Bloomberg - Microsoft Is Designing Its Own Chips for Servers, Surface PCs

Ya people saying ARM will never replace x86 are in for a shock. Microsoft will probably also be willing to hook OEM windows manufacturers up to ensure Windows ARM becomes the standard.
 
None of this matters unless Microsoft commits and actually can stick to a direction. A problem they've had for some time.

Windows Phone is a good example of this. They built an entire platform for 7/8 Phone and then later Windows 10 phone. Barely supported it, then killed it. And that's after their previous gen before that. They didn't bother to support it and apparently didn't realize going in that it was going to take a long time for people to adopt it and get apps.

The only way forward is with a lot of resources, very well designed and engineered chips, and someone at the top that is making decisions that aren't made by a comittee but with a very specific purpose in order to ensure that both their software and hardware teams are working in tandem and towards the same goals. These are traditionally things that Microsoft doesn't do well. And I'm reminded of that every time I boot into Windows 10 to play CP2077.
 
Last edited:
Ya people saying ARM will never replace x86 are in for a shock. Microsoft will probably also be willing to hook OEM windows manufacturers up to ensure Windows ARM becomes the standard.
lol guess we'll see, maybe, since the "articles" are simply someones ramblings and rumors.
 
  • Like
Reactions: erek
like this
None of this matters unless Microsoft commits and actually can stick to a direction. A problem they've had for some time.

Windows Phone is a good example of this. They built an entire platform for 7/8 Phone and then later Windows 10 phone. Barely supported it, then killed it. And that's after their previous gen before that. They didn't bother to support it and apparently didn't realize going in that it was going to take a long time for people to adopt it and get apps.
The only way forward is with a lot of resources, very well designed and engineered chips, and someone at the top that is making decisions that aren't made by a comittee but with a very specific purpose in order to ensure that both their software and hardware teams are working in tandem and towards the same goals. These are traditionally things that Microsoft doesn't do well. And I'm reminded of that every time I boot into Windows 10 to play CP2077.
If they go after their cloud server chip supply... hopefully that means they are going to be going hard on performance. The consumer side is a big question mark of course.... I can see MS going after Intel (x86) replacement. With Apple doing what they are doing... and likely picking up a good bit of market share. MS has no real worries in regards to any anti trust stuff. If MS had tried to make their own CPUs 20 years ago they would have ended up in court and risked being broken up. Now.... especially if Apple manages to show 20% market share by the time MS gets this ready for prime time. Clear sailing. Pretty sure MS has the legal go ahead to compete with Intel just like AMD does as a fabless CPU manufacturer.

See your point though MS has a long history of half assing things and killing them early. The next few years are going to get messy and wonderful. lol
 
None of this matters unless Microsoft commits and actually can stick to a direction. A problem they've had for some time.

Windows Phone is a good example of this. They built an entire platform for 7/8 Phone and then later Windows 10 phone. Barely supported it, then killed it. And that's after their previous gen before that. They didn't bother to support it and apparently didn't realize going in that it was going to take a long time for people to adopt it and get apps.

The only way forward is with a lot of resources, very well designed and engineered chips, and someone at the top that is making decisions that aren't made by a comittee but with a very specific purpose in order to ensure that both their software and hardware teams are working in tandem and towards the same goals. These are traditionally things that Microsoft doesn't do well. And I'm reminded of that every time I boot into Windows 10 to play CP2077.

I think this go around MSFT won't fall into the same mistakes. The move to its own processor is a page taken straight from Amazon AWS playbook. However, with the Graviton, I can only run Linux, whereas with MSFT's own ARM processor, it could run Windows in the future, so Enterprise that want to continue to run Windows Server in Azure will be able to do so at a reduced cost. Graviton is a bit weak for now, but for something that we need to test quickly, it is quite cheap. AWS Graviton utilization has increased year over year.
https://twitter.com/LucidCap/status/1338208878278926338

AMD EC2 instances are about 10% cheaper than Intel. Intel is not going to do well in the near future, or maybe even further down the road with cloud being all the rage nowadays.
 
None of this matters unless Microsoft commits and actually can stick to a direction. A problem they've had for some time.

Windows Phone is a good example of this. They built an entire platform for 7/8 Phone and then later Windows 10 phone. Barely supported it, then killed it. And that's after their previous gen before that. They didn't bother to support it and apparently didn't realize going in that it was going to take a long time for people to adopt it and get apps.

