You don't win by having a unique CPU. You win by having the right solution and pricing. That is why moving to Intel was such a huge step. Columnists no longer had to write about the "megahertz myth" and Apple no longer had to do spend their marketing $ on explaining how great PowerPC was or waiting for Motorola to produce them. It was all a non-issue to users and resellers. Apple won big time by doing so. They could sell based on design and their OS - which in part won by not being Vista. You didn't need Windows + Mac as your Mac could run Windows.
Will it work this time?
1. Hardware+OS iPhone killed Nokia but the only disruption on PC since Windows 95 was the failure of Vista and the change in K-12 from Mac+Win to 60% ChromeOS. Users just don't change platform for a new CPU. Did ordinary iPhone 4 users care that they had the first A-series CPU or did they discuss Antennagate?
2. Software Users + companies do stick to their software. The only major change in software preference within the last 10-15 years is from Internet Explorer to Chrome. Last time Apple changed happened to be with MS under maximum legal pressure to avoid being split and during Vista. There are no such tailwinds at this point.
To those that want Intel for VMs or even to run Windows in Boot Camp (I do both), the answer is an Intel daughterboard or external box like an eGPU as an option, we had them before we can have them again.
Last time was when most used bulky desktops. It won't work when you're on the road or if you need that special app in a classroom. It will be super expensive and not very popular - just like last time. Apple could perhaps do some kind of Thunderbolt accessory but the current solution comes a zero cost. Any new setup will require software or hardware with cost and complexity. It all reduces the value of the Mac.
Nonsense and typical columnist speculation.
There is no evidence that any of this is true or that it would make any sense to do.
Actually a lot of this is true. That fact that that the information has been condensed into an article of speculation doesn’t make the info any less true. The fact is Apple has put a lot of effort into its App Store system to better support multiple architectures. At the moment those architectures are basically of the same CPU family.
What i I think the articles author missed though is why would Apple want to do this. I really think it comes down to the customer logic the author touched upon. In the future ARM cores or X86 cores will not be the area where the most real estate will be dedicated. In fact those old cores will not even be the most important parts of a chip. Rather I’m expecting a steady increase in the importance of AI acceleration blocks and other special function hardware. The actual CPU cores will steadily become proportionally smaller on the SoC. We can see this today in Apples SoC Where the amount of space dedicated to “””cores””” grows steadily smaller. I could even see Apple adopting chiplet tech to break up the SoC with the ARM cores separated from all the new stuff. This would allow them to rapidly advance the new technologies like Neural Engine or advanced GPUs. In five years time we may find us more focused on Neural Engine performance than ARM cores. In the end nobody will care about the ARM cores.
I was once convinced Apple would switch to ARM in Macs, but if Apple were going to do this, they should have done it already. With TSMC's process advantage and Apple's strong design team, they could have ARM SOCs in Macs today that smoke Intel's offerings. Even poor undercapitalized AMD has taken a lead in everything but games over Intel by combining their Zen design with TSMC 7nm. With vastly more resources, Apple's design team could surely beat AMD's, and by extension, have a big lead over Intel right now. If Apple were going to go ARM in Macs, that's the way to do it -- with a big performance lead to get people excited and willing to accept the loss of x86 compatibility.
The problem is that by this time next year, Intel will have moved onto Ice Lake on their 10nm+ design (similar to TSMC 7nm+). Ice Lake offers a big improvement over Skylake. Also, Intel is very well aware of the value of specialized units (NPUs etc), and so they will be offering that type of functionality as well. It will therefore be harder next year for Apple to post a big performance victory over Intel.
And maybe that's why Apple hasn't made the switch and won't make the switch -- as frustrating as Intel's missteps have been, Intel is still a very formidable designer and manufacturer of CPUs. Perhaps Apple looked at Intel's roadmap a few years ago and decided they just couldn't beat it, or at least not beat it by enough to warrant the risk of a switch. It's annoying that Intel's roadmap was delayed, but the ultimate conclusion might be the same -- it's just not worth it.
Nonsense and typical columnist speculation.
