NVidia and MediaTek to collaborate on ARM gaming laptops

Also, the only console in the 1990s that used MIPS was the original PlayStation, and only the GameCube, Wii, and PS3 used PowerPC in the 2000s.

N64, Dreamcast, and PS2 were all MIPS.

dont forget the original Xbox with its x86 processor and the Xbox 360 with a triple core PowerPC proc!

also, though it’s a bit more of a tangent, I believe most of the portables have all been arm based, but none arm64 until the switch?
 
dont forget the original Xbox with its x86 processor and the Xbox 360 with a triple core PowerPC proc!

also, though it’s a bit more of a tangent, I believe most of the portables have all been arm based, but none arm64 until the switch?
Crap, that's right, the 360 was also PowerPC - thanks for the reminder! :D
 
Oh, so you do think CPU power will still matter. :)
not so much CPU, I mean our Payroll/HR/Timesheet/Pension system for 2000+ still has the recommended requirements that an i3 with 8gb of ram can tackle it no problem. But that's asking for a HDD failure, networking and read/write, and backup is far more important than CPU at this stage. And security I mean having that running on a laptop which most offices use now for user mobility is just asking for somebody to lose it in an Uber and have that disappear to god knows where. That deserves to live on a server in a closet with UPS, backups, and blah blah blah. The only thing that should be running client-side is the client which again you can get away with a low-end i3 and 4GB of ram if you really want to subject your staff to that punishment, I mean Chrome has higher requirements than most office applications at this stage. But CPU will always "matter" but there is a soft floating point where after so much power it become irrelevant because CPU is not the bottleneck.

So really CPU matters based on the use case, but we've reached a stage where that requirement is pretty low for the day-to-day essentials, it's the entertainment factors that put the heavy load on consumer hardware at this point.
 
That's a fascinating theory and I bet people running payroll or ERPs or the like (just as a couple of OTTOMH examples) would disagree.
Tell those people I said they're stupid. Like many people already pointed out if you need serious compute power you won't be doing it on a laptop. Also nowadays any video encoding is done through hardware acceleration, not on the CPU. At least it shouldn't be done on the CPU. Web browsers use 3D acceleration as well as hardware video decoding for any video watching.
But why will it slump? ARM for the Mac isn't the same as for Linux and Windows. The "headaches" aren't really there; you can generally assume software and peripherals will work, and Rosetta-translated apps run well enough as a rule. If they don't work, there's usually a good reason for it (eGPUs are the biggest concern right now).
Windows x86 is a force that even Microsoft can't control. Any application or program ever made will most certainly be done for Windows x86, and in many cases, only for Windows x86. As a Linux user this sucks, but as a Mac ARM user this double sucks. That doesn't mean that all Windows x86 applications won't run well on ARM, it just increases the odds that it won't. Just the nature of emulation.
Besides, given that sales of some of Apple's most popular systems are surging right now, saying it won't be welcome flies in the face of evidence.
Apple users tend to leap before they look.
Please drop the childish fan wars. Macs are fine.
No.
How many Mac users do you know are playing Cyberpunk 2077 right now, or desperately wish they could? Not many
Firstly, you and I don't know that. Secondly, that's not my point. Cyberpunk 2077 is just one example of what M1 owners can expect. Assassin's Creed Valhalla doesn't work, and Borderlands 3 is unplayable. Nvidia has decades of experience working on graphics and whatever Nvidia's mobile ARM SoC they make will likely be better than Apple's M1. Also Nvidia owns ARM now so that matters too. It isn't just AMD and Intel that Apple should worry about, as Nvidia is likely going to produce something far better than the M1, despite that I also feel it'll fail at the market it's trying to attract.
And an iPhone isn't a Mac, just as your Android phone isn't the same as your PC.
I think we'll see iPhones and iPads with Apple M1 hardware in them.
Even if AMD doesn't mess up, they are still giving a lot of competition with x86-64, and without them, ARM would have been the dominant CPU ISA much sooner.
The same could have happened with PowerPC. The thing about x86 is that IBM made sure that Intel didn't dominate that market and forced them to let AMD compete. We're still benefiting from this today. Keep in mind that Apple's M1 is only this good because x86 is still good. The moment ARM is on top is when nobody is going to try and engineer it better. This is why Intel today is a joke because they didn't fund enough R&D, or put R&D into other markets since Intel dominated the PC market. Remember when Intel tried to make x86 on mobile phones and tablets? Intel wasted a lot of money and gave up fairly quickly.
I don't think ARM is going to stay exclusively in the mobile space forever, as they (excluding Apple's ARM) are gaining a lot of traction in the proprietary and non-proprietary server market, supercomputers, and niche workstations.
Probably more to do with cost and freedom, as Intel doesn't let anyone make whatever they want. Remember the Fujitsu A64FX? That thing is badass but it's mostly a general purpose GPU. It'll be interesting to see what Nvidia allows or doesn't allow others to make with ARM. You know they're going to be dicks about it?
Well, Rosetta 2 is technically a translation application, which is one level lower than emulation if I remember correctly.
Translation is much faster and more efficient, and requires far less work, than straight emulation like what Microsoft was doing with Windows RT.
From what I understand the performance is mostly from the Apple M1 having hardware similar to x86. Mostly in the memory department. AVX512 might be something that not many people consider with x86 as Yuzu author was able to "accelerate ARM's vector bit-reversal into a just 2 fast instructions". This means that x86 doesn't need special hardware to emulate ARM, as AVX512 can basically do it all. The only problem is that it gets the Intel chips really hot and consume a lot of power.

Will Nvidia try to make their SoC ARM devices play nice with x86? Hell no, they'll force developers to develop for ARM. Kinda like how Sega did with the Saturn, and look how well that turned out. Nvidia should really put more money into software development if they want quicker adoption. At least Apple is doing that fairly well.
 
Last edited:
ARM won't stagnate just because there on top. As long as no one company buys ARM and does anything crazy. Yes if Nvidia gets ARM and starts releasing their own chips with new ISA as they announce it then yes ARM will die. As long as they or who ever helms ARM stays the course ARM will never stagnate. Even in Mobile where there really hasn't been any competition ISA wise in a decade we see progress. Its simple any company can license the ISA and build a chip. Something like 10 companies have architecture licenses allowing them to build their own cores following nothing but the ISA white papers.

x86s downfall isn't that its old or slow or bad really. Its that Intel has held too tightly to the ISA. Yes AMD and Cyrix got overflow work that lead to licenses... but otherwise Intel has kept anyone else out. Nvidia wanted a license around the time they where producing x86 chipsets and when Intel said buzz off... they bought 3DFX. Which had nothing to do with wanting 3DFX... they simply wanted the real3d patents 3DFX owned so they could try and squeeze a x86 license out of Intel. Intel would rather pad Nvidias profit margin for a decade then part with a license. Which imo is exactly why Jensen is hot on owning ARM. He is coming for Intels server cookies... and if he can be seen as ARM it will make it all that much sweeter should ARM diminish Intel.