The only way forward is with a lot of resources, very well designed and engineered chips, and someone at the top that is making decisions that aren't made by a comittee but with a very specific purpose in order to ensure that both their software and hardware teams are working in tandem and towards the same goals. These are traditionally things that Microsoft doesn't do well. And I'm reminded of that every time I boot into Windows 10 to play CP2077.
I bought into Windows phone. Lost much respect for how quickly they dropped it. I wish people would give me millions of dollars to make dumbass arbitrary decisions.
 
Fuck Microshit chips. This is just another step towards totally eroding general purpose and 100% owned hardware.
People are freaking out about the "Great Reset" over in soapbox, well this is it: a future where you own nothing, where everything is leased, where a corporation can remotely shut off or disable your CPU because reasons.

Again, fuck this shit. Stallman was right.
 
Fuck Microshit chips. This is just another step towards totally eroding general purpose and 100% owned hardware.
People are freaking out about the "Great Reset" over in soapbox, well this is it: a future where you own nothing, where everything is leased, where a corporation can remotely shut off or disable your CPU because reasons.

Again, fuck this shit. Stallman was right.
A dark cyberpunk future, indeed... :borg:
 
If they go after their cloud server chip supply... hopefully that means they are going to be going hard on performance. The consumer side is a big question mark of course.... I can see MS going after Intel (x86) replacement. With Apple doing what they are doing... and likely picking up a good bit of market share. MS has no real worries in regards to any anti trust stuff. If MS had tried to make their own CPUs 20 years ago they would have ended up in court and risked being broken up. Now.... especially if Apple manages to show 20% market share by the time MS gets this ready for prime time. Clear sailing. Pretty sure MS has the legal go ahead to compete with Intel just like AMD does as a fabless CPU manufacturer.

See your point though MS has a long history of half assing things and killing them early. The next few years are going to get messy and wonderful. lol
I would certainly seem like every megacorp is done with Intel's bumbling and repeated failures at this point, and really, who could blame them.
It's too bad that AMD is having capacity issues (also good that they are at full capacity, though - very telling), because they are the only company left that is pushing x86-64 forward, and in far better ways than Intel ever has or could; mad props to Lisa Su and all those working with and under her.

ARM is going to be the future, though this isn't quite the general-purpose future we were hoping for.
I get the feeling that 20-30 years from now we are going to be pushing RISC-V to crush ARM, the same way we are hoping ARM crushes x86-64 right now.

What I find very ironic, is that in the 1990s, the pendulum was swinging towards x86 systems that could run and do everything, and pushed away from the processor/ISA specific proprietary systems of the 1980s and earlier.
The pendulum is now swinging back the other way into that proprietary area, where even if the ISA is the same, each individual company's systems will only run specific operating systems and their respective software for said systems.

History repeats itself.
Up next, the world moves back to dumb terminals and time-sharing on mainframes, plus the return of the minicomputer! (links provided for those not over the age of 70) :D
 
Fuck Microshit chips. This is just another step towards totally eroding general purpose and 100% owned hardware.
People are freaking out about the "Great Reset" over in soapbox, well this is it: a future where you own nothing, where everything is leased, where a corporation can remotely shut off or disable your CPU because reasons.

Again, fuck this shit. Stallman was right.
Yep. They are training the masses with all these subscription services. The end game is that everyone has a dumb terminate they use to login to the cloud. You only get access to what ever you pay for. Do you want internet? $4.99 month please. You want to use so and so software? $4.99 a month please. You want to play the latest AAA game? $100 for the game plus $4.99 to $99.99 based on the performance level you want. That is the future.
 
Microsoft has been using ARM servers internally for a while. Azure runs on TX2 chips.
 
Up next, the world moves back to dumb terminals and time-sharing on mainframes, plus the return of the minicomputer! (links provided for those not over the age of 70) :D
Minicomputers?
MMMM, Data General Nova:
Z2ssMfm.jpg
I ripped off their color scheme for my home lab's rack.
More fun than the not-quite-black gray of modern computers, or the plain beige of the 90's.
 
History repeats itself.
Up next, the world moves back to dumb terminals and time-sharing on mainframes, plus the return of the minicomputer! (links provided for those not over the age of 70) :D
Really between virtualization, web apps, Citrix, and RDP thin clients we arrived there years ago. It is far easier to secure a single server and have it accessed only through secured protocols while the user side has no access directly to sensitive data. Set them to wipe themselves on log off or reboot and give the local user 0 administrative functions and you have greatly decreased your attack surface. I have a sneaking suspicion that secure infrastructure is going to be a much more important and front facing issue in 2021.
 
