What will be the PC response to Apple's M1 ?

[Erehy Dobon]

My guess is that the big players will double down on their cloud computing initiatives and ignore trying to compete with Apple's tight hardware-software integration and (fairly) walled application garden.

And not just the office/productivity arena.

Streaming audio (Apple Music, Spotify, Tidal, etc.) and video (YouTube, Disney+, Amazon Video, AppleTV+, Netflix, etc.) are just the entry points.

Stuff like gaming (Nvidia GeForce, Google Stadia, online Xbox Game Pass). VR like Oculus Quest 2.

At some point in the not too distant future, a quality user experience will have less to do with the specs of the device in your hand versus the cloud computing server support that those devices connect to.

 

[deconstruct60]

Perhaps this is a dumb question
But is there any actual reason why Microsoft could not sit down with some amazing chip designers and task them with using ARM cores + custom design and tell them to build a chip, or SOC which is specifically tailored to run Windows and accelerate all aspects of Windows?

Don't have to go to "random" designers.... they could just go to ARM with cash.

https://community.arm.com/developer...ors-ip-blog/posts/arm-cortex-x-custom-program

Similarly what is special that Windows actually needs. The farther up the OS stack from the kernel there isn't much extra special about Windows.

There would be some upside of commissioning at baseline design with Pluton built into a standard design.

Meet the Microsoft Pluton processor – The security chip designed for the future of Windows PCs | Microsoft Security Blog

In collaboration with leading silicon partners AMD, Intel, and Qualcomm Technologies, Inc., we are announcing the Microsoft Pluton security processor. This chip-to-cloud security technology, pioneered in Xbox and Azure Sphere, will bring even more security advancements to future Windows PCs and...

www.microsoft.com

But that is "down" from the kernel; not up .

 

[amartinez1660]

It’s not just the hardware though, it is the tight integration that makes a Mac special.
For example my PC smashes my macs in terms of speed and is great to use. However the integration is terrible and it feels super clunky. So it’s great for productivity in certain apps but I much prefer a Mac for ‘flow’.
No hardware can beat this.

I switched to 100% Mac for this exact reason... when the thing works (granted sometimes it bugs but very rarely), all the flows like copy on the phone paste on the Mac, take a screenshot then it appears on the iPad Pro to be annotated, take a quick break still wearing the headphones (that ALWAYS worked and didn’t need any sort of driver tinkering), do a hey Siri while at it and ask it to increase the volume to 75%, messages, calls, camera always works (again, no drivers tinkering), etc etc etc.

Heck, the very first migration assistant from an older MacBook that I used half and half between bootcamp and MacOS, when I saw that the whole thing transferred without issues and I was up and running in minutes with all the software I was already using while the windows part was looking so painful to transfer... I just left it there, never did it.

 

[amartinez1660]

A response would require a level of cooperation that hasn't really worked in the past. Recall back to the days of PowerPC. Apple was going to be but one vendor and MacOS just one option. We saw CHRP, and Macs adopting PCI, SATA, and other PC standards. There were of course, the clones licensing Mac OS from different vendors. Unfortunately, it fizzled, Steve canceled licensing upon his return, and Apple wound up going to Intel a decade later.

You would need buy-in from the CPU makers. They would need to agree on a common ISA. They would also need to agree on specific components to put in the chips. Part of the reason M1 does so well at video editing is because there is hardware acceleration for encoding/decoding different formats. Likewise, there is the Neural Engine, so macOS devs can start targeting that as well. But Intel, AMD, MediaTek, Nvidia, etc. would need to agree upon a baseline feature set, and be willing to share potential improvements with competitors to avoid fragmentation.

Then, you'd need MS, Red Hat, Canonical, Oracle, etc. to agree to support new features and provide frameworks and APIs for third-party devs.

All of that so that third-party developers have a way of utilizing the improvements without too much fuss.

Apple wasn't able to make it work a few decades past. I'm not sure it would be possible now, especially not with the way Nvidia tends to do business - but that could change?

RISC-V is a neat academic exercise, but we're still a decade or so away from competing with current chips.

On the hardware accelerator chips front, the industry did embrace the use of GPUs as a way to accelerate 3D graphics out of the CPU. What I see is Apple is doing a similar move but on a lot of different tasks.

 

[Fomalhaut]

But is there any reason why PC's/Windows could not move to ARM ?

Why could we not get high end ARM PC's and Msoft have Windows running on it, and app makers recompiling their apps for Windows ARM in the same way they are for Apple's ARM ?