There is no evidence that any of this is true or that it would make any sense to do.
I think there was clearly lack of hardware understanding, especially the last few sentences on CPU ISA. And the cost involves in building Desktop Class CPU and GPU. Could Apple do it, Yes. Does it make any sense? Well no one have given a decent argument yet.
The Mac App Store is not mandatory, and its App Submission does not force the use of OpCode, not to mention the compiler optimisation are ISA specific. Which we have known and confirmed on multiple WWDC occasions for years. ( And yet AI still repeat those nonsenses )
It talks about a Mix of ARM ( only ) and Intel ( Mix ARM ) Model without specifically mentioning it. An ARM Only Mac is basically an iPad with Keyboard.
I thought all of these were Obvious, especially after WWDC when Apple has double down on Pro ( Real Pro ) and iPadOS.
And the reason why they never used AMD? AMD wasn't any good in the last 15 years. It isn't until Zen 2 ( Only Released two weeks ago ) were they competitive on multiple levels.
Apple already has desktop class CPU cores and has had for the last two A-series releases. There is not a huge incremental cost in adding the additional logic to support the rest of a desktops needs. Basically you are talking about adding more I/O, cache and wider access to RAM. While not trivial such adjustments, to make a desktop variant, it is also not impossible.
Ive always said that a switch to ARM cores in Apples products will not be about ARM cores but rather about building what they want into the SoC. SoC are what massivePCB where to computers in the 1980’s, it is the place that engineers add tech to make your device unique.
Asfor laptops the A series is real close to being good enough for them also. All you really need are two USB-C ports supporting TB. Well that and maybe more ports for RAM. This is not a massive overhaul to the A-Series and could in fact still be used in the iPad. It simply wouldn’t be a massive effort on Apples part to start using Apples ARM chips in Macs.
If Apple releases an ARM Mac it isn't gonna be just an old Macbook with an ARM inside. It's gonna be something revolutionary and game changing.
Certainly. The funny thing is, all Macs with the T2 chip are already “ARM Macs” (Except the iMac, which next generation will include T2 most probably.) That can be read in two ways: “Apple is gradually transitioning to ARM” and “Apple reserves its desktop capacity ARM to iPad Pros and the T2(3, 4...n) is all the ARM Mac you can get for at least five years”.
As other posters pointed out, if SoC integration is more important than the number of cores, then there is no reason to believe that Apple will compete with Intel on CPU cores/speed in such a destructive way and won’t let Intel to adapt to Apple’s new logic boards / SoCs.
It's interesting to read these comments and see which ones criticize Daniel and which ones support his thoughts. Lots of commenters with few posts think Daniel is either crazy or dumb yet they probably don't realize Daniel's been performing technical reviews of Apple products practically since day 1. He knows Apple's history and has a lot of inside information and contacts. I've also seen him speak and he knows his stuff. Those with more posts either like him or don't but I believe they do understand he has the knowledge to make these assertions. Apple makes decisions based on what they feel they can sell. They offer products that won't sell a lot to keep themselves in certain businesses (video, audio, etc.). Apple provided a way to run Windows natively (best OTS PC to run Windows happens to be a Mac and tests show this) and a few companies provided virtual engines to run Windows and unix/linux as well so Apple users can have the compatibility they require (simply because software vendors hate having to deal with a company (Apple) who cares more about security than many application programmers do).
I have been waiting for ARM-based cluster devices from Apple that could provide native platforms for computational-intensive applications that could fit in a shoebox (sans power supply) and deliver more power than a bank or traditional servers. Apple's GPUs appear to be able to run without creating a lot of heat so I'd love to see Apple create its own eGPU "shoebox" with the equivalent of 1000 GPUs with some kind of cooling. The A12x Bionic has 7 GPU cores along with 8 CPU cores, all packed in a very small chip. There's no reason why Apple couldn't design the mother of all GPU board containing a huge number of dedicated GPU cores (same for CPU cores). I know there's a lot more that goes into a motherboard than these two devices but Apple has, as a few others have stated, desktop performance coming out of the latest A-series chips and iOS devices. I could see an MacBook Air with an A12X Bionic chip and extra solid state memory knocking the socks off the existing MacBook Air.