Frankly, Microsoft doesn't have much choice here.

Windows on ARM has struggled in no small part because Microsoft is chained to Qualcomm, and Qualcomm's PC-oriented chips are... underpowered, to put it mildly. There are other reasons, of course, but Microsoft was never going to get PC users to adopt ARM if x86 chips were always faster.

And that's a real problem when Apple's in-house silicon is not only much faster than Qualcomm's chips, but outperforms comparable Intel CPUs (and in some cases AMD's latest). Imagine if Microsoft did nothing — there's a real chance Apple could walk away with certain categories thanks to clear advantages in performance and battery life. I picture Microsoft execs having nightmares where reviewers say "if you want the most powerful computer in this class, you have to get a Mac."

The problem, as others have noted, is that this might represent Microsoft's classic too-little-too-late response à la Windows Phone or Zune. A reluctant acknowledgment that rivals are doing something better when it should have anticipated possible issues years ago. I don't think the Mac is about to take over the PC world — not when the company's cheapest system is $699 — but there is a chance the industry could look very different in the long run.
 
Frankly, Microsoft doesn't have much choice here.

Windows on ARM has struggled in no small part because Microsoft is chained to Qualcomm, and Qualcomm's PC-oriented chips are... underpowered, to put it mildly. There are other reasons, of course, but Microsoft was never going to get PC users to adopt ARM if x86 chips were always faster.

And that's a real problem when Apple's in-house silicon is not only much faster than Qualcomm's chips, but outperforms comparable Intel CPUs (and in some cases AMD's latest). Imagine if Microsoft did nothing — there's a real chance Apple could walk away with certain categories thanks to clear advantages in performance and battery life. I picture Microsoft execs having nightmares where reviewers say "if you want the most powerful computer in this class, you have to get a Mac."

The problem, as others have noted, is that this might represent Microsoft's classic too-little-too-late response à la Windows Phone or Zune. A reluctant acknowledgment that rivals are doing something better when it should have anticipated possible issues years ago. I don't think the Mac is about to take over the PC world — not when the company's cheapest system is $699 — but there is a chance the industry could look very different in the long run.
For our 2021 hardware refresh for the first time in a decade Apple is back on the table for fleet wide. We are now down to one piece of software that isn’t Mac native or already a web app but now works flawlessly in Citrix. For the 8 users who need it getting this spun up is a no brainer.
Now that our firewall hardening process is complete we have 12 months of logged network traffic usage and outside the Ubuntu machines Apple is the platform that is not only generating the least amount of non user initiated traffic but also coming in with the lowest TCO. It certainly doesn’t hurt that their machines are perceived as being nicer by staff which really need the moral boost.
Microsoft needs to show my bosses something pretty snazzy if they want us to be using their OS much longer, the licensing and security costs on them seem to go up faster than not. We’re going to keep on with Azure, O365, and our SharePoint Online services of course as the Microsoft Canada data centres are the only ones FOIPPA compliant but Intel isn’t doing anybody any favours and AMD can’t meet supply demands.
 
Yep. They are training the masses with all these subscription services. The end game is that everyone has a dumb terminate they use to login to the cloud. You only get access to what ever you pay for. Do you want internet? $4.99 month please. You want to use so and so software? $4.99 a month please. You want to play the latest AAA game? $100 for the game plus $4.99 to $99.99 based on the performance level you want. That is the future.
Yes but in many cases it ends up being cheaper than flat out buying the software more than not if you were one of the users who kept updated. But gone are the days of buying a single copy and using it for a decade and change.
 
For our 2021 hardware refresh for the first time in a decade Apple is back on the table for fleet wide. We are now down to one piece of software that isn’t Mac native or already a web app but now works flawlessly in Citrix. For the 8 users who need it getting this spun up is a no brainer.
Now that our firewall hardening process is complete we have 12 months of logged network traffic usage and outside the Ubuntu machines Apple is the platform that is not only generating the least amount of non user initiated traffic but also coming in with the lowest TCO. It certainly doesn’t hurt that their machines are perceived as being nicer by staff which really need the moral boost.
Microsoft needs to show my bosses something pretty snazzy if they want us to be using their OS much longer, the licensing and security costs on them seem to go up faster than not. We’re going to keep on with Azure, O365, and our SharePoint Online services of course as the Microsoft Canada data centres are the only ones FOIPPA compliant but Intel isn’t doing anybody any favours and AMD can’t meet supply demands.