There is no inherent reason prevently Windows moving to ARM other than what we already know:
1) Windows on current non-Apple ARM chips like the Qualcomm Snapdragon has lacklustre performance. You would need Microsoft to team up with a chip design company to build their own version of the M1, optimized for MS Windows.
2) There is such a massive amount of x86/x64 software out there, that it would require a huge effort to migrate it all, and many developers simply won't bother - leaving consumers stuck on Windows. Unless Microsoft develops their own version of Rosetta 2 of course.
3) Do they have business-driven need to do this? If a large proportion of computer buyers start to move to Apple and MacOS, then this would be the impetus, but Apple is not positioned at the budget end of the market, so Intel/AMD/Microsoft can still command a large part of the market.

 

[Fomalhaut]

Once Intel is running on 5nm then the performance advantage will evaporate. We live in a Windows centric world so it’s great for Apple to have a lead but the incompatibilities of these ARM machines mean they’re not useful to most normal users. Mac OS X will continue to be a niche product, even more so due to its inability to run Windows x64 too.

That's a challenging and unsubstantiated statement. What incompatibilities are you referring to?

I received my M1 Mini today and have been doing my normal work on it all day. I haven't encountered a single "incompatibility" with very standard productivity software.

Mail (including MS Outlook), Slack, MS Office, Skype, Dropbox, OneDrive, LastPass, Spotify, Visual Studio Code, Carbon Copy Cloner, Windows Remote Desktop, Final Cut Pro, Paragon NTFS, and of course Safari and Chrome browsers, where let's face it, most of us can do the majority of our work if necessary.

All of the Intel apps I've installed appear to work without issue, and run well.

Yes, there is some software that isn't working yet, especially for developers, e.g. Docker.

What is your definition of a "normal" user? I'm genuinely curious.

In my opinion, a "normal user", will want a computer for work, study, communication and entertainment.

A lot of productivity work can be done with web-based tools, which is the trend for all sorts of business tools (accounting, CRM, expenses, time-management, reports, presentations....). Most study involves note-taking, research, and writing. Maybe some specialist apps for science & engineering. Communication: video conferencing, chat / IM, e-mail. Entertainment: streaming media, web-pages, photo/video editing, and maybe games, which is the only category I would concede is weak on the M1.

So sorry, I simply don't accept your argument. Happy to hear your opinion though!

 

[dmccloud]

From the outside looking in, it appears that Pluton is an example of Microsoft trying to compete with Apple's T2 chip it has been using with their Intel-based systems. The issue with that approach is that Redmond is going where the puck has been instead of where it's going. Microsoft's one big weakness with respect to Windows on ARM is that they are a software company first and foremost, and lack the experience and capabilities required to design their own CPUs or SoCs. Pluton is first and foremost a set of security technologies that can be integrated into existing x86 and Qualcomm CPUs, but it is not something that Microsoft is designing on their own. It would be interesting to get an indepth comparison of Pluton and the T2 or even the M1s secure module to see how they resemble (and differ from) each other, mainly to see how much these two approaches diverge from each other. (*cough* Anandtech, here's a story idea for you)

We believe that processors with built-in security like Pluton are the future of computing hardware. With Pluton, our vision is to provide a more secure foundation for the intelligent edge and the intelligent cloud by extending this level of built-in trust to devices, and things everywhere.

This line from the Microsoft press release makes me wonder if this alone might prevent WoA from ever running on Apple Silicon. Apple is not going to share its secret sauce with competitors, and if Microsoft makes Pluton support required for future OS updates, then it is likely that WoA on Mac could be dead before it ever actually even began. Ironically, the "Seven properties of highly secure devices" Microsoft touts its Azure Sphere as the first IoT device to meet the criteria of is a Microsoft research project, so they're merely reaching their own internal goal rather than any sort of industry-wide standard.

 

[Argon_]

There is no inherent reason prevently Windows moving to ARM other than what we already know:
1) Windows on current non-Apple ARM chips like the Qualcomm Snapdragon has lacklustre performance. You would need Microsoft to team up with a chip design company to build their own version of the M1, optimized for MS Windows.
2) There is such a massive amount of x86/x64 software out there, that it would require a huge effort to migrate it all, and many developers simply won't bother - leaving consumers stuck on Windows. Unless Microsoft develops their own version of Rosetta 2 of course.
3) Do they have business-driven need to do this? If a large proportion of computer buyers start to move to Apple and MacOS, then this would be the impetus, but Apple is not positioned at the budget end of the market, so Intel/AMD/Microsoft can still command a large part of the market.