As for gaming on ARM... ya I don't see the issue. As you point out.... 90% of anything that requires acceleration is using non CPU parts to do it. That includes games... unless I was dreaming that any game made in the last 20 years requires a GPU. Developers target GPUs not CPUs. There is really nothing in any modern game that requires x86 to operate. All the calls are to Shader languages and very very high level object programming. Hell probably more then 3/4 of all AAA titles are using the same 2 major engines. (Which both fully support compiling out to other ISAs already to cover mobile) It really wouldn't be that hard to switch game developers over to ARM. For the most part it would involve pretty minor work and a recompile. Outside of perhaps ID (and I don't think its true since Carmack moved on) no one is programming directly for CPU stuffs. Low level GPU stuff sure... but low level CPU anything ??? No... it can all be recompiled if they had a reason to.
 
Firstly, you and I don't know that. Secondly, that's not my point. Cyberpunk 2077 is just one example of what M1 owners can expect. Assassin's Creed Valhalla doesn't work, and Borderlands 3 is unplayable. Nvidia has decades of experience working on graphics and whatever Nvidia's mobile ARM SoC they make will likely be better than Apple's M1. Also Nvidia owns ARM now so that matters too. It isn't just AMD and Intel that Apple should worry about, as Nvidia is likely going to produce something far better than the M1, despite that I also feel it'll fail at the market it's trying to attract.
It's a reasonable statement given that most Mac purchases are laptops, and laptops that clearly aren't meant for gaming at that.

And do you really think someone who already has an Intel Mac is fretting over whether or not AC Valhalla or Borderlands 3 will run properly? They'd already have to go into Boot Camp just to run these titles, and again, the number of Mac users who both have a powerful-enough system and the determination to play those games is rather small. I have a higher-end iMac that could, but most Mac users don't buy iMacs, let alone high-end ones.

Also, it's too soon to speculate about what NVIDIA can manage if it succeeds in buying ARM (remember, the deal hasn't even been approved so far). Owning the architecture designer helps, but that isn't a magic fairy dust that automatically makes the chips faster. And besides, the first fruits of an NVIDIA/ARM combo are years away. This potential product will compete against a future Apple chip, not the M1.


I think we'll see iPhones and iPads with Apple M1 hardware in them.
Nope. The M1 is based on Apple's A14; there wouldn't actually be much to gain from it, especially since the M1 was mainly optimized to accommodate Macs.
 
It's a reasonable statement given that most Mac purchases are laptops, and laptops that clearly aren't meant for gaming at that.

And do you really think someone who already has an Intel Mac is fretting over whether or not AC Valhalla or Borderlands 3 will run properly? They'd already have to go into Boot Camp just to run these titles, and again, the number of Mac users who both have a powerful-enough system and the determination to play those games is rather small. I have a higher-end iMac that could, but most Mac users don't buy iMacs, let alone high-end ones.
Don't forget that ThunderBolt eGPUs can be used on the Intel-based Macs for gaming, which a surprising amount of people use.
For those with just the vanilla Mac, or ARM-based Macs, yes, gaming is most likely not their first priority.
 
It's kinda hard to tell how x86 vs ARM works in efficiency since there's no Apple to AMD/Intels comparison.

We have no x86 legacy, like 32-bit support and things like that,” said Hegde. “We are able to optimize our code, and our core area is significantly smaller [as a result]. Just to give you an idea, in the previous generation, if you look at ThunderX2, compared to AMD or Skylake, for the same process node technology [we get] roughly 20% to 25% smaller die area. That translates into lower power. When we move to 7nm with ThunderX3, our core compared to AMD Rome’s 7nm is roughly 30% smaller.”

https://www.hpcwire.com/2020/03/17/marvell-talks-up-thunderx3-and-arm-server-roadmap/

That was back before the PS4 and Xbox One was released. Also who cares?

The same reasons why they preferred ARM then continue being valid today.

I don't think people understand how Apple going x86 was such a good move.

I don't think people understand how Apple going ARM is such a good move.
 
ARM is certainly more efficient at lower power, but the higher the power envelope the smaller the performance gap becomes. I mean if you take the latest 10w ARM cores and put them against the latest 10w x86 cores from either Intel or AMD and those x86's get pummeled, you bring both those to 250w and suddenly that performance gap closes and now you find them basically trading back and forth. So yes ARM is very scalable and that is what makes it great but in the sub 15w power envelope is where it really stands out.

You are entirely right. The x86 tax has a fixed part and a variable part. The larger and more complex the design, the less weights the fixed portion of the tax and the ARM-x86 efficiency gap is reduced, but the gap is still significant at the ~ 200W level.
 
Don't forget that ThunderBolt eGPUs can be used on the Intel-based Macs for gaming, which a surprising amount of people use.
For those with just the vanilla Mac, or ARM-based Macs, yes, gaming is most likely not their first priority.
I wouldn't say there's that many eGPU users. Surprising, maybe, but not enough for Apple to lament lost customers. Odds are it'll more than make up for that with all the additional people buying MacBooks.
 
Well, I got the "no M1 in an iPad" bit wrong. Still not expecting it in an iPhone, and this doesn't mean the iPad will run Mac apps, but it also means iPads with up to 16GB of RAM... egads! They already make good use of 6GB.

One important note from Apple's event: Tim Cook noted that M1 Macs already make up most of Apple's computer sales. So much for ARM-based Macs prompting a decline.
 
I'd be more interested in seeing if they extend this to desktops also. If they are going to focus on laptops only, and ARM is the future of gaming... well I'm not looking forward to regressing to the point where a 17" monitor is considered large and you "upgrade" by tossing your entire computer (and monitor) in the trash every few years, buying a new one instead. That's laptop gaming.
With a library of over a thousand games I am not going to be jumping on the ARM bandwagon so soon for gaming. It's guaranteed that most games will not see an update for ARM support. I'm also not interested in emulation unless it can be proven that there is no loss in performance compared to running natively on x86. That includes latency in addition to framerate and pacing.
Nvidia and Collaborate in the same sentence, that was worth a laugh right there.
NVIDIA is one of the leading contributors and promoter members in the Khronos Group. You know, that group that maintains that little known cross-platform graphics API everybody loves called Vulkan.
 
