News Arm-Based Laptops Gaining Share Despite PC Market Weakness: Report

[Admin]

Arm SoCs to power 25% of laptops in five years, says Counterpoint Research.

Arm-Based Laptops Gaining Share Despite PC Market Weakness: Report : Read more

 

[bit_user]

I just hope some ARM-based models make a credible play for mid-market. I want an ARM-based laptop, but I'm sure not going to pay extra for one.

Qualcomm has so far been positioning theirs as a premium product, and I think that's a huge mistake. Premium status must be earned. They'll do much better starting out by offering a superior value. Once their reputation has been established, then they can make a play for the up-market segment.

Oh, and I don't care for Windows 11. Give me decent mainstream Linux support, too. If they can support ChromeOS, that should be easy.

 

[RichardtST]

Yeah. Bound to happen. Personally, I absolutely hate ARM assembly. Like, who the heck needs a barrel shifter or an assembler that can't handle constants with more than a few bits set at a time? OMG. Such garbage. That said, it is still smaller and faster than the x86 instruction set. ARM should, in theory, eventually blow away x86. Of course ARM isn't without it's own structural and bloating problems, which is why riscV is lurking... The x86 instruction architecture is so overbloated with useless instructions that it will, as is natural, collapse under it's own weight. AMD and Intel have one chance, right now, to strip it hard and remove 90% of the instruction set, restructure for speed and surprise us all.... or go the way of the dinosaurs.

 

[PlaneInTheSky]

My ARM laptop simply has much better battery life. I don't have the noisy fans like on x86 either.

I could sort of mitigate the fan noise on x86 by turning off Turbo in the BIOS, but the laptops are still much noisier than ARM.

If Steam Deck had been ARM based like Nintendo Switch, I might have kept it. But Steam deck is currently a giant pointless battery hogging x86 PC, I hate the noisy fans and I returned it.

A mobile device without good battery life is pointless.

 

[JamesJones44]

Yeah. Bound to happen. Personally, I absolutely hate ARM assembly. Like, who the heck needs a barrel shifter or an assembler that can't handle constants with more than a few bits set at a time? OMG. Such garbage. That said, it is still smaller and faster than the x86 instruction set. ARM should, in theory, eventually blow away x86. Of course ARM isn't without it's own structural and bloating problems, which is why riscV is lurking... The x86 instruction architecture is so overbloated with useless instructions that it will, as is natural, collapse under it's own weight. AMD and Intel have one chance, right now, to strip it hard and remove 90% of the instruction set, restructure for speed and surprise us all.... or go the way of the dinosaurs.

I 100% agree with this, they should have done it with AMD64 IMO. If I were Intel/AMD I would start weening people off legacy instructions/software. There is very little reason why a DOS 1.0 application needs to run natively anymore for example, let MS and other OS makers emulate those modes if they still want to support them. They are going to need to do it for ARM anyway so Intel/AMD should use the opportunity to introduce a slim/efficient IS. Even development wise it only took about a year for most software compilers to add ARM support once Apple shifted, so it's not like it would be hard for languages to adapt, heck it would probably easier now that they already support two different ISs (in some cases 3 as some already support riscV), adding a 3rd/4th would probably be minimal effort.

 

[TerryLaze]

This is based on apple laptops taking off, and I doubt that apple fans will grow by such an amount in 5 years. Every apple fan already got one, if they replace their current one it will not increase the overall market share.
Beyond that the arm based laptops will, at most, take the place of the eepc which I doubt ever had a big share.
If ARM had such a big potential then AMD and Intel would be the first companies to release products in that market, ryzen was supposed to be a hybrid platform with ARM, they dropped that real fast.

 

[JamesJones44]

This is based on apple laptops taking off, and I doubt that apple fans will grow by such an amount in 5 years. Every apple fan already got one, if they replace their current one it will not increase the overall market share.
Beyond that the arm based laptops will, at most, take the place of the eepc which I doubt ever had a big share.
If ARM had such a big potential then AMD and Intel would be the first companies to release products in that market, ryzen was supposed to be a hybrid platform with ARM, they dropped that real fast.

Qualcomm and Microsoft are partnered to release ARM based laptops using the same tech Apple did. It's being held up right now by Qualcomm and Arm Holdings fighting over price, but that will get resolved at some point in the next year most likely. I don't think you can ignore that. MS has updated all of their tool chains to support ARM. Apps from the MS Store can already contain both ARM and x86 fat binaries. Clearly MS thinks there is an opportunity with ARM.

 

[bit_user]

I absolutely hate ARM assembly. Like, who the heck needs a barrel shifter or an assembler that can't handle constants with more than a few bits set at a time? OMG. Such garbage.

