hardware translation of x86 instructions they could get back in the game.
The new Intel chip had better not have x86 instructions. x86 offers zero advantage these days when developers compile their code and the correct instructions are automatically created for whatever architecture is in use. x86 is all downside now. It is baggage from an ancient past. It is a large part of the reasons why Intel CPUs are bigger, slower and run hotter than Apple Silicon.
Apple is really the only OEM that can do this. For example, how could HP or Dell make a move like this without Microsoft? They couldn’t. Controlling both the hardware and the operating system makes this possible.
Then there’s the problem of Intel mocking Apple’s move to its own silicon. Remember, though, that Samsung has a long history of mocking both Apple and its users yet Apple still does business with them. Business is business.
You seem to be forgetting that Apple was in exactly the same situation as HP and Dell until just recently. It still hasn't even fully transitioned yet.
Dell has been offering non-Microsoft OS options for 20 years. HP also has non-Microsoft options.
The possibility of creating in-house chipsets is not beyond either of them. Itanium began life at HP.
The bigger question is if the investment is worth it to them. At the moment, it seems they think it isn't. Just like Apple did until they decided to go it alone.
This isn't that different to PowerPC at the beginning. I wouldn't go out on a limb now and say how Apple might be progressing with regards to the rest of the industry in 10 years.
Of course Apple isn't the only vendor of non-x86 processors. And of course it isn't the only vendor of ARM ISA chipsets either. Nor is it the only vendor controlling both the OS and the hardware.
? I know three engineers who work for Intel, one of them complained to me for years that the only solution was adding more capacitors which adds more heat. All of this was true but Apples approach wins out because the engineers I talked to at Intel didn’t see a way forward and now the CEO knows today that they’re going to be destroyed by these new chips, obliterated is a more correct term to use…
Of course you do. There are no capacitors in CPU (cores) for a start, and if you meant transistors you'll find that Apple's M1 has ~16bn transistors, twice as many as the 24-core Xeon 8180. That's ballooned from 3.3bn in the A10.
Adding transistors doesn't necessarily = more heat, because they're not all constantly switching - which is the only time they are dissipating heat. Application specific silicon can reduce the heat produced but increase the transistor count because there are less total transistor flips for a particular piece of code to execute: it's more efficient. And as above, if that silicon is idle it's not using power.
Intel CPUs are so inefficient because they are essentially a CISC interpreter ontop of a RISC CPU. Plus due to backward compatibility, there are thousands of SIMD extensions that are used by barely anything but can't be removed due to the few customers that do need them.
Quick update to those transistor numbers:
M1 Pro features 33.7 billion transistors M1 Max features 57 billion transistors
Apple is all about the control of what they do.....engineering, development and delivery. Right now the one thing Apple is fine with third parties is manufacturing. And that's the only path for Intel right now since Apple will never ever go back to buying an expensive general purpose CPU or GPU since they will never meet their specific needs. And right now, none of them have the power or especially the power per watt that Apple is doing.
Too little, too late. Me thinks they should’ve fought harder to KEEP Apple’s business rather than fight to get Apple’s business. But it wouldn’t have mattered much, Apple isn’t a priority with Intel to begin with.
GeorgeBMac said: Apple will increasingly integrate their product lines into one cohesive whole. That means that the Mac line will no longer be an orphan, the weird uncle who's only invited on holidays. It will become an integral part of a family of products in which each interacts, supports and strengthens the others.
But that cannot happen if it would return to using non-Apple hardware that is essentially incompatible with Apple's other core products: AppleWatch, iPhone and iPad.
The hardest and most failure prone aspect of any computer system is its interaction with other computer systems. The more those barriers between systems are eliminated the more stable and powerful they become. Putting what is essentially an iPhone processor in Macs went a very long way to breaking down those barriers. And that is an advantage that Intel, AMD or even Samsung can ever match.
The problem here is software. It doesn't much matter if Apple gains a uniform platform, if they fail at getting all the (important) software that current exists on Wintel over to the Mac. I'm hoping the speed differential tempts a bunch of them, but ultimately, this is still a big task Apple faces.