Mr. Dilger's article almost makes me feel sorry for Intel. Almost.
What's interesting about Intel isn't that they failed to recognize the corner they'd painted themselves into with the x86 architecture -- it's that they DID recognize, and tried to solve it, and failed. They tried to get into other chip fabrications like broadband, and failed. They tried IA-64, and failed. They tried the Atom, and failed. For broadband and Atom it became clear to the industry that Intel's solution wasn't good enough, but for IA-64 and Itanium Intel fell into the classic trap of having a superior product that others wouldn't invest in to use. Microsoft wasn't going for it and neither were the other industry leaders.
You'd think that someone at Intel would learn from all this failure -- Apple (well, really, Jobs, and to a fair extent Cook and Ive) certainly learned from failure, which is why we got the iMac, iPod, iPhone, Mac OS X, etc. Failure, if you survive it, is a good teacher. What has Intel learned? Darned if I know.
they probably don't realize Daniel's been performing technical reviews of Apple products practically since day 1. He knows Apple's history and has a lot of inside information and contacts.
Is it "Apple will do this" or "Apple should do this"? I'm sure they can do it. This reads like an opinion on why Apple should do it and we're discussing is "does it make sense". There can be several views on this. Daniel doesn't exactly sport a track record of nuanced or unbiased reporting. Which is fine for an op-ed.
I have been waiting for ARM-based cluster devices from Apple that could provide native platforms for computational-intensive applications that could fit in a shoebox (sans power supply) and deliver more power than a bank or traditional servers.
The Mac Pro is for computational intensive stuff. Apple only did updates to the motherboard in 2009, 2013, and 2019. It has no focus at all. Why update it now to kill it off a few months later? It would make the new Mac Pro a new Mac IIvi (and if you don't know the story behind that... then you're new to the game). That would really upset the remaining Pro customers. Jobs discontinued Xserve, the storage systems, and the Apple on-site services. It will take 5+ years to get that started. Right now server business moves to AWS and Azure - often without a classic OS.
But Apple could perhaps use their server farms in a Google Stadia like way and move gaming + heavy stuff to the cloud that could run on A-series equipment. And they could move the Mac GPU to their own design. And they could make an iPad in a shell design (eMate). That wouldn't have an impact on the Mac or engineering resources. But switching all of Mac... it would take time away from their i-series hardware teams and the teams designing their processors. With Mac getting little attention today it would be even worse. We have seen how iOS can delay macOS. Same would happen on hardware.
Nonsense and typical columnist speculation.
There is no evidence that any of this is true or that it would make any sense to do.
Why it would make sense for Apple is clear... it's a cogent discussion... unlike your comment. Just because you can't understand or comprehend a thesis doesn't make it nonsense
Mr. Dilger's article almost makes me feel sorry for Intel. Almost.
What's interesting about Intel isn't that they failed to recognize the corner they'd painted themselves into with the x86 architecture -- it's that they DID recognize, and tried to solve it, and failed.
Intel is the Boeing of semiconductors. Their products were not always great but they had the ability to introduce new process types (more transistors within the same space) and scale manufacturing. They lost that edge to TSMC 5 years ago. Now AMD can use TSMC to produce 7nm x86 units (Ryzen 3) while Intel is stuck on 14nm - and won't be on 7nm before 2022/2023. Apple could switch to AMD. Why haven't they?
Nice to read if you are locked in the thermal corner of a prosumer laptop. But sorry, from the iMac Pro perspective I see no interest at all in this prospected evolution.
What? You don't want to run your software an order of magnitude faster for two orders of magnitude less money?
A significant factor in the PPC > x86 switch was Rosetta. It is much easier to emulate RISC PPC with its relatively small instruction set than it is CISC x86, and now x64. PPC apps running in Rosetta weren't much slower than the native ones, but that was also partly offset by the Intel CPUs being much, much faster than PPC ones. The A-series CPUs are quick, and in a less power and thermally constrained environment no doubt even quicker - but CISC emulation on RISC architectures is excruciatingly slow, no matter how fast the native CPU. Remember Connectix's Virtual PC? That emulated an x86 machine on PPC. Installing Win98 took 3 or 4 hours even on a G5. Of course API level emulation a-la Rosetta has less overhead, but it's still slow.