I think it's pretty cool. Since instructions words are 32-bits (which is part of the key for making them fast & efficient to decode), immediate operands are necessarily limited to 12 bits. However, the immediate encoding scheme does something clever, by effectively giving you an 8-bit mantissa with a 4-bit exponent. The fact that it does a rotate seems like a clever hack to let you cover a few more cases than you could hit with a straight shift-left.

ARM immediate value encoding

alisdair.mcdiarmid.org

Of course ARM isn't without it's own structural and bloating problems, which is why riscV is lurking...

RISC-V didn't arise because ARM was bloated, and that's also not why people are adopting it. They're adopting it because it's royalty-free and you're also free to pick & choose which extentions to support and you can even integrate your own custom extensions.

ARM is very rigid on all of these points. Your main point of flexibility with ARM is simply to choose which revision of the ISA to support, and then there are just a couple optional extensions, like SVE (although SVE2 is now a mandatory part of ARMv9-A).

AMD and Intel have one chance, right now, to strip it hard and remove 90% of the instruction set, restructure for speed and surprise us all.... or go the way of the dinosaurs.

It's a lot of work to add toolchain, OS, library, and application support for a new ISA. I highly doubt either Intel or AMD will invent a new one.

If/when they leave x86, they'll probably jump on the RISC-V bandwagon. Before, they might've gone with ARM, but ever since Nvidia tried to buy it, I think they would've steered clear. Even now that the acquisition is off, the crazy lawsuit ARM leveled at Qualcomm for their Nuvia cores/acquisition is possibly worse for any CPU/SoC-makers thinking of designing their own cores.

Even development wise it only took about a year for most software compilers to add ARM support once Apple shifted, so it's not like it would be hard for languages to adapt,

No...
🤦

ARM had done decades of work, leading up to that point. And even Apple had over a decade of work invested in ARM, from their iPods and iPhones. And even with all of that groundwork in place, it was probably more like 3 years to port over MacOS and all of their in-house software, SDKs, and libraries.

...not to mention Google and all of the Android developers, the Raspberry Pi community, all of the ARM embedded develpers, etc. All of them helped contribute to porting and improving compilers and software packages on ARM. Certainly some billions of engineering hours have gone into the ARM software ecosystem, even by the point when Apple started porting MacOS to it.

 

[bit_user]

If ARM had such a big potential then AMD and Intel would be the first companies to release products in that market, ryzen was supposed to be a hybrid platform with ARM, they dropped that real fast.

AMD launched the Opteron A1100 before the market was really ready for ARM servers.

https://www.amd.com/system/files/documents/hierofalcon-product-brief.pdf

And the reason they furloughed the K12 is both the market immaturity and because they were virtually at the brink of bankruptcy. So, they wisely kept Zen and put the K12 on the back burner.

 

[Deleted member 14196]

I wouldn’t mind one, but yeah I don’t want to pay a premium price but I do want a good one

 

[Sleepy_Hollowed]

I’m literally awaiting for Lenovo to update their ARM laptops to see if I can have my Linux workstation that has long battery life.

 

[TerryLaze]

Qualcomm and Microsoft are partnered to release ARM based laptops using the same tech Apple did. It's being held up right now by Qualcomm and Arm Holdings fighting over price, but that will get resolved at some point in the next year most likely. I don't think you can ignore that. MS has updated all of their tool chains to support ARM. Apps from the MS Store can already contain both ARM and x86 fat binaries. Clearly MS thinks there is an opportunity with ARM.

Just saying....

(MS surface sponsorship tablets used as stands for ipads... )

People that are attracted to arm already have smartphones and tablets, how much need is there for them to have the same thing in a larger form factor without it being much, or even any, faster?!

 

[JamesJones44]

No...
🤦

ARM had done decades of work, leading up to that point. And even Apple had over a decade of work invested in ARM, from their iPods and iPhones. And even with all of that groundwork in place, it was probably more like 3 years to port over MacOS and all of their in-house software, SDKs, and libraries.

...not to mention Google and all of the Android developers, the Raspberry Pi community, all of the ARM embedded develpers, etc. All of them helped contribute to porting and improving compilers and software packages on ARM. Certainly some billions of engineering hours have gone into the ARM software ecosystem, even by the point when Apple started porting MacOS to it.

I'm specifically talking about Tool Chains. Java, .NET, Rust, Swift, Node, etc. Yes, I agree the shift took a very long time from start to finish (investment in general started in the 90s (arguably late 80s)), but when Apple first released the A12z dev machine in a Mac Mini there were no freely available JVMs that supported ARM, .NET did not support compiling into ARM, Rust did not support compiling into ARM, etc. Once the M1 MacBook Air officially launched about 6 months later more than half of those supported ARM, fast forward to the M1 Pro/Max and all of the supported programs targeted for ARM. The shift for toolchains was very fast. My point is, now that all of those Tool Chains support multiple CPU architectures that they never did before, adding additional will be easier from a technical point of view (especially those that use frontends like LLVM).