Interesting! I'm sure your employer is in a better position for this than others (not every company can justify Macs due to price or software), but this is how Apple makes its inroads. Companies that can justify Macs ask why they should buy Windows PCs that are slower or more troublesome, and Microsoft's user base erodes.

I do realize Microsoft won't entirely mind given its focus on the cloud, and you even hinted as much — why worry about someone buying Windows licenses when they've signed up for much more lucrative cloud services like Azure and Office 365? It's just wild to think Mac deployments are making much more sense these days.
 
Interesting! I'm sure your employer is in a better position for this than others (not every company can justify Macs due to price or software), but this is how Apple makes its inroads. Companies that can justify Macs ask why they should buy Windows PCs that are slower or more troublesome, and Microsoft's user base erodes.

I do realize Microsoft won't entirely mind given its focus on the cloud, and you even hinted as much — why worry about someone buying Windows licenses when they've signed up for much more lucrative cloud services like Azure and Office 365? It's just wild to think Mac deployments are making much more sense these days.
That's the rub, over 5 years a $600 laptop costs significantly more than a $1,100 mac. By the time I have Win 10 Enterprise, my AV license, MDM license, that $600 dell is now closer to $800. But now work in the management costs associated with most win10 devices, update and patch management mostly and the annual costs of said MDM and AV and software alone brings them to parity on costs. But over that 5-year span, I will not see a Mac user, I will not deal with how an intel update broke the wireless, I won't be dealing with a broken Ricoh driver update killed printing to tray 3 again, and so on and so on. And the big one right now is Mac's screens look great in crap fluorescent lighting where I have a crapload of users complaining about eye strain and headaches from their screens. They get them at a far higher rate than the Apple users, the HR issues alone would make most happy if we could bring that down.
 
In a world where I can run a whole company with just my phone... what the fuck is MS doing with their time?
 
In a world where I can run a whole company with just my phone... what the fuck is MS doing with their time?

I wish we could run my whole company with my phone. On the flip side though, I've managed to get it to the point where we barely need Microsoft. Just a few old applications left - thankfully they run just fine in VMs.

Windows is - by far - the biggest IT issue at our company. Linux users are smart enough to fix issues themselves, and MacOS just works.

Windows users: "This printer sucks! It never works right!"
Me: You said that about the last three printers too
Linux/Mac user: just slack me the file, I'll print it for you

And it always prints perfectly. You can pretty much swap any task in for printing and the same principle applies.
 
Last edited:
In a world where I can run a whole company with just my phone... what the fuck is MS doing with their time?
Ensuring that folks can continue to run their apps on everything from phones to laptops, and reducing the dependence on a specific architecture.

Yes, it's weird that a "software company" would make hardware, but it actually aligns more with their goals of getting their stuff everywhere, seamlessly.
 
And the big one right now is Mac's screens look great in crap fluorescent lighting where I have a crapload of users complaining about eye strain and headaches from their screens.
I suppose that management isn't considering upgrading the lighting, either.
 
I suppose that management isn't considering upgrading the lighting, either.
We have we’re going all LED, but it’s going to take like 5 years to bang out all sites. Replacing that many ballasts is expensive and the disposal of that many bulbs and we have so many spares they are trying to schedule upgrades as the supply dwindles. So when it’s all done we don’t have too many extras that are going with them. But by the time they are all done we will be looking at putting out bids for our next refresh in 2026 I suspect screen tech and our operating environment will be drastically different by then so it likely will no longer be an issue.
But before any of that starts we need to finish the solar panel installs, maintenance is shutting the buildings down on the 23, 24’th so they can get the capacitor banks in for smoothing the panel output. BC hydro charges huge sums for dumping unregulated power back onto the grid.

EDIT:
They requested to do it on those dates, they were not asked too. Most of the team is single, and to compensate for it 2.5 time was authorized they just don’t know that yet.
 
Last edited:
Frankly, Microsoft doesn't have much choice here.

Windows on ARM has struggled in no small part because Microsoft is chained to Qualcomm, and Qualcomm's PC-oriented chips are... underpowered, to put it mildly. There are other reasons, of course, but Microsoft was never going to get PC users to adopt ARM if x86 chips were always faster.