... That's a war in which Apple will increasingly be just another player as more and more computers and servers are switched over to ARM based processors and leave the world of x86 behind.
Maybe if you're in the server business, but you don't seem many high end CAD, 3D, engineering, etc. apps running on ARM. They run on Windows, mostly (which we could run on our Macs). That's the real hurdle Apple needs to overcome, and the advantage Intel has.
What Apple has done is really impressive, but if you can't run the software you need to run, it doesn't much matter. The question is how successful Apple will be in courting all these Wintel packages to modernizing their apps and creating Apple Silicon/Metal versions. It isn't impossible, but Apple hasn't been great at that aspect.
The new Intel chip had better not have x86 instructions. x86 offers zero advantage these days when developers compile their code and the correct instructions are automatically created for whatever architecture is in use.
It can't be nearly that easy or all those Wintel packages would simply compile a Mac version.
avon b7 said: This isn't that different to PowerPC at the beginning. I wouldn't go out on a limb now and say how Apple might be progressing with regards to the rest of the industry in 10 years.
I think it is way different in that Apple isn't a relatively small company talking tech industry giants into co-developing a platform. They now have some of the best chip designers and are the major player, making their own stuff just how they want to make it.
His battle is not to win Apple back (they're gone forever from the world of x86).
The war he is fighting is x86 versus ARM based processors.
That's a war in which Apple will increasingly be just another player as more and more computers and servers are switched over to ARM based processors and leave the world of x86 behind.
It's analogous to the makers of carriages after the internal combustion engine challenged the horse. Many fought it and went out of business -- while others welcomed the change and stuck an internal combustion engine in their carriages making "horseless carriages".
I suspect Intel is looking at all of its options right now.
You need to see the forest for the trees. He was asked about Apple and said what any reasonable person would say about getting back (or technically not losing out completely) an old customer.
And it's also not about x86 v ARM, but that's another discussion.
For Intel, it's ALL about x86 vs ARM.
Currently, everybody can switch to the newer technology (Microsoft, HP, Dell, Lenovo, etc...) fairly easily. But not Intel. For them it means walking away from their bread & butter and taking on some serious competition -- competition that they may not survive.
But, to think that Apple would ever go back to Intel (even if they had an ARM processor) is only seeing the trees not the forest: For Apple, putting a souped up iPhone chip in their PC line allows them to do what they do best: integrate their product lines into a single, seamless whole.
GeorgeBMac said: Apple will increasingly integrate their product lines into one cohesive whole. That means that the Mac line will no longer be an orphan, the weird uncle who's only invited on holidays. It will become an integral part of a family of products in which each interacts, supports and strengthens the others.
But that cannot happen if it would return to using non-Apple hardware that is essentially incompatible with Apple's other core products: AppleWatch, iPhone and iPad.
The hardest and most failure prone aspect of any computer system is its interaction with other computer systems. The more those barriers between systems are eliminated the more stable and powerful they become. Putting what is essentially an iPhone processor in Macs went a very long way to breaking down those barriers. And that is an advantage that Intel, AMD or even Samsung can ever match.
The problem here is software. It doesn't much matter if Apple gains a uniform platform, if they fail at getting all the (important) software that current exists on Wintel over to the Mac. I'm hoping the speed differential tempts a bunch of them, but ultimately, this is still a big task Apple faces.
... That's a war in which Apple will increasingly be just another player as more and more computers and servers are switched over to ARM based processors and leave the world of x86 behind.
Maybe if you're in the server business, but you don't seem many high end CAD, 3D, engineering, etc. apps running on ARM. They run on Windows, mostly (which we could run on our Macs). That's the real hurdle Apple needs to overcome, and the advantage Intel has.
What Apple has done is really impressive, but if you can't run the software you need to run, it doesn't much matter. The question is how successful Apple will be in courting all these Wintel packages to modernizing their apps and creating Apple Silicon/Metal versions. It isn't impossible, but Apple hasn't been great at that aspect.