Also, people who are switching to Mac can still use the Mac as a PC if they need to. It provides a comfort blanket. As soon as Apple switched to x86, Mac sales took off.
I wouldn't call running Windows comfortable, not even in another universe. Its best to get rid of it. Running ppc apps under Rosetta was slow, very slow and some apps didn't run at all. Running CISC on RISC or vise versa isn't inherently more difficult. It isn't guaranteed to be symmetrical but that doesn't depend on CISC or RISC (this is nowadays an outdated distinction) or the number of instructions one or the other has. I would say that a 64 bit instruction set (or not) is a more important notion when translating instruction sets. The internal state of the processor and how easily it is represented on another (processor) is also an important notion. I would expect to see a difference in efficiency even per instruction. All in all I expect that on average only a few instructions are needed to translate one instruction set to another no matter what. Current processors are extremely fast so a factor 5 or so will not be noticed when running most apps.
Edit: one fun fact to consider is that Intel already hardware translates its (CISC) instruction set(s) on its internal RISC instruction set at full clock speed, so its certain that this isn't a slow option.
A significant factor in the PPC > x86 switch was Rosetta. It is much easier to emulate RISC PPC with its relatively small instruction set than it is CISC x86, and now x64. PPC apps running in Rosetta weren't much slower than the native ones, but that was also partly offset by the Intel CPUs being much, much faster than PPC ones. The A-series CPUs are quick, and in a less power and thermally constrained environment no doubt even quicker - but CISC emulation on RISC architectures is excruciatingly slow, no matter how fast the native CPU. Remember Connectix's Virtual PC? That emulated an x86 machine on PPC. Installing Win98 took 3 or 4 hours even on a G5. Of course API level emulation a-la Rosetta has less overhead, but it's still slow.
Also, people who are switching to Mac can still use the Mac as a PC if they need to. It provides a comfort blanket. As soon as Apple switched to x86, Mac sales took off.
I wouldn't call running Windows comfortable, not even in another universe. Its best to get rid of it. Running ppc apps under Rosetta was slow, very slow and some apps didn't run at all. Running CISC on RISC or vise versa isn't inherently more difficult. It isn't guaranteed to be symmetrical but that doesn't depend on CISC or RISC (this is nowadays an outdated distinction) or the number of instructions one or the other has. I would say that a 64 bit instruction set (or not) is a more important notion when translating instruction sets. The internal state of the processor and how easily it is represented on another (processor) is also an important notion. I would expect to see a difference in efficiency even per instruction. All in all I expect that on average only a few instructions are needed to translate one instruction set to another no matter what. Current processors are extremely fast so a factor 5 or so will not be noticed when running most apps.
If it was so easy why Microsoft has failed with Surface RT? Why desktop Windows applications didn’t run on RT?
A significant factor in the PPC > x86 switch was Rosetta. It is much easier to emulate RISC PPC with its relatively small instruction set than it is CISC x86, and now x64. PPC apps running in Rosetta weren't much slower than the native ones, but that was also partly offset by the Intel CPUs being much, much faster than PPC ones. The A-series CPUs are quick, and in a less power and thermally constrained environment no doubt even quicker - but CISC emulation on RISC architectures is excruciatingly slow, no matter how fast the native CPU. Remember Connectix's Virtual PC? That emulated an x86 machine on PPC. Installing Win98 took 3 or 4 hours even on a G5. Of course API level emulation a-la Rosetta has less overhead, but it's still slow.
Also, people who are switching to Mac can still use the Mac as a PC if they need to. It provides a comfort blanket. As soon as Apple switched to x86, Mac sales took off.