And that's a real problem when Apple's in-house silicon is not only much faster than Qualcomm's chips, but outperforms comparable Intel CPUs (and in some cases AMD's latest). Imagine if Microsoft did nothing — there's a real chance Apple could walk away with certain categories thanks to clear advantages in performance and battery life. I picture Microsoft execs having nightmares where reviewers say "if you want the most powerful computer in this class, you have to get a Mac."

The problem, as others have noted, is that this might represent Microsoft's classic too-little-too-late response à la Windows Phone or Zune. A reluctant acknowledgment that rivals are doing something better when it should have anticipated possible issues years ago. I don't think the Mac is about to take over the PC world — not when the company's cheapest system is $699 — but there is a chance the industry could look very different in the long run.
I sold computers for a living 20 some years back. $699 for a computer would have been basically impossible. Everything is relative. :)

But yes Apple isn't going after the ele cheapo..... However the ele cheapo category is still not a Intel/AMD lock.

As I see it for anyone in the market for a sub $500 machine.... a chromebook is probably the best option. With a high to medium range smartphone being the second best option. The shit box emachine type desktop is a distant who cares... and a cheap windows laptop is an option today but really every year the cheap windows laptops look crappier and crappier vs the chromebook options.

Once you hit the mid range price range right now... hell ya Apples M1 powered AIR and pro laptops are probably the best DEAL in the market. Best battery life, one of the best bang for the bucks on the screen.... and a OS that makes very good use of a small amount of RAM. The M1 powered stuff is frankly probably best in class bang for the buck. Which seems crazy but also seems pretty much the way it is.

If Apple releases a M1+ or M2 to power higher end notebooks and mid range desktop imac options. Then we end up at a place where imo it gets really hard to recommend windows for much other then gaming. Even the arguments of X or Y piece of windows software requires me to run windows is going to evaporate over the next year or two. Now that Apple has both exciting hardware, and most importantly market share (its going up and M1... and its follow up big brothers are going to drive Apples share up faster then MS is comfortable with for sure) software developers that may not have been supporting MacOS are going to. (and imo everything major is already on MacOS anyway)

It does feel like MS may be too late.... I think if they where smart. They would mend their fences with Nvidia. Ya I hate Nvidia as much as the next Nvidia hater... but if they are going to end up being ARM. It is probably in Microsofts interest to get them to design a proper Apple chip competitor that can excite consumers. If the dangle their cloud service contract in front of Nvidia they may even get a decent deal. Nvidia could easily build ARM/Ampere chips for MS cloud... and a ARM/Ampere/AI SOC for their consumer devices. Who else can bring a windows friendly version of the Apple chips other then them. Perhaps AMD.... but AMD if where being honest has little incentive to kill x86 now that they are the champs for today and probably the next few years. Nvidia is motivated to become the Windows ARM consumer and server supplier.
 
I sold computers for a living 20 some years back. $699 for a computer would have been basically impossible. Everything is relative. :)

But yes Apple isn't going after the ele cheapo..... However the ele cheapo category is still not a Intel/AMD lock.

As I see it for anyone in the market for a sub $500 machine.... a chromebook is probably the best option. With a high to medium range smartphone being the second best option. The shit box emachine type desktop is a distant who cares... and a cheap windows laptop is an option today but really every year the cheap windows laptops look crappier and crappier vs the chromebook options.

Once you hit the mid range price range right now... hell ya Apples M1 powered AIR and pro laptops are probably the best DEAL in the market. Best battery life, one of the best bang for the bucks on the screen.... and a OS that makes very good use of a small amount of RAM. The M1 powered stuff is frankly probably best in class bang for the buck. Which seems crazy but also seems pretty much the way it is.

If Apple releases a M1+ or M2 to power higher end notebooks and mid range desktop imac options. Then we end up at a place where imo it gets really hard to recommend windows for much other then gaming. Even the arguments of X or Y piece of windows software requires me to run windows is going to evaporate over the next year or two. Now that Apple has both exciting hardware, and most importantly market share (its going up and M1... and its follow up big brothers are going to drive Apples share up faster then MS is comfortable with for sure) software developers that may not have been supporting MacOS are going to. (and imo everything major is already on MacOS anyway)