NVIDIA is one of the leading contributors and promoter members in the Khronos Group. You know, that group that maintains that little known cross-platform graphics API everybody loves called Vulkan.
It's no wonder that the british government is investigating Nvidia's Arm acquisition on national security grounds. Everyone knows Nvidia's pattern when it comes to producing closed proprietary technology standards. A one off khronos fluke won't detract from the well known Nvidia operating procedure.
 
Last edited:
ARM won't stagnate just because there on top.
It's happened before to many CPU architectures and it'll happen to ARM. Once something is established then moving away from it is difficult and therefore the manufacturer can get lazy and cheap about improvements. ARM may have already hit that point since it's everywhere in Android and iOS devices. Apple's M1 is their own design because Apple needs to compete with Intel. ARM did need to get bought out by Nvidia to prevent itself from imploding. Nvidia is certainly going to push for improvements on ARM since they too need to worry about a future of APU's. The only thing pushing ARM forward are companies who are looking to create their own walled gardens and maintain them.
x86s downfall isn't that its old or slow or bad really. Its that Intel has held too tightly to the ISA.
Eventually that might be the downfall of x86. The strength of x86 is the IBM compatible fiasco that gave rise to the open nature of x86. Nobody else can make an x86 CPU, but you're also not tied to the ecosystem by any manufacturer. ARM on the other hand is generally tied to an ecosystem but anyone can make a CPU.
As for gaming on ARM... ya I don't see the issue. As you point out.... 90% of anything that requires acceleration is using non CPU parts to do it. That includes games... unless I was dreaming that any game made in the last 20 years requires a GPU. Developers target GPUs not CPUs. There is really nothing in any modern game that requires x86 to operate. All the calls are to Shader languages and very very high level object programming. Hell probably more then 3/4 of all AAA titles are using the same 2 major engines. (Which both fully support compiling out to other ISAs already to cover mobile) It really wouldn't be that hard to switch game developers over to ARM. For the most part it would involve pretty minor work and a recompile. Outside of perhaps ID (and I don't think its true since Carmack moved on) no one is programming directly for CPU stuffs. Low level GPU stuff sure... but low level CPU anything ??? No... it can all be recompiled if they had a reason to.
You're right but you still need the software to be ported to ARM and to an OS. Windows x86 dominates so hard in this area that Nvidia isn't going to come along with an ARM based laptop and put so much as a dent in it. You would need an ARM cpu with a 1:1 emulation of x86 to even get anywhere. If every game developer released the source code to their games and allowed the community to port them, then ARM might have a chance.

The idea that Nvidia could be using MediaTek as a way to push for Geforce Now maybe a sad reality of what this business venture is all about. I'm sure even Nvidia doesn't have enough hubris to try and take on Windows x86.

It's a reasonable statement given that most Mac purchases are laptops, and laptops that clearly aren't meant for gaming at that.
That's not a reasonable statement at all.
And do you really think someone who already has an Intel Mac is fretting over whether or not AC Valhalla or Borderlands 3 will run properly? They'd already have to go into Boot Camp just to run these titles, and again, the number of Mac users who both have a powerful-enough system and the determination to play those games is rather small.
No? Pretty sure that's my point, in that Intel mac users don't have the problems of running x86 games like the M1 users would.
Also, it's too soon to speculate about what NVIDIA can manage if it succeeds in buying ARM (remember, the deal hasn't even been approved so far). Owning the architecture designer helps, but that isn't a magic fairy dust that automatically makes the chips faster. And besides, the first fruits of an NVIDIA/ARM combo are years away. This potential product will compete against a future Apple chip, not the M1.
When I say M1, that includes future ARM based products from Apple. I don't think many people here understand how much Apple shit the bed by going ARM. Intel makes a lot of technologies that modern computers use, including Thuderbolt. Intel made it clear that they own Thunderbolt. Apple already had problems with Nvidia a long time ago, and now Nvidia is buying ARM which will impact Apple in some way. AMD is the only supplier of GPU's that Apple could use since Intel and Nvidia will obviously not be working with them, and AMD right now is too busy supplying their products to Apple's competitors. While the Apple GPU isn't bad, but it pales in comparison to what AMD and Nvidia offer.

If Nvidia gets serious about making good SoC's with their graphics, then Apple is going to have other competitors to deal with. Nvidia was already working on ARM products, so anything Nvidia is doing is not years away. Maybe months away, as they are working on something for Nintendo. It maybe months away to see an ARM based Nvidia server product, but we're not talking about those products.
Nope. The M1 is based on Apple's A14; there wouldn't actually be much to gain from it, especially since the M1 was mainly optimized to accommodate Macs.
In the essence of being cheap and lazy, I think Apple would benefit from manufacturing as few different kinds of chips as they could. Having multiple products based on the M1 would be a smart move.
 
Last edited:
You talk about ARM like there is only one ARM manufacturer. Your right Apple went after Intel... they in fact replaced Intel chips in their own machines. But do you not believe someone is going to come after M1? It seems clear to me Nvidia will... Microsoft is building a ARM CPU for surface machines... which will also target not only Apple but also Intel as they will also be replacing Intel. If Intel becomes uncompetitive, then Microsoft and Apple duel it out... perhaps with Nvidia Samsung or any other player that wants to. x86 is locked down... we got AMD through a legal fluke that got them a license.

I would also suggest x86 isn't open at all. How can you call it open... open to who. A handful of companies are allowed to what build mother boards ? Nvidia got chased out of chipsets over a decade ago as did Via. Intel does not foster a open platform. AMD and Intel are not even interchangeable any more.
 
Intel dropped the ball hard, AMD is struggling to meet the demand they currently have and can’t take on supplying much more. Manufacturers have to either beg AMD to find them supply, put in an under performing Intel chip, or build their own chip. Apple made a call and I think it was a good one with the M1, and it will give them an advantage especially as more Mac developers release their ARM native applications. For the PC market space it’s going to be hard for ARM to make inroads, not impossible but not easy, there’s already so much to try and support there and x86 was the one constant, unless Microsoft and the Linux kernel guys can come up with one hell of a translation layer (not an emulator) that is going to be one hell of a hard sell.

Apple had an advantage here that their latest 2020 models weren’t running the fastest Intel options so it was possible to release an ARM chip that ran comparably or better. So they can say their M1 variants are upgrades. In the PC space unless something drastic happens I don’t see that same tipping point arriving any time soon for general computing.
 