A MacBook SE a few years from now might gather some of that market share. Trim fat from the current Air wherever possible. Non backlit keyboard, legacy chip, 128 GB storage ~$749. That might be doable, when considering the money saved by using Apple silicon.

 

[thenewperson]

A MacBook SE a few years from now might gather some of that market share. Trim fat from the current Air wherever possible. Non backlit keyboard, legacy chip, 128 GB storage ~$749. That might be doable, when considering the money saved by using Apple silicon.

I'm thinking 799, but same. I don't think they even need to remove some of those things. Just re-use design + components after a few years.

 

[Fomalhaut]

I'm thinking 799, but same. I don't think they even need to remove some of those things. Just re-use design + components after a few years.

Yes, it's certainly possible if Apple wants to play in this space. TBH even the current M1 is likely to compare quite well to future Intel 12th or 13th gen chips in 2-3 years' time, if we consider that it already beats the 11th gen Tiger Lake, and Intel may hope for a 10% year-on-year improvement at best.

 

[UBS28]

Just wondering.
Generally in any technology something is created which appears better in way to what's currently available.
This has been going on since man invented the wheel/rollers.
Someone else saw this and copied it, hopefully over time improving it.
We as a species continued to do this till we found ourselves where we are today.

So Apple launches the M1 which has benefits in real world usage for a specific type of machine.
With the promise that more powerful machines will follow.

The Windows/Linux (PC) world will of course see this, and if promised improvements at the higher ends do materialize will of course change also to gain similar benefits.
I would be crazy, would it not to assume the PC world will simply ignore ARM and the M1 if it does deliver what people are expecting.

So, what do you think will happen, over, perhaps the next decade let's say?
I know AMD are reported to be working on something a little like the M1 in some ways.

I would guess, at some point Intel will do something, albeit very late.

What are your thoughts?

1. AMD, their old 15W CPU is outperforming the M1 in multiscore. So the new AMD low powered chips will be even faster.
2. Microsoft is now focussed on cloud computing. So in theory you can run the most demanding applications even from your phone if you want to (atleast something like this is what is in the mind of Microsoft).

 

[marcuzt]

I am still hoping for more open RISC-based motherboards. Either ARM or RISCV or similar.
Huawei has one in the pipes, but looks rather like slim pickings in that market. Due to software needs to be specific for each type of architecture. There is no RISC-software, so someone (like apple does) need to create their own standard of hardware that will run linux or their own OS.

Few companies are willing to do this because there is no market for it now.

 

[Fomalhaut]

.
I would be crazy, would it not to assume the PC world will simply ignore ARM and the M1 if it does deliver what people are expecting.

So, what do you think will happen, over, perhaps the next decade let's say?
I know AMD are reported to be working on something a little like the M1 in some ways.

I would guess, at some point Intel will do something, albeit very late.

What are your thoughts?

I wouldn't be surprised at all if some people, even manufacturers, will just ignore ARM and M1, even if it demonstrably superior in many cases. You've no doubt seen the denial on tech forums and YouTube comments about how "the iPad chip can't possibly be better than a proper laptop/desktop chip", and "AMD Zen xxx or Intel Gen yyyy will blow the Apple usurper away....".

For some reason technology choice seems to appeal to irrational partisanship...I've never understood this, but looking at recent political events, I guess logic and demonstrable facts have little to do with what people align themselves to....

 

[Fomalhaut]

1. AMD, their old 15W CPU is outperforming the M1 in multiscore. So the new AMD low powered chips will be even faster.
2. Microsoft is now focussed on cloud computing. So in theory you can run the most demanding applications even from your phone if you want to (atleast something like this is what is in the mind of Microsoft).

I think the Apple M1 is a great achievement and is a landmark development for ARM-based CPUs in consumer computing...the first significant player in the space between low-power mobile devices and ARM-server hardware in many years (since Acorn Archimedes, maybe?)

How much this will impact the x86 market remains to be seen. The reason Apple could claim that the M1 is faster than 98% of PCs sold last year, is because the majority of PCs are low-price (and low-spec), and Apple probably won't want to slum it with the budget PC brigade.

If AMD can make even a few of the efficiency improvements that Apple has achieved they will do well, and will probably offer a wider choice of processors than Apple will.