It does feel like MS may be too late.... I think if they where smart. They would mend their fences with Nvidia. Ya I hate Nvidia as much as the next Nvidia hater... but if they are going to end up being ARM. It is probably in Microsofts interest to get them to design a proper Apple chip competitor that can excite consumers. If the dangle their cloud service contract in front of Nvidia they may even get a decent deal. Nvidia could easily build ARM/Ampere chips for MS cloud... and a ARM/Ampere/AI SOC for their consumer devices. Who else can bring a windows friendly version of the Apple chips other then them. Perhaps AMD.... but AMD if where being honest has little incentive to kill x86 now that they are the champs for today and probably the next few years. Nvidia is motivated to become the Windows ARM consumer and server supplier.
This is in CDN funds but any laptop $600 or less is basically a Chromebook with windows and maybe an i3 not a Celeron. $600-900 and you are into your base line machines where minor features are going to be different but they are all basically the same. $900+ is where things start to get interesting and the special features can start to show you see nicer screens, GPU’s battery life they are going to start diverging and actually offering something different. The Apple 13” M1 based MacBooks don’t really have an equal here at that price point. There are things that offers different features, maybe more Ram and CPU but a cheaper screen, large HDD and GPU but shite battery and variations there but you get into a lot of subjectives into this range and people can start picking what they want and need and can start shopping towards priorities rather than generalities. Apple has thrown a serious punch and Microsoft has taken notice, Dell, HP, and Lenovo are all scrambling I will not be surprised if come May, Intel announces steep discounts on their lower power i3 and i5 variants to allow them to offer something that can be competitive in the price/aesthetics market of the 13” ultrabook.
 
  • Like
Reactions: ChadD
like this
This is in CDN funds but any laptop $600 or less is basically a Chromebook with windows and maybe an i3 not a Celeron. $600-900 and you are into your base line machines where minor features are going to be different but they are all basically the same. $900+ is where things start to get interesting and the special features can start to show you see nicer screens, GPU’s battery life they are going to start diverging and actually offering something different. The Apple 13” M1 based MacBooks don’t really have an equal here at that price point. There are things that offers different features, maybe more Ram and CPU but a cheaper screen, large HDD and GPU but shite battery and variations there but you get into a lot of subjectives into this range and people can start picking what they want and need and can start shopping towards priorities rather than generalities. Apple has thrown a serious punch and Microsoft has taken notice, Dell, HP, and Lenovo are all scrambling I will not be surprised if come May, Intel announces steep discounts on their lower power i3 and i5 variants to allow them to offer something that can be competitive in the price/aesthetics market of the 13” ultrabook.
Ya the Dell 13 inchers have been very popular. I also love that you can get them with Ubuntu out of the box. However Apple is basically destroying them in value right now.... value, Apple. Implausible but true. lol

The issue for the windows oems... is ya they need someone to produce Apple level hardware, and they also need Microsoft to build a non shit version of Windows ARM. Its a tall order. We'll have to see if MS is up to the challenge. Why I think they should make nice with Nvidia. Set them the goal of windows M1 chip.... make it clear if they nail it in a year they get the MS cloud contracts. Then MS needs to focus on what they are good at (or should be) making a worthy version of ARM windows. Probably the only way they get something out to compete for real with Apple before Apple moves on to the M1 replacement. Problem I see for MS doing their own chip... by the time they get something out the door Apple will be ready to replace their M1 with a newer more powerful version if need be. I know Apple hasn't been to fast to upgrade their hardware in the past... however now that they build it all, I could see them being able to iterate more often. I wouldn't be shocked to see them replacing the Mini/Air/Pro as early as next year with a M1+. Perhaps not a complete redesign, but as the fab end gets better why not slide a couple extra cores in or bump the freq.
 
I bought into Windows phone. Lost much respect for how quickly they dropped it. I wish people would give me millions of dollars to make dumbass arbitrary decisions.

I did as well. The UX and ecosystem was years ahead of its time on release, certainly far ahead of android. The music and app services were second to none. The death blows were lack of support and hardware (although there were a few winners: the Dell Venue Pro was a dream).
 
I did as well. The UX and ecosystem was years ahead of its time on release, certainly far ahead of android. The music and app services were second to none. The death blows were lack of support and hardware (although there were a few winners: the Dell Venue Pro was a dream).
still using an acer jade primo with 10 on it. was sad they killed it(win phn) two weeks after i bought it....
 
None of this matters unless Microsoft commits and actually can stick to a direction. A problem they've had for some time.