  • Like
Reactions: ChadD
like this
It's happened before to many CPU architectures and it'll happen to ARM. Once something is established then moving away from it is difficult and therefore the manufacturer can get lazy and cheap about improvements. ARM may have already hit that point since it's everywhere in Android and iOS devices. Apple's M1 is their own design because Apple needs to compete with Intel. ARM did need to get bought out by Nvidia to prevent itself from imploding. Nvidia is certainly going to push for improvements on ARM since they too need to worry about a future of APU's. The only thing pushing ARM forward are companies who are looking to create their own walled gardens and maintain them.

Eventually that might be the downfall of x86. The strength of x86 is the IBM compatible fiasco that gave rise to the open nature of x86. Nobody else can make an x86 CPU, but you're also not tied to the ecosystem by any manufacturer. ARM on the other hand is generally tied to an ecosystem but anyone can make a CPU.

You're right but you still need the software to be ported to ARM and to an OS. Windows x86 dominates so hard in this area that Nvidia isn't going to come along with an ARM based laptop and put so much as a dent in it. You would need an ARM cpu with a 1:1 emulation of x86 to even get anywhere. If every game developer released the source code to their games and allowed the community to port them, then ARM might have a chance.

The idea that Nvidia could be using MediaTek as a way to push for Geforce Now maybe a sad reality of what this business venture is all about. I'm sure even Nvidia doesn't have enough hubris to try and take on Windows x86.


That's not a reasonable statement at all.

No? Pretty sure that's my point, in that Intel mac users don't have the problems of running x86 games like the M1 users would.

When I say M1, that includes future ARM based products from Apple. I don't think many people here understand how much Apple shit the bed by going ARM. Intel makes a lot of technologies that modern computers use, including Thuderbolt. Intel made it clear that they own Thunderbolt. Apple already had problems with Nvidia a long time ago, and now Nvidia is buying ARM which will impact Apple in some way. AMD is the only supplier of GPU's that Apple could use since Intel and Nvidia will obviously not be working with them, and AMD right now is too busy supplying their products to Apple's competitors. While the Apple GPU isn't bad, but it pales in comparison to what AMD and Nvidia offer.

If Nvidia gets serious about making good SoC's with their graphics, then Apple is going to have other competitors to deal with. Nvidia was already working on ARM products, so anything Nvidia is doing is not years away. Maybe months away, as they are working on something for Nintendo. It maybe months away to see an ARM based Nvidia server product, but we're not talking about those products.

In the essence of being cheap and lazy, I think Apple would benefit from manufacturing as few different kinds of chips as they could. Having multiple products based on the M1 would be a smart move.
You make some good points but Apple is in a unique position with their ARM license in that they own it, it would be a difficult legal battle for NVidia to take that away and their attempts to do so would doom the platform. The new iMac’s still have Thunderbolt, though be it v3 and not v4 but it’s there. And the USB 4 specifications are dangerously close to matching TB on the feature front while offering greater compatibility and if FireWire taught us anything it’s that Apple has no qualms with dropping IO support.

NVidia should actually be rooting for Apple here, if NVidia gets the license than Apples wins become their wins. Anybody who advances the platform is doing everybody else a favour because of how competitive and open the ARM platform is through its licenses and that license set is where ARM’s strengths really come out.
 
You talk about ARM like there is only one ARM manufacturer. Your right Apple went after Intel... they in fact replaced Intel chips in their own machines. But do you not believe someone is going to come after M1?
Right now it seems AMD doesn't care about the M1, and it makes sense. AMD hasn't had much of a market in Apple products when it comes to CPU's, and clearly Apple didn't consider AMD when they were upset with Intel. It doesn't look like Apple will be using AMD GPU's either. For AMD their main competitor is still Intel, because Intel still has massive control over the market. The only manufacturer who could compete with Apple is AMD and they don't care about a market they can't penetrate.
It seems clear to me Nvidia will...
Nvidia will but Nvidia is going after the gaming market, where Apple has largely ignored. It won't be hard for Nvidia to make an ARM SoC that will outperform the M1, but it'll be harder to penetrate the Windows x86 market with a Windows ARM platform. Assuming Nvidia will use Windows in their future products.
Microsoft is building a ARM CPU for surface machines... which will also target not only Apple but also Intel as they will also be replacing Intel. If Intel becomes uncompetitive, then Microsoft and Apple duel it out... perhaps with Nvidia Samsung or any other player that wants to. x86 is locked down...
Microsoft's Surface products are a joke, and the ARM versions are laughably bad. The whole purpose of these devices were to directly compete against Apple since nobody else made similar products like Apple, but there's a reason for this. People who want overpriced thin and light products will go Apple, and nothing else. The Windows market is dominated by cheap and expandable products. This is also why Chromebooks are doing so well, as the cheap and light products are what consumers want, and those consumers generally don't care what OS or CPU exists in these devices.
we got AMD through a legal fluke that got them a license.
No we didn't. We got AMD because IBM wanted a second source to the CPU's manufactured for them. They signed a 10 year agreement that Intel never honored. IBM wanted a second source of these 8086 chips and Intel agreed but eventually fought back and lost a court case. IBM didn't want Intel to have exclusive control over x86, which is why we had AMD, Cyrix, and even IBM made their own x86 CPU's.


I would also suggest x86 isn't open at all. How can you call it open... open to who. A handful of companies are allowed to what build mother boards ? Nvidia got chased out of chipsets over a decade ago as did Via. Intel does not foster a open platform. AMD and Intel are not even interchangeable any more.
Neither x86 or ARM are truly open, but both have their ups and downs. x86 CPU manufacturing is closed to mostly Intel and AMD, but because of IBM's mistakes and the IBM clones, the actual platform that x86 mostly sits on which is IBM Clones is open. ARM is open to anyone to manufacture, but most ARM devices exist in a closed ecosystem that actively prevents anyone from installing any software they please. This is why Android is so segmented in OS versions because it's up to each hardware manufacturer to supply an update to these devices, and of course it would be unwise to support a product that is no longer profitable. In fact, the lack of support would be better to promote people to buy new products with the newer Android OS. Apple doesn't have these issues but as we can see, Apple updates aren't always to the benefit of their customers when they can actually slow down the device with updates. Apple updates can even break devices if you replaced the screen or other parts that aren't from Apple, thus giving Apple total control over a product you supposedly own.