Your point about cloud computing is also relevant. This has been my professional focus for the last 5 or so years, and it has shown me that the power of your client computer is becoming less relevant for many tasks.

1) An increasing percentage of business software is accessed via web-apps, not desktop apps. I recently joined a new cloud consultancy company - none of our business software (expenses, time-management, documentation, recruiting, e-mail, productivity, file storage etc.) depends on desktop apps. It is all web-based running various SaaS (software-as-a-service) apps.
2) Our customer development, test and production environments are all running on cloud infrastructure, and we access these via web-browsers, command-line windows, or light-weight IDEs. We don't need to run any heavyweight software like databases, app-servers on the local machines.

Although it wouldn't be my choice, I *could* do my whole job on a Chromebook or an iPad...but I prefer to have the choice to install some "real" software locally if I want to test some ideas locally, such as node.js or Python apps, and there are some occasions where you don't have Internet connectivity to run web-apps (not often these days).

 

[SlCKB0Y]

For some reason technology choice seems to appeal to irrational partisanship...I've never understood this, but looking at recent political events, I guess logic and demonstrable facts have little to do with what people align themselves to...

Humans have always been, and will always be, innately tribal.

 

[SlCKB0Y]

An increasing percentage of business software is accessed via web-apps, not desktop apps. I recently joined a new cloud consultancy company - none of our business software (expenses, time-management, documentation, recruiting, e-mail, productivity, file storage etc.) depends on desktop apps. It is all web-based running various SaaS (software-as-a-service) apps.

I've been in this area for more than a decade and as the industry shifted from bare metal, to virtualisation to cloud, the software we used followed suit. Across all of our technical teams in terms of desktop apps, their jobs can pretty much be done with a web browser, an SSH client and their text editor of choice (local or otherwise) and very occasionally an RDP client with almost all of them running either Mac laptops with a few on Linux

Everything else is web based, either running on our own infrastructure, AWS or SaaS. Atlassian's apps, Slack, Zapier, GApps, Zabbix/Nagios/Data Dog, PagerDuty, Zendesk, BambooHR, Monday/Asana, Graphana, Xero, Salesforce, Mailchimp, GitHub, Zoom and many more.

The only other really required local apps are Excel for the finance team (Sheets doesn't cut it for them but for the rest of us its fine), and Adobe CS for our design teams and they are all run on Macs.

It' so liberating being apart of a company not tied to Windows and desktop apps.

 

[k2k koos]

I did hear that RISC-V is starting to look pretty amazing, and can do things on a fraction of the power that ARM can do.
And being open, has some amazingly clever people working on it, without the restrictions ARM has on it's design.

No idea how RISK-V will develop into the future, but it's all exciting stuff.
And yes, it needs someone the design the hardware to match the software.

Remember that 'secret' corner in Apple's HQ during Steve Jobs introduction of the Intel switch, where they had Mac OSX as it was then still known, running on Intel in secret for several years? I'm going to bet that there is a similar 'lab' at least experimenting with RISC-V running Mac OS, and developing it, just in case the ARM's race is looking like to come to an end. After all, RISC-V is an open arcitecture, and that open design is what Microsoft and the wider PC customers like. So in the not too distant future, I would not be surprised to see the PC world runnning RISC-V, with Apple doing their own thing, until it becomes clear that RISC-V is starting to nibble away the lead or threathens to overtake it. It won't happen next year, but by the end of this decade, we may well see yet another transition...

 

[the8thark]

The topic is perhaps less to do with the M1 alone and more the integration between it and the rest of the system.

This topic is perhaps less do do with PCs in general (all Apple Macs are PCs) and more to do with, if Intel and AMD will try to innovate their product lineup in the medium to long term future.

We also need to realise that now Intel and AMD are in different market segments. Apple offers the all in one end solution. You buy the Mac. We all do. Intel and AMD on the other hand sell the parts wholesale to the industry to manufacture PCs and individual hobbyists who like to make their own PCs.

Because of these differing needs of the customers, it allows Apple to have the specific co-processors on the M1 specific to what MacOS and the Apple customers need. The Intel and AMD ifferings need to be suitable for a much wider customer base, ie their chips being placed in many different PCs. Thus their product needs to be more generalised to be good for every use case scenario.

Intel and AMD need to be the jack of all trades but it also means they are masters of none.
Apple is master of one thing, the Mac and useless to everything else (ie all non Apple PCs).