 

[bit_user]

Just saying....

(MS surface sponsorship tablets used as stands for ipads... )

Spoiler

When is that from? And how do you know those were the ARM-based surface tablets? They could've been the x86 versions.

Anyway, we don't know why. Maybe some people just want to use on air what they're most comfortable with. Because, live TV. Or maybe the surface battery was depleted. We really don't know.

People that are attracted to arm already have smartphones and tablets, how much need is there for them to have the same thing in a larger form factor without it being much, or even any, faster?!

But it is. If you look up the specs of Qualcomm's laptop SoC, they have more X-series high-performance cores than their smartphone SoCs. Also, they're clocked higher, because power and cooling are both more capacious.

Most importantly, Qualcomm will initially ship Nuvia cores in their laptop SoC, only. So, when those things finally launch, next year, do expect to see significantly better performance.

 

[bit_user]

when Apple first released the A12z dev machine in a Mac Mini there were no freely available JVMs that supported ARM, .NET did not support compiling into ARM, Rust did not support compiling into ARM, etc. Once the M1 MacBook Air officially launched about 6 months later more than half of those supported ARM,

Even that work surely didn't all start right when Apple shipped that mini. And I can assure you that .NET did support ARM, just maybe not on Mac. Microsoft has been working on running various flavors of Windows on ARM since Windows Phone, and probably even way back to Windows CE.

My point is, now that all of those Tool Chains support multiple CPU architectures that they never did before, adding additional will be easier from a technical point of view (especially those that use frontends like LLVM).

You know that Java didn't even start on x86, right? It was developed by a company called Sun, and it started on their own SPARC CPUs.

As for LLVM, Google has been exclusively using that on Android for more than 5 years. I was actually doing a bit of native development on Android, when they dropped support for GCC. I don't know why Rust wouldn't have supported compiling on ARM, but I imagine it probably wasn't a lot of work to fix.

My point is that I'd be pretty shocked if either Intel or AMD introduced a completely new ISA. Especially after the black eye Intel got from IA64.

 

[TerryLaze]

When is that from? And how do you know those were the ARM-based surface tablets? They could've been the x86 versions.

It is from x86, it was just to show that just because MS is doing something it doesn't mean that it will do something to the market.
It was a pretty big thing
They paid a lot of money and everybody kept calling them ipads and kept using their ipads, propped up on the sponsored surfaces.
https://www.chicagotribune.com/busi...ft-surface-loss-0126-biz-20160125-column.html

But it is. If you look up the specs of Qualcomm's laptop SoC, they have more X-series high-performance cores than their smartphone SoCs. Also, they're clocked higher, because power and cooling are both more capacious.

Cost? Power draw? performance?
They would all have to be much better than the x86 options for people to choose something new and, for them, unproven.

 

[bit_user]

It is from x86, it was just to show that just because MS is doing something it doesn't mean that it will do something to the market.
It was a pretty big thing
They paid a lot of money and everybody kept calling them ipads and kept using their ipads, propped up on the sponsored surfaces.
https://www.chicagotribune.com/busi...ft-surface-loss-0126-biz-20160125-column.html

That article is from 2016. I think it's more of a distraction than actually relevant to the discussion, here.

I think the main point about MS and ARM is that how well Windows runs on ARM could give ARM a big boost in gaining traction in the laptop and PC desktop space. All we need now is for the next XBox to be ARM-based, and it'd be the beginning of the end for x86.

 

[TerryLaze]

That article is from 2016. I think it's more of a distraction than actually relevant to the discussion, here.

So what happened in the last 6 years that would make an MS product more desirable now?!

I think the main point about MS and ARM is that how well Windows runs on ARM could give ARM a big boost in gaining traction in the laptop and PC desktop space.

No, that would give MS good traction in cutting into the android OS monopoly.
That's a huge amount of money that will only get bigger and bigger and MS wants a piece of that action.
ARM laptops and desktops will only be appealing to people that already use arm.

If I would want to do my everyday tasks and save energy doing it I could get a docking station for my very old smartphone and browse the net and read emails on a monitor using keyboard and mouse.

All we need now is for the next XBox to be ARM-based, and it'd be the beginning of the end for x86.

The xbox 360 and the PS3 both had some kind of risc cpu.
Developers whined so long about it that both companies switched to x86 so that devs could stop devving for risc.
They are not going to switch back anytime soon, certainly not until arm can hit the performance level of whatever the next gen of x86 consoles would hit.
They won't release a console that is slower then the current one, or just about as fast.