AMD is struggling to meet the demand they currently have and can’t take on supplying much more. Manufacturers have to either beg AMD to find them supply, put in an under performing Intel chip, or build their own chip.
If I were AMD, this is a problem I would welcome. Sucks for consumers but great for AMD's bottom line.
Apple made a call and I think it was a good one with the M1,
Everyone knew this was going to happen because it just made sense for Apple. Apple wasn't happy with Intel either, since Intel was lagging behind other manufacturers, but Apple could have gone with AMD if they didn't want to disrupt their consumers work flow. Going ARM is a great move for Apple, but a bad move if you're a consumer of Apple Mac products. It will certainly divide the customer base while giving M1 users a worse experience overall.
For the PC market space it’s going to be hard for ARM to make inroads, not impossible but not easy, there’s already so much to try and support there and x86 was the one constant, unless Microsoft and the Linux kernel guys can come up with one hell of a translation layer (not an emulator) that is going to be one hell of a hard sell.
What ARM needs is a desktop product that works similar to x86, which nobody has or wants to make. To make an ARM motherboard with a socket that you can install in a case would standardize ARM and that's not what manufacturers want. Manufacturers wants walled gardens that they can control. Linux has no future unless Microsoft screws up, and I'm not sure when that will happen. While I use Linux daily and enjoy using it, I can't see a future without native apps being ported over to it. Trying to get Wine working is a 50/50 chance for each application.
 
Last edited:
You make some good points but Apple is in a unique position with their ARM license in that they own it, it would be a difficult legal battle for NVidia to take that away and their attempts to do so would doom the platform. The new iMac’s still have Thunderbolt, though be it v3 and not v4 but it’s there. And the USB 4 specifications are dangerously close to matching TB on the feature front while offering greater compatibility and if FireWire taught us anything it’s that Apple has no qualms with dropping IO support.

NVidia should actually be rooting for Apple here, if NVidia gets the license than Apples wins become their wins. Anybody who advances the platform is doing everybody else a favour because of how competitive and open the ARM platform is through its licenses and that license set is where ARM’s strengths really come out.
Nvidia will never take any licenses away. But ya Apples is untouchable. People forget that Apple is not just a license holder they where actually in on ARM all the way back to the first Acorn chip. Apple helped develop the first mobile ARM chip all the way back in 92. If Apple had been the Apple of today in the late 90s I'm sure ARM would have become a Apple division all the way back then. They did however cover their legal bases for eternity on their own use of the ISA they helped develop.

I am sure Nvidia is rooting for Apple hard. If M1... M2 ect win market share. The PC industry will be clamoring for ARM chips. Who better to provide them then the company that actually owns ARM. Even if that falls through Nvidia will still has the same parts on offer anyway. Owning ARM is just about mindshare. Seeing them come now at the server market owning ARM makes tons of sense. When the bean counters are signing multi million dollar contracts for big iron, it will help if the company they are dealing with is seen as ARM itself. No worries about buying a dead end ARM server chip that will loose support. There where big players that tested the waters on earlier ARM server plays and basically got burnt. Owning ARM makes that a easier sell for Nvidia.
 
Nvidia will never take any licenses away. But ya Apples is untouchable. People forget that Apple is not just a license holder they where actually in on ARM all the way back to the first Acorn chip. Apple helped develop the first mobile ARM chip all the way back in 92. If Apple had been the Apple of today in the late 90s I'm sure ARM would have become a Apple division all the way back then. They did however cover their legal bases for eternity on their own use of the ISA they helped develop.

I am sure Nvidia is rooting for Apple hard. If M1... M2 ect win market share. The PC industry will be clamoring for ARM chips. Who better to provide them then the company that actually owns ARM. Even if that falls through Nvidia will still has the same parts on offer anyway. Owning ARM is just about mindshare. Seeing them come now at the server market owning ARM makes tons of sense. When the bean counters are signing multi million dollar contracts for big iron, it will help if the company they are dealing with is seen as ARM itself. No worries about buying a dead end ARM server chip that will loose support. There where big players that tested the waters on earlier ARM server plays and basically got burnt. Owning ARM makes that a easier sell for Nvidia.
And the more people making them the more people licensing themselves which makes NVidia money. And the more the platform advances the more NVidia can roll those advancements into their Server and Workstation platforms. Consumer PC space makes big money on volume but NVidia wants the server, datacenter, supercomputer market. That’s where the future and the money is moving too. More competition in the consumer space will race prices down, which is great, but custom SOC’s for high end specialized tasks is a big growth industry.
 
Right now it seems AMD doesn't care about the M1, and it makes sense. AMD hasn't had much of a market in Apple products when it comes to CPU's, and clearly Apple didn't consider AMD when they were upset with Intel. It doesn't look like Apple will be using AMD GPU's either. For AMD their main competitor is still Intel, because Intel still has massive control over the market. The only manufacturer who could compete with Apple is AMD and they don't care about a market they can't penetrate.

Nvidia will but Nvidia is going after the gaming market, where Apple has largely ignored. It won't be hard for Nvidia to make an ARM SoC that will outperform the M1, but it'll be harder to penetrate the Windows x86 market with a Windows ARM platform. Assuming Nvidia will use Windows in their future products.

Microsoft's Surface products are a joke, and the ARM versions are laughably bad. The whole purpose of these devices were to directly compete against Apple since nobody else made similar products like Apple, but there's a reason for this. People who want overpriced thin and light products will go Apple, and nothing else. The Windows market is dominated by cheap and expandable products. This is also why Chromebooks are doing so well, as the cheap and light products are what consumers want, and those consumers generally don't care what OS or CPU exists in these devices.

No we didn't. We got AMD because IBM wanted a second source to the CPU's manufactured for them. They signed a 10 year agreement that Intel never honored. IBM wanted a second source of these 8086 chips and Intel agreed but eventually fought back and lost a court case. IBM didn't want Intel to have exclusive control over x86, which is why we had AMD, Cyrix, and even IBM made their own x86 CPU's.



Neither x86 or ARM are truly open, but both have their ups and downs. x86 CPU manufacturing is closed to mostly Intel and AMD, but because of IBM's mistakes and the IBM clones, the actual platform that x86 mostly sits on which is IBM Clones is open. ARM is open to anyone to manufacture, but most ARM devices exist in a closed ecosystem that actively prevents anyone from installing any software they please. This is why Android is so segmented in OS versions because it's up to each hardware manufacturer to supply an update to these devices, and of course it would be unwise to support a product that is no longer profitable. In fact, the lack of support would be better to promote people to buy new products with the newer Android OS. Apple doesn't have these issues but as we can see, Apple updates aren't always to the benefit of their customers when they can actually slow down the device with updates. Apple updates can even break devices if you replaced the screen or other parts that aren't from Apple, thus giving Apple total control over a product you supposedly own.