True, and that's why I've (also) been saying that Apple needs to work with Microsoft to get Windows back into Bootcamp on their newer devices. And, it's not just for the high end engineering apps but also for the custom stuff used by schools and businesses that were developed for and are only supported as Windows based software.
But, it sounds like some of the virtual platforms may fill that hole.
The new Intel chip had better not have x86 instructions. x86 offers zero advantage these days when developers compile their code and the correct instructions are automatically created for whatever architecture is in use.
It can't be nearly that easy or all those Wintel packages would simply compile a Mac version.
avon b7 said: This isn't that different to PowerPC at the beginning. I wouldn't go out on a limb now and say how Apple might be progressing with regards to the rest of the industry in 10 years.
I think it is way different in that Apple isn't a relatively small company talking tech industry giants into co-developing a platform. They now have some of the best chip designers and are the major player, making their own stuff just how they want to make it.
IBM & Motorola were hardly "small companies". But they were up against the HP/Intel/Microsoft cartel -- not to mention a media that was enamored by those new guys and hammered anything the old school attempted.
GeorgeBMac said: True, and that's why I've (also) been saying that Apple needs to work with Microsoft to get Windows back into Bootcamp on their newer devices. And, it's not just for the high end engineering apps but also for the custom stuff used by schools and businesses that were developed for and are only supported as Windows based software.
But, it sounds like some of the virtual platforms may fill that hole.
Hmm, I never thought about an ARM-based Bootcamp if Windows succeeds on that front. Interesting though. I just kind of assumed it was all about VM software now.
GeorgeBMac said: IBM & Motorola were hardly "small companies". But they were up against the HP/Intel/Microsoft cartel -- not to mention a media that was enamored by those new guys and hammered anything the old school attempted.
Oh, no, I meant Apple was a relatively small player at that point (by comparison). They were just trying to leverage those players (who also had something to gain) in an attempt to better compete against Wintel. It's a night and day difference now (which is why I responded to that post).
It will be interesting to see the role the media plays this time around. Apple (the company) is a media-darling this time around, yet still fairly biased against Macs when talking about anything business or IT.
Comments
Intel's only hope of winning Apple business is as a foundry, earning foundry margins.
Dell has been offering non-Microsoft OS options for 20 years. HP also has non-Microsoft options.
The possibility of creating in-house chipsets is not beyond either of them. Itanium began life at HP.
The bigger question is if the investment is worth it to them. At the moment, it seems they think it isn't. Just like Apple did until they decided to go it alone.
This isn't that different to PowerPC at the beginning. I wouldn't go out on a limb now and say how Apple might be progressing with regards to the rest of the industry in 10 years.
Of course Apple isn't the only vendor of non-x86 processors. And of course it isn't the only vendor of ARM ISA chipsets either. Nor is it the only vendor controlling both the OS and the hardware.
M1 Pro features 33.7 billion transistors
M1 Max features 57 billion transistors
Maybe if you're in the server business, but you don't seem many high end CAD, 3D, engineering, etc. apps running on ARM. They run on Windows, mostly (which we could run on our Macs). That's the real hurdle Apple needs to overcome, and the advantage Intel has.
What Apple has done is really impressive, but if you can't run the software you need to run, it doesn't much matter. The question is how successful Apple will be in courting all these Wintel packages to modernizing their apps and creating Apple Silicon/Metal versions. It isn't impossible, but Apple hasn't been great at that aspect.
I think it is way different in that Apple isn't a relatively small company talking tech industry giants into co-developing a platform. They now have some of the best chip designers and are the major player, making their own stuff just how they want to make it.
Oh, no, I meant Apple was a relatively small player at that point (by comparison). They were just trying to leverage those players (who also had something to gain) in an attempt to better compete against Wintel. It's a night and day difference now (which is why I responded to that post).
It will be interesting to see the role the media plays this time around. Apple (the company) is a media-darling this time around, yet still fairly biased against Macs when talking about anything business or IT.