I wouldn't call running Windows comfortable, not even in another universe. Its best to get rid of it. Running ppc apps under Rosetta was slow, very slow and some apps didn't run at all. Running CISC on RISC or vise versa isn't inherently more difficult. It isn't guaranteed to be symmetrical but that doesn't depend on CISC or RISC (this is nowadays an outdated distinction) or the number of instructions one or the other has. I would say that a 64 bit instruction set (or not) is a more important notion when translating instruction sets. The internal state of the processor and how easily it is represented on another (processor) is also an important notion. I would expect to see a difference in efficiency even per instruction. All in all I expect that on average only a few instructions are needed to translate one instruction set to another no matter what. Current processors are extremely fast so a factor 5 or so will not be noticed when running most apps.
If it was so easy why Microsoft has failed with Surface RT? Why desktop Windows applications didn’t run on RT?
Thats a good question. Its maybe best to ask MS. I think this has noting to do with translating instruction set(s), but with (memory) capacity and processor speed. So, this is possibly by design (which probably means in this case that some manager pushed his/her faulty ideas and didn't listen to engineering).
The primary message of this article is that Apple is heading in this direction, and it sure seems to be correct in that regard. Has anyone seen any indication of Apple actively entrenching or reinforcing its use of x86, anywhere?
This is what I meant when I mentioned that Microsoft was helping to solve the boot camp issue. Theoretically, Window 10 on ARM would work for the desperate one or two occasional work apps some Mac users need, at least.
Comments
Will it work this time?
1. Hardware+OS
iPhone killed Nokia but the only disruption on PC since Windows 95 was the failure of Vista and the change in K-12 from Mac+Win to 60% ChromeOS. Users just don't change platform for a new CPU. Did ordinary iPhone 4 users care that they had the first A-series CPU or did they discuss Antennagate?
2. Software
Users + companies do stick to their software. The only major change in software preference within the last 10-15 years is from Internet Explorer to Chrome. Last time Apple changed happened to be with MS under maximum legal pressure to avoid being split and during Vista. There are no such tailwinds at this point.
Last time was when most used bulky desktops. It won't work when you're on the road or if you need that special app in a classroom. It will be super expensive and not very popular - just like last time. Apple could perhaps do some kind of Thunderbolt accessory but the current solution comes a zero cost. Any new setup will require software or hardware with cost and complexity. It all reduces the value of the Mac.
What will it add?
What i I think the articles author missed though is why would Apple want to do this. I really think it comes down to the customer logic the author touched upon. In the future ARM cores or X86 cores will not be the area where the most real estate will be dedicated. In fact those old cores will not even be the most important parts of a chip. Rather I’m expecting a steady increase in the importance of AI acceleration blocks and other special function hardware. The actual CPU cores will steadily become proportionally smaller on the SoC. We can see this today in Apples SoC Where the amount of space dedicated to “””cores””” grows steadily smaller. I could even see Apple adopting chiplet tech to break up the SoC with the ARM cores separated from all the new stuff. This would allow them to rapidly advance the new technologies like Neural Engine or advanced GPUs. In five years time we may find us more focused on Neural Engine performance than ARM cores. In the end nobody will care about the ARM cores.
The problem is that by this time next year, Intel will have moved onto Ice Lake on their 10nm+ design (similar to TSMC 7nm+). Ice Lake offers a big improvement over Skylake. Also, Intel is very well aware of the value of specialized units (NPUs etc), and so they will be offering that type of functionality as well. It will therefore be harder next year for Apple to post a big performance victory over Intel.
And maybe that's why Apple hasn't made the switch and won't make the switch -- as frustrating as Intel's missteps have been, Intel is still a very formidable designer and manufacturer of CPUs. Perhaps Apple looked at Intel's roadmap a few years ago and decided they just couldn't beat it, or at least not beat it by enough to warrant the risk of a switch. It's annoying that Intel's roadmap was delayed, but the ultimate conclusion might be the same -- it's just not worth it.
Ive always said that a switch to ARM cores in Apples products will not be about ARM cores but rather about building what they want into the SoC. SoC are what massivePCB where to computers in the 1980’s, it is the place that engineers add tech to make your device unique.