If I were AMD, this is a problem I would welcome. Sucks for consumers but great for AMD's bottom line.

Everyone knew this was going to happen because it just made sense for Apple. Apple wasn't happy with Intel either, since Intel was lagging behind other manufacturers, but Apple could have gone with AMD if they didn't want to disrupt their consumers work flow. Going ARM is a great move for Apple, but a bad move if you're a consumer of Apple Mac products. It will certainly divide the customer base while giving M1 users a worse experience overall.

What ARM needs is a desktop product that works similar to x86, which nobody has or wants to make. To make an ARM motherboard with a socket that you can install in a case would standardize ARM and that's not what manufacturers want. Manufacturers wants walled gardens that they can control. Linux has no future unless Microsoft screws up, and I'm not sure when that will happen. While I use Linux daily and enjoy using it, I can't see a future without native apps being ported over to it. Trying to get Wine working is a 50/50 chance for each application.


Of course AMD isn't too worried about Apple right now. They are a relatively small company... Lisa shelved AMDs K1 ARM chips because they don't have the resources to do 20 things at once. Right now Intel dropped the ball in x86 of course AMD is focused on eating as much of that market share as they can before Intel comes back with something to halt the bleed or potentially reverse it even. AMD don't have time to update K1 right now and go after ARM.... but that is still something I think they may revisit at some point. I wouldn't be shocked to hear AMD talking about ARM again in the future. For sure if NVidias Grace makes in roads. I think its more likely AMD revives K1 into a new Epyc ARM part... not suggesting they will dump x86 of course not but I can see them doing both. That was in fact their plan before Lisa decided they had to narrow their focus for financial reasons.

As for the windows platform... I don't really think it will be that hard to penetrate the windows market with ARM. Not as hard as some of us PC geeks expect. 99% of the software every day people use (not counting games in that) already runs perfectly fine on ARM... or can easily be recompiles. I mean Microsofts own software runs on Apple M1... as does a bunch of Adobe stuff now (and the entire adobe lineup will by the time Windows ARM gets a relaunch). Hell most regular people are using mostly open source stuff these days... stuff that almost all has ARM compiles around already. The average person wanting to install their Email client, MS word, Perhaps a basic adobe software... the VLCs filezilas Mozillas Thunderbirds ect ect all have ARM compiles now and building windows ARM compiles for those that don't have is as simple as a recompile. The software world we live in today is mostly ISA agnostic. The industry has been building with cross platform tools for a decade now. Sure there are a few examples of oddities. Very little of that type of stuff will require top tier performance... it just has to run reasonably. So if Nvidia builds server class arm chips for PC use, there is a good chance they will run even those very old or very obscure bits at the same speed as all but the highest end x86 chips anyway. The truth is ARM is a much faster ISA its just the way it is... the fact that basically parts without active cooling can hang with x86 chips at all at this point should be all the proof people need. Acrons first ARM chips where a revelation.... they destroyed Intel way back in the 80s. If the technology sector was a meritocricy we would have been running ARM for 30 years already. Intel at that point was all but untouchably.

I don't disagree with you on the Qcom powered Surface devices. There biggest issue is their over priced. But like M1... Microsoft is working on a more purpose built chip. Qcom at the end of the day basically just put a phone processor in a laptop chasse. We'll see what happens with MS... they aren't going to give up on Windows ARM. I am not sure their own chip is going to save that idea or not.... it probably will come down to what Alder lake actually is. If it is a x86 that does what Intel plans... a x86 chip that can compete with M1 on power and performance. Then perhaps MS surface continue being the nothing impactful skus they are.

At the end of the day I think the future of ARM still sort of rides on Intel. I believe Alder lake is going to be a big huge belly flop that will probably be the end of Intel as the big dominant consumer chip company. At that point AMD continues stealing OEM contracts... and Laptops if not desktops start switching to ARM.
 
And the more people making them the more people licensing themselves which makes NVidia money. And the more the platform advances the more NVidia can roll those advancements into their Server and Workstation platforms. Consumer PC space makes big money on volume but NVidia wants the server, datacenter, supercomputer market. That’s where the future and the money is moving too. More competition in the consumer space will race prices down, which is great, but custom SOC’s for high end specialized tasks is a big growth industry.
Indeed I don't think Nvidia really wants ARM for anything consumer facing. Other then all the web cloud stuff people use everyday will be running they hope on Nvidia hardware not Intel or AMD. They are probably just fine with Intel and AMD continuing to sell consumer CPUs for awhile. At some point I'm sure they come for those markets... but your right the real money is powering the machines people are accessing all day. :)
 
No? Pretty sure that's my point, in that Intel mac users don't have the problems of running x86 games like the M1 users would.
It seems like you've missed the point.

Most Mac users don't buy their systems with gaming in mind — certainly not the usual Steam/EGS variety. It doesn't matter that the M1 won't run games as well as an Intel-based system if they're buying that MacBook Air mainly to write reports or attend Zoom calls. And that extends to desktops, too... I bought my iMac so I could juggle a lot of apps for work, not because I had visions of playing AC Valhalla.

Remember, most Macs sold now are already ARM-based... if gaming was actually a major concern, you'd see that reflected in sales ratios.


When I say M1, that includes future ARM based products from Apple. I don't think many people here understand how much Apple shit the bed by going ARM. Intel makes a lot of technologies that modern computers use, including Thuderbolt. Intel made it clear that they own Thunderbolt. Apple already had problems with Nvidia a long time ago, and now Nvidia is buying ARM which will impact Apple in some way. AMD is the only supplier of GPU's that Apple could use since Intel and Nvidia will obviously not be working with them, and AMD right now is too busy supplying their products to Apple's competitors. While the Apple GPU isn't bad, but it pales in comparison to what AMD and Nvidia offer.
It's too soon to say if Apple has done anything wrong. GPU performance will matter toward the higher end, but we also haven't seen how Apple tackles high-end GPUs yet. I'm not expecting it to trounce AMD or NVIDIA overnight... but I wouldn't assume it'll trail far behind, either. And if NVIDIA does succeed in buying ARM, it'll likely be subject to a lot of conditions that prevent it from claiming an unfair advantage.
 
It seems like you've missed the point.