Will Intel and AMD change their thinking and move into more specified chips for their larger customers or will they stay with their jack of all trades and just up the cores and ghz to get more performance out of them?

 

[dogslobber]

That's a challenging and unsubstantiated statement. What incompatibilities are you referring to?

There's a reason that Macs are not on the approved vendor list for most companies. E.g. harder to provide IT services for, incompatible with most Enterprise software. It's cute they work for you but you're a mere data point.

 

[KShopper]

In the short term, nothing. For majority of people and the market, Windows is still the only choice, so whatever hardware are offered, people will buy it, regardless of how "bad" they are. Legacy compatibility is the key that lock people in.

There are three types of mainstream computer purchasers that I can see:

1. Forward-thinking productivity workers and technology consumers that see the value of an integrated technology platform, with different devices targeted at

There's a reason that Macs are not on the approved vendor list for most companies. E.g. harder to provide IT services for, incompatible with most Enterprise software. It's cute they work for you but you're a mere data point.

I disagree with this statement. High-function information workers gravitate to the Mac for good reason.

IBM Announces Research Showing Mac Enables Greater Productivity and Employee Satisfaction at IBM

Findings affirm that offering choice delivers strategic value in addition to IT efficiency and lower total cost of ownership...

www.globenewswire.com

 

[dogslobber]

I disagree with this statement. High-function information workers gravitate to the Mac for good reason.

I.B.M. It's not the 1990s anymore and they're totally irrelevant to any IT buying decision not involving tape drives.

 

[deconstruct60]

Remember that 'secret' corner in Apple's HQ during Steve Jobs introduction of the Intel switch, where they had Mac OSX as it was then still known, running on Intel in secret for several years? I'm going to bet that there is a similar 'lab' at least experimenting with RISC-V running Mac OS, and developing it, just in case the ARM's race is looking like to come to an end.

I doubt that. Mac OS X pragmatically ran on Intel before it was "Mac OS X". NeXT had NextStep running on Intel when Apple bought them. They never stopped making/building it on Intel after Apple bought them. So as NextStep mutated to Mac OS X it was just same thing , different day as far as the builds went. They just didn't ship it publicly.

The builds were done as a "sanity check" for implementation specific feature/assumption creep. And it was also an "out". For the next couple of years x86-64 or perhaps some other ARM implementor would be a bigger sanity check than RISC-V.

Two major differences now :

1. 100M x86-64 install base. Apple has to support x86-64 macOS anyway for more than a couple of years. So Apple already has an "alternative" platform implementation to build against that is part of their "day job". They don't need another secret build to do a sanity check on. When Apple gets to the point of winding down x86-64 support to nothing then there would be an "empty" slot for the alternative build.

So this is all quite similar to when NeXTStep already ran on Intel and easiest thing to do is just build when have been building.... just don't ship it. ( similar if Apple opens a much wider fork between ARM instruction set evolution and Apple's implementation. )


2. Currently, it isn't clear if Nvidia is going to get ARM. That may not pass regulators. If it does pass and there are contractual "hands off and independent" operations restrictions on Nvidia moving ARM out of the path they have been on for last couple of decacdes, then the window for RISC-V isn't going to be the same. ARM future uncertainty is opening the door on RISC-V.


Actually there is a 3rd difference in that x86 "world" is under heavy assault. That actually might spur either AMD or Intel (or both) to do a somewhat revolutionary jump and dump lots of the legacy baggage holding x86_64 back. Dump all the 32-bit stuff. Dump most of the hodge podge of vector solutions ( SSE , SSE4 , AVX v1 , .... ) . Stop making bleeding edge processors to run 1990-2005 code well in 2022+

Only if AMD, Intel, and Nvidia muck it all up and make RISC-V a major investment pool for the rest of the industry ... then yeah ... fire up a RISC-V build. It would take several years of those three (and Nvidia getting control of ARM) to set the preconditions for that though.

After all, RISC-V is an open arcitecture, and that open design is what Microsoft and the wider PC customers like. So in the not too distant future, I would not be surprised to see the PC world runnning RISC-V, with Apple doing their own thing, until it becomes clear that RISC-V is starting to nibble away the lead or threathens to overtake it. It won't happen next year, but by the end of this decade, we may well see yet another transition...

Apple can just 'fork' off of the ARM instruction set with their architectural license. Apple could just stop buying the future arch evolutionary improvements. That is about just as 'free' ( 'free beer' ) as RISC-V.