Asfor laptops the A series is real close to being good enough for them also. All you really need are two USB-C ports supporting TB. Well that and maybe more ports for RAM. This is not a massive overhaul to the A-Series and could in fact still be used in the iPad. It simply wouldn’t be a massive effort on Apples part to start using Apples ARM chips in Macs.
If Apple releases an ARM Mac it isn't gonna be just an old Macbook with an ARM inside. It's gonna be something revolutionary and game changing.
As other posters pointed out, if SoC integration is more important than the number of cores, then there is no reason to believe that Apple will compete with Intel on CPU cores/speed in such a destructive way and won’t let Intel to adapt to Apple’s new logic boards / SoCs.
I have been waiting for ARM-based cluster devices from Apple that could provide native platforms for computational-intensive applications that could fit in a shoebox (sans power supply) and deliver more power than a bank or traditional servers. Apple's GPUs appear to be able to run without creating a lot of heat so I'd love to see Apple create its own eGPU "shoebox" with the equivalent of 1000 GPUs with some kind of cooling. The A12x Bionic has 7 GPU cores along with 8 CPU cores, all packed in a very small chip. There's no reason why Apple couldn't design the mother of all GPU board containing a huge number of dedicated GPU cores (same for CPU cores). I know there's a lot more that goes into a motherboard than these two devices but Apple has, as a few others have stated, desktop performance coming out of the latest A-series chips and iOS devices. I could see an MacBook Air with an A12X Bionic chip and extra solid state memory knocking the socks off the existing MacBook Air.
There can be several views on this. Daniel doesn't exactly sport a track record of nuanced or unbiased reporting. Which is fine for an op-ed.
The Mac Pro is for computational intensive stuff. Apple only did updates to the motherboard in 2009, 2013, and 2019. It has no focus at all. Why update it now to kill it off a few months later? It would make the new Mac Pro a new Mac IIvi (and if you don't know the story behind that... then you're new to the game). That would really upset the remaining Pro customers. Jobs discontinued Xserve, the storage systems, and the Apple on-site services. It will take 5+ years to get that started. Right now server business moves to AWS and Azure - often without a classic OS.
But Apple could perhaps use their server farms in a Google Stadia like way and move gaming + heavy stuff to the cloud that could run on A-series equipment. And they could move the Mac GPU to their own design. And they could make an iPad in a shell design (eMate). That wouldn't have an impact on the Mac or engineering resources. But switching all of Mac... it would take time away from their i-series hardware teams and the teams designing their processors. With Mac getting little attention today it would be even worse. We have seen how iOS can delay macOS. Same would happen on hardware.
Its best to get rid of it.
Running ppc apps under Rosetta was slow, very slow and some apps didn't run at all.
Running CISC on RISC or vise versa isn't inherently more difficult. It isn't guaranteed to be symmetrical but that doesn't depend on CISC or RISC (this is nowadays an outdated distinction) or the number of instructions one or the other has.
I would say that a 64 bit instruction set (or not) is a more important notion when translating instruction sets. The internal state of the processor and how easily it is represented on another (processor) is also an important notion.
I would expect to see a difference in efficiency even per instruction.
All in all I expect that on average only a few instructions are needed to translate one instruction set to another no matter what.
Current processors are extremely fast so a factor 5 or so will not be noticed when running most apps.
Edit: one fun fact to consider is that Intel already hardware translates its (CISC) instruction set(s) on its internal RISC instruction set at full clock speed, so its certain that this isn't a slow option.
Its maybe best to ask MS.
I think this has noting to do with translating instruction set(s), but with (memory) capacity and processor speed.
So, this is possibly by design (which probably means in this case that some manager pushed his/her faulty ideas and didn't listen to engineering).
https://www.google.com/amp/s/www.theverge.com/platform/amp/2018/11/16/18098230/microsoft-windows-on-arm-64-bit-app-support-arm64
This is what I meant when I mentioned that Microsoft was helping to solve the boot camp issue. Theoretically, Window 10 on ARM would work for the desperate one or two occasional work apps some Mac users need, at least.