Most Mac users don't buy their systems with gaming in mind — certainly not the usual Steam/EGS variety. It doesn't matter that the M1 won't run games as well as an Intel-based system if they're buying that MacBook Air mainly to write reports or attend Zoom calls. And that extends to desktops, too... I bought my iMac so I could juggle a lot of apps for work, not because I had visions of playing AC Valhalla.

Remember, most Macs sold now are already ARM-based... if gaming was actually a major concern, you'd see that reflected in sales ratios.



It's too soon to say if Apple has done anything wrong. GPU performance will matter toward the higher end, but we also haven't seen how Apple tackles high-end GPUs yet. I'm not expecting it to trounce AMD or NVIDIA overnight... but I wouldn't assume it'll trail far behind, either. And if NVIDIA does succeed in buying ARM, it'll likely be subject to a lot of conditions that prevent it from claiming an unfair advantage.
Apple has some strong points in the video department because the Metal API is very well documented and pretty good. The apple arcade is slowly growing and I really do hope to see some larger titles in there sooner than not.
 
I am sure Nvidia is rooting for Apple hard. If M1... M2 ect win market share.
Nvidia is but Nvidia still hates Apple because Apple has been working against Nvidia for over 10 years now. Great that through Apple ARM has a chance to enter the PC market, but also not through Nvidia.
The PC industry will be clamoring for ARM chips. Who better to provide them then the company that actually owns ARM.
Nobody is clamoring for ARM chips, at least not yet. Good chance that later on Nvidia will cock block Apple through some method. It's not hard to see Nvidia with ARM producing some new technology that only through Nvidia could anyone acquire.
Owning ARM is just about mindshare.
No it isn't. If I asked you what SoC you have in your phone, good chance you're Googling that info. You're not even sure if it's a Qualcomm, Exynos, or an Apple based SoC. If I asked you what SoC you have in your Apple product you may know. You probably know what's in your PC. By no means do you associate it with ARM as ARM isn't exactly an advertised brand.

This is why I say that consumers don't know and don't care. You could put anything in their devices and as long as it isn't slow they don't mindshare. This is where being able to build your own ARM system would create mindshare, but as it is nobody knows what's in most consumer electronics. They might know Nvidia's ARM devices, but that's because of Nvidia's mindshare on x86 systems.
Seeing them come now at the server market owning ARM makes tons of sense. When the bean counters are signing multi million dollar contracts for big iron, it will help if the company they are dealing with is seen as ARM itself. No worries about buying a dead end ARM server chip that will loose support. There where big players that tested the waters on earlier ARM server plays and basically got burnt. Owning ARM makes that a easier sell for Nvidia.
Nvidia knows that at some point the future is APU's or graphics embedded with CPU's. Since Nvidia has no x86 license then ARM makes sense to pursue.

Of course AMD isn't too worried about Apple right now. They are a relatively small company... Lisa shelved AMDs K1 ARM chips because they don't have the resources to do 20 things at once. Right now Intel dropped the ball in x86 of course AMD is focused on eating as much of that market share as they can before Intel comes back with something to halt the bleed or potentially reverse it even. AMD don't have time to update K1 right now and go after ARM.... but that is still something I think they may revisit at some point. I wouldn't be shocked to hear AMD talking about ARM again in the future. For sure if NVidias Grace makes in roads. I think its more likely AMD revives K1 into a new Epyc ARM part... not suggesting they will dump x86 of course not but I can see them doing both. That was in fact their plan before Lisa decided they had to narrow their focus for financial reasons.
I think AMD is ignoring ARM because of the runaway success of their Ryzen CPU's. ARM was a backup plan if they fail to produce a competitive x86 CPU. I wouldn't be shocked if Qualcomm is working with AMD to produce SoC's because of AMD's graphics. There's already talk of AMD doing what Apple did and move the memory closer to the CPU to reduce latency and increase IPC. This of course sucks since you can't upgrade ram but it does give a performance edge, and does lower power consumption.
As for the windows platform... I don't really think it will be that hard to penetrate the windows market with ARM. Not as hard as some of us PC geeks expect. 99% of the software every day people use (not counting games in that) already runs perfectly fine on ARM... or can easily be recompiles. I mean Microsofts own software runs on Apple M1... as does a bunch of Adobe stuff now (and the entire adobe lineup will by the time Windows ARM gets a relaunch). Hell most regular people are using mostly open source stuff these days... stuff that almost all has ARM compiles around already. The average person wanting to install their Email client, MS word, Perhaps a basic adobe software... the VLCs filezilas Mozillas Thunderbirds ect ect all have ARM compiles now and building windows ARM compiles for those that don't have is as simple as a recompile.
Yes and as a Linux user I do find myself not needing applications that are closed source much. The problem with the M1 is Apple themselves as not all developers want to spend the time to port their applications to a platform that doesn't conform to standards. I know the Dolphin emulator author is working on a M1 port but it'll be using OpenGL and MoltanVk. RCPS3 doesn't even have a Mac port due to the Metal API having limitations and MoltanVK just not good enough for performance. Developers solution for Apple users who wants to run RPCS3 is to install Windows or Linux. Yuzu got started to port to the Apple M1 but as always the problem seems to be MoltanVK. Citra has been ported but it isn't working too well, probably due to OpenGL being the only API it supports and Apple's OpenGL is left for dead.

A common theme you'll see here is that many developers do not like working with the Metal API, and MoltanVK has issues. This is because nobody wants to learn an API that only works on a very small group of users. Also the M1 is so new that there's bound to be issues. These are issues that might be ironed out with time but that's something you have to deal with.
I don't disagree with you on the Qcom powered Surface devices. There biggest issue is their over priced. But like M1... Microsoft is working on a more purpose built chip. Qcom at the end of the day basically just put a phone processor in a laptop chasse. We'll see what happens with MS... they aren't going to give up on Windows ARM. I am not sure their own chip is going to save that idea or not.... it probably will come down to what Alder lake actually is. If it is a x86 that does what Intel plans... a x86 chip that can compete with M1 on power and performance. Then perhaps MS surface continue being the nothing impactful skus they are.
The problem with Windows ARM was that Microsoft forced users to get apps from the Microsoft Store, which didn't sit well for a lot of people. They also had like no x86 backwards compatibility. They fixed these issues now, but x86 performance is terrible. Microsoft was too concerned about pushing people into a walled garden that they forgot they needed people in the garden first before you could lock them in.
At the end of the day I think the future of ARM still sort of rides on Intel. I believe Alder lake is going to be a big huge belly flop that will probably be the end of Intel as the big dominant consumer chip company. At that point AMD continues stealing OEM contracts... and Laptops if not desktops start switching to ARM.
I don't expect anything good from Intel in 2021. Alder lake is just a small stop in Intel pretending to stay relevant. 2022 is when Intel really starts to get their shit together. Intel has been riding on their Tic Toc since Sandy Bridge for a long time and I expect Intel to have some new massive performing CPU with Xe graphics sometime next year. It won't be on 14nm or 10nm as these technologies are too far behind compared to Apple's 5nm and AMD's 7nm. I expect Intel to partner with someone to manufacture their new chips, while massively overhauling their fabs to move to a more aggressive manufacturing process.