Windows Phone is a good example of this. They built an entire platform for 7/8 Phone and then later Windows 10 phone. Barely supported it, then killed it. And that's after their previous gen before that. They didn't bother to support it and apparently didn't realize going in that it was going to take a long time for people to adopt it and get apps.

The only way forward is with a lot of resources, very well designed and engineered chips, and someone at the top that is making decisions that aren't made by a comittee but with a very specific purpose in order to ensure that both their software and hardware teams are working in tandem and towards the same goals. These are traditionally things that Microsoft doesn't do well. And I'm reminded of that every time I boot into Windows 10 to play CP2077.

Exactly this! I do not see that even if Microsoft goes all in on these new chips and stuff that will actually mean anything. Off the top of my head, Microsoft Surface, Microsoft Band Smartwatch, Kinect 2 and also the Windows Phone / Mobile platform are all dead because of Microsoft letting them die, simple as that. (These were all really good products.) Oh, and Cortana when it first came out was good as well but they have let it die as well.

Therefore, if Microsoft were to abandon X86 and go full on ARM, I could not see the Windows OS surviving. I think they would not know how to follow through with an ARM processor of their own and if they have nothing to fall back on, that is when Linux might grab a large part of the desktop OS market.
 
Ya people saying ARM will never replace x86 are in for a shock. Microsoft will probably also be willing to hook OEM windows manufacturers up to ensure Windows ARM becomes the standard.
If there is a good x86-64 emulator and the performance is good, it could succeed at that.

If VMware/Microsoft/Nutanix AHV write hypervisors for it, that really could be the end of Intel as we know it. Since it sounds like (at the moment) that these chips have lower single core IPC but the chip scales up into many cores, the server virtualization market is where they can really hit Intel's bottom line.
 
  • Like
Reactions: ChadD
like this
If there is a good x86-64 emulator and the performance is good, it could succeed at that.

If VMware/Microsoft/Nutanix AHV write hypervisors for it, that really could be the end of Intel as we know it. Since it sounds like (at the moment) that these chips have lower single core IPC but the chip scales up into many cores, the server virtualization market is where they can really hit Intel's bottom line.
Ya Intel is in serious trouble on the server front. There not the only x86 option, AMD has been eating into that market as well. Intel still has some advantages in software stack optimization... Intel has some great software support on the Linux side. AMD needs to step that up if there going to steal the real juicy customers. But ya AMD and Intel are both going to be dealing with serious ARM competition going forward. Its not just MS, it seems everyone has a ARM server play. With ARM cloud servers starting to get used more and more the software side is flushing out pretty quickly. I think its probably some years before x86 is pushed completely out of the server market.... ARM is gaining market share fast though in those markets, I think it isn't crazy to think in a few years that market could easily be 50/50. At that point if ARM keeps gaining its just a matter of how long x86 hangs on at all really.
 
Ya Intel is in serious trouble on the server front. There not the only x86 option, AMD has been eating into that market as well. Intel still has some advantages in software stack optimization... Intel has some great software support on the Linux side. AMD needs to step that up if there going to steal the real juicy customers. But ya AMD and Intel are both going to be dealing with serious ARM competition going forward. Its not just MS, it seems everyone has a ARM server play. With ARM cloud servers starting to get used more and more the software side is flushing out pretty quickly. I think its probably some years before x86 is pushed completely out of the server market.... ARM is gaining market share fast though in those markets, I think it isn't crazy to think in a few years that market could easily be 50/50. At that point if ARM keeps gaining its just a matter of how long x86 hangs on at all really.
This has been an upcoming problem for a while and I'm sure both AMD and Intel have something in the works. The problem is that for most stuff out there ARM is "fast enough", and large core counts and power efficiency are starting to be a much higher priority. ARM now has enough market penetration that the software exists and there are communities and companies investing more and more into it. This isn't a problem that is going away for either AMD or Intel and I have a hard time believing that they are short-sighted enough to not realize this.
 
This has been an upcoming problem for a while and I'm sure both AMD and Intel have something in the works. The problem is that for most stuff out there ARM is "fast enough", and large core counts and power efficiency are starting to be a much higher priority. ARM now has enough market penetration that the software exists and there are communities and companies investing more and more into it. This isn't a problem that is going away for either AMD or Intel and I have a hard time believing that they are short-sighted enough to not realize this.

ARM is fastest in many stuff. The fastest supercomputer in the world is ARM. Altra Q80-33 is the fastest server for several server and HPC workloads. Apple M1 is the fastest mobile chip in its kind...
 
Back
Top