If Nvidia ( or whoever ARM's fuure owner turns out to be ) screws up the licensing model perhaps Apple will change. But paying ARM to lay down some common shared baseline..... Apple isn't really hurting for cash to throw into the common R&D pot.

It is unclear whether RISC-V is going to follow a balkanized *BSD path of open source or there will be some kind of "glue" that holds Linux more coherent together commercially. And any open architecture that took on GPL like zealotry Apple would avoid like the plague.





P.S. I wouldn't be surprised if someone at Apple did a build of kernel or something limited on RISC-V for giggles. If only to figure out a rough scope of how much tool chain work there is to do there. But it probably isn't in the "Plan B" stage for now. And all of the 32-bit RISC-V implementations are in the "no fly" zone.

 

[spiderman0616]

The whole premise of the OP's original question is flawed. "The PC" can't just simply respond to the M1, because no single company owns the PC platform. Intel, AMD, Nvidia, Microsoft, and all the third party PC makers that are any kind of successful all have a death grip on the PC platform. It doesn't make any sense that they would be able to just respond to this in a unified way.

Even if you just reduce this problem down to AMD and Intel, who are fierce competitors, the problem spirals out of control quickly. What happens when they each make their own SoC style chip like the M1? Does Microsoft have to maintain versions of Windows/Office/etc. for each one? Do they have to cooperate with each other to make sure they're both making a chip that is compatible with everything? What does Nvidia do in that situation as far as gaming GPUs? Serious gamers are still going to want those. Does Nvidia have to maintain skus for each kind of SoC out there in the PC ecosystem? The M1 is a custom designed system that was created to run a specific software/hardware platform. The PC isn't really like that in a lot of situations, especially gaming.

I'm also not sure what makes anyone think Microsoft would be able to keep up with that fiasco when they are still struggling to put Windows on ARM in a way that's anywhere near acceptable for serious professionals or gamers. And this isn't my attempt to sit here and rip on Microsoft. I couldn't care less what they do with Windows because I don't use it. Microsoft is gonna Microsoft. But this is a REALLY hard problem to solve when you have all these different 3rd party vendors involved. It would be very pie-in-the-sky thinking to assume that they're all going to magically get along and respond to Apple any time soon.

 

[Argon_]

The whole premise of the OP's original question is flawed. "The PC" can't just simply respond to the M1, because no single company owns the PC platform. Intel, AMD, Nvidia, Microsoft, and all the third party PC makers that are any kind of successful all have a death grip on the PC platform. It doesn't make any sense that they would be able to just respond to this in a unified way.

Even if you just reduce this problem down to AMD and Intel, who are fierce competitors, the problem spirals out of control quickly. What happens when they each make their own SoC style chip like the M1? Does Microsoft have to maintain versions of Windows/Office/etc. for each one? Do they have to cooperate with each other to make sure they're both making a chip that is compatible with everything? What does Nvidia do in that situation as far as gaming GPUs? Serious gamers are still going to want those. Does Nvidia have to maintain skus for each kind of SoC out there in the PC ecosystem? The M1 is a custom designed system that was created to run a specific software/hardware platform. The PC isn't really like that in a lot of situations, especially gaming.

I'm also not sure what makes anyone think Microsoft would be able to keep up with that fiasco when they are still struggling to put Windows on ARM in a way that's anywhere near acceptable for serious professionals or gamers. And this isn't my attempt to sit here and rip on Microsoft. I couldn't care less what they do with Windows because I don't use it. Microsoft is gonna Microsoft. But this is a REALLY hard problem to solve when you have all these different 3rd party vendors involved. It would be very pie-in-the-sky thinking to assume that they're all going to magically get along and respond to Apple any time soon.

The PC building component of this is an interesting discussion.

I wonder if there is a viable intermediate between the current system of tower PC assembly, and full SoC. Imagine a motherboard with a CPU socket, a GPU socket, and RAM in between. You could get the modularity that PC builders want, while having 'unified memory' and a faster interface than PCIe. It would also do great things for GPU cooling.

 

[Internaut]

History teaches us that the rest of the PC business has plenty of time to respond. The first Commodore Amiga was in many ways better than pretty much any PC or Mac of its time. Where is the Amiga now? The Acorn Archimedes* was the fastest personal computer in the world, at launch. Where is Acorn?**

* First ARM; mid 80s.
** They became ARM and the Acorn RISC Machine became Advanced RISC Machine.