The only way I could see Intel failing is if they pretend that Alder lake is the holy grail for 2+ years and they think 10nm ought to be enough for anyone. They'd have to be real fucking stupid to do something like that.
 
I don't expect anything good from Intel in 2021. Alder lake is just a small stop in Intel pretending to stay relevant. 2022 is when Intel really starts to get their shit together. Intel has been riding on their Tic Toc since Sandy Bridge for a long time and I expect Intel to have some new massive performing CPU with Xe graphics sometime next year. It won't be on 14nm or 10nm as these technologies are too far behind compared to Apple's 5nm and AMD's 7nm. I expect Intel to partner with someone to manufacture their new chips, while massively overhauling their fabs to move to a more aggressive manufacturing process
I'm thinking late 2022 and early 2023, their DDR5 stuff will probably shape up pretty well (not sure about first-gen) Intel has too many resources to remain where they are but holy crap baskets batman did they ever manage to screw this one up. Intel has already signed agreements with TSMC for 5nm and 3nm for what is rumored to be i3's and some i9's or maybe Xeon W's. But TSMC did confirm Intel's commitment to purchasing on their 5nm node and will be getting in there before AMD does.
 
The only way I could see Intel failing is if they pretend that Alder lake is the holy grail for 2+ years and they think 10nm ought to be enough for anyone. They'd have to be real fucking stupid to do something like that.
To be fair - and I have no actual idea as to which way things will go - we also thought they’d be really stupid to keep trying to refine 14nm and refreshing skylake for so long and that they’d have some kind of generational leap coming in 2019, after Ryzen had been out for a bit.

and now their marketing (and various tech sites that buy into it) are talking about how Intel’s 10nm is plenty good enough compared to TSMC’s 7nm... My best guess is that Windows/Linux systems wind up with a mishmash of environments for different use cases and Apple’s big advantage will come from their consistency.

whether that turns into a market share advantage or whatever for ARM, who knows.
 
AMD has had TSMC on the plate for Zen4/5nm a long time before talk of intel and TSMC was even around. Like 2019.
Maybe but Intel has already spent the money so they are in there, and their chips are expected on market before AMD's, Intel is already using TSMC's 5nm to pick up the slack from their 7nm fabs on their DoE commissioned supercomputers, so they just rolled that forward for some consumer parts and agreed for 3nm in H2 of 2022.
 
I think AMD is ignoring ARM because of the runaway success of their Ryzen CPU's. ARM was a backup plan if they fail to produce a competitive x86 CPU.

This is a rewrite of the history. Check the talks, the roadmaps, and the official plans when Rory Read was CEO. Not only the x86 core was unnamed then, but the ARM core was officially called K12, because it was the successor to K5, K6, K7, K8, K9, and K10.
 
“Instead, Intel, as prime contractor, moved on to a new version of Aurora (A21), with full exascale (a billion billion operations per second) compute power combining Intel 10nm CPUs and “Ponte Vecchio” 7nm GPUs, along with the Cray Shasta architecture for 2021, a centerpiece of the U.S. exascale strategy.”

https://insidehpc.com/2020/07/intel-delays-7nm-chip-again-what-does-it-mean-for-aurora-exascale/

Intel is using a combination of their 7nm and TSMC’s 5nm for the GPU’s not the CPU’s.

https://www.anandtech.com/show/1645...hio-xehpc-power-on-posts-photo-of-server-chip

I am trying to find the news release where Intel announced they were using TSMC’s 5nm process to cover the shortfall of their 7nm after the DoJ threatened a lawsuit over yet another delay on Aurora from the issues on their 7nm fab.
Edit:
Found this one but I’m trying to find the original not an article about an article.
https://wccftech.com/exclusive-intel-ponte-vecchio-gpu-not-on-tsmc-6nm/
 
Last edited:
  • Like
Reactions: Mega6
like this
I'm thinking late 2022 and early 2023, their DDR5 stuff will probably shape up pretty well (not sure about first-gen) Intel has too many resources to remain where they are but holy crap baskets batman did they ever manage to screw this one up. Intel has already signed agreements with TSMC for 5nm and 3nm for what is rumored to be i3's and some i9's or maybe Xeon W's. But TSMC did confirm Intel's commitment to purchasing on their 5nm node and will be getting in there before AMD does.
Probably but later this year AMD is going to release Warhol which is going to be 6nm and be a slight performance improvement over Zen3. That's going to happen later this year, and likely be before Intel has anything with 5nm and 3mn. AMD's Zen4 is looking to be really impressive with it being on 5nm and having a massive performance boost compared to Zen3. If you needed a reason not to spend money on overpriced hardware this year, that's a good enough reason for me.
 
Probably but later this year AMD is going to release Warhol which is going to be 6nm and be a slight performance improvement over Zen3. That's going to happen later this year, and likely be before Intel has anything with 5nm and 3mn. AMD's Zen4 is looking to be really impressive with it being on 5nm and having a massive performance boost compared to Zen3. If you needed a reason not to spend money on overpriced hardware this year, that's a good enough reason for me.
Turns out Intel is already shipping their Ponte Vechio to customers and that is a combination of Intel 7nm and TSMC 5nm depending on the sku ordered.
But yeah not being to get the hardware I thought I needed has turned out to be a blessing in disguise.
 
dont forget the original Xbox with its x86 processor and the Xbox 360 with a triple core PowerPC proc!

also, though it’s a bit more of a tangent, I believe most of the portables have all been arm based, but none arm64 until the switch?

Bring back Hitachi CPUs!
 
Perhaps this is a bit off-topic, but I figured it would be good example worth mentioning:




The M1 absolutely smokes the similarly spec'd AMD and Intel Surfaces in both synthetic and real-world applications, and in Rosetta 2 as well.
ARM definitely has a future, and there is a reason NVIDIA is going after it.

As for x86-64, it certainly has an uphill battle, and we have still yet to see a true desktop-class, or at least a higher TDP, variant of the M1.
 
Back
Top