 

[bit_user]

So what happened in the last 6 years that would make an MS product more desirable now?!

Which? You mean the x86 surface? I can't really comment on MS Surface, but Windows on ARM is about more than Surface.

As for Windows on ARM, what's changed recently is that Windows 11 now supports emulation of x86-64 on ARM. Previously, they only supported emulation of 32-bit x86 apps.

No, that would give MS good traction in cutting into the android OS monopoly.

We've been through this, before. ChromeOS is not Android. Very different. Nobody is running Android on laptops.

More to the point, ChromeOS mostly dominates the bottom end of the market, which isn't very interesting to MS, because margins are so low. That's one of the things that killed Windows on phones - there's no way they could compete with "free".

Anyway, it appears your argument isn't with me, but rather the authors of this study.

ARM laptops and desktops will only be appealing to people that already use arm.

You mean like in their phones? That's 100% of people with smartphones.

Based on the article, Apple's transition seems to be going very well. So, that's a good sign for ARM and bad news for x86.

It's entirely plausible corporate users could transition from Windows on x86 to Windows on ARM rather seamlessly.

If I would want to do my everyday tasks and save energy doing it I could get a docking station for my very old smartphone and browse the net and read emails on a monitor using keyboard and mouse.

Good luck with that, but I once tried connecting a monitor, mouse, and keyboard to an Android tablet and it wasn't much better than just using the tablet. Still nowhere near an experience like ChromeOS or Windows.

The xbox 360 and the PS3 both had some kind of risc cpu.

PowerPC.

Developers whined so long about it that both companies switched to x86 so that devs could stop devving for risc.

There are a few reasons for this, but the difference with ARM is that most major game engines already support ARM, in order to enable gaming on phones, tablets, and Nintendo Switch. And, for similar reasons, a lot of game developers have experience with ARM that they didn't have with PowerPC.

They are not going to switch back anytime soon,

If Microsoft wanted to spur games to natively support Windows on ARM, it would be a strategic move they could make. I'm not saying it will happen, but it it would be a game changer (excuse the pun) if it did.

They won't release a console that is slower then the current one, or just about as fast.

I wonder how ARM's X3 or V2 cores compare with leading x86 cores. Within a console CPU's power budget, they might be very comparable.

A big selling point might be cost, since ARM cores tend to be much smaller than x86.

 

[Amdlova]

Apple has make a good move with m1 booth is I'm same league of Intel amd and nvidia. 5% upgrade per evolution I'm done with this market.

 

[TerryLaze]

Which? You mean the x86 surface? I can't really comment on MS Surface, but Windows on ARM is about more than Surface.

As for Windows on ARM, what's changed recently is that Windows 11 now supports emulation of x86-64 on ARM. Previously, they only supported emulation of 32-bit x86 apps.

Stop talking about how windows/MS looks at things.
MS can try to increase the market share but only customers really do change it,
Customers don't want to emulate something that they can run fine natively.
MS failed 6 years ago running windows natively and you argue that they will do better emulating it... ?!

We've been through this, before. ChromeOS is not Android. Very different. Nobody is running Android on laptops.

More to the point, ChromeOS mostly dominates the bottom end of the market, which isn't very interesting to MS, because margins are so low. That's one of the things that killed Windows on phones - there's no way they could compete with "free".

Oh, so MS is after the super small market share of people that use chromeOS pc/laptops...makes much more sense than MS trying to get into the absolutely huge android market.

Anyway, it appears your argument isn't with me, but rather the authors of this study.

The authors of this article are completely out of touch and think that the annual increase that happened once from apple fans buying apple products during release is going to be an ongoing trend.

You mean like in their phones? That's 100% of people with smartphones.

Based on the article, Apple's transition seems to be going very well. So, that's a good sign for ARM and bad news for x86.

It's entirely plausible corporate users could transition from Windows on x86 to Windows on ARM rather seamlessly.

No it's a good sign for apple, but we all knew this already, apple has their audience in a death grip, it was the same when they switched from mc68000 to powerPC and then when they switched from powerPC to x86.
Those changes didn't change anything for the overall market share of these CPUs.

Good luck with that, but I once tried connecting a monitor, mouse, and keyboard to an Android tablet and it wasn't much better than just using the tablet. Still nowhere near an experience like ChromeOS or Windows.

But that's the point, you can do it.
And for most people "just using the tablet" is everything they will ever need.
If they need anything better they get a real laptop or desktop.
As I said, the best case ARM is going to take the place of the eeePCs.

PowerPC.


There are a few reasons for this, but the difference with ARM is that most major game engines already support ARM, in order to enable gaming on phones, tablets, and Nintendo Switch. And, for similar reasons, a lot of game developers have experience with ARM that they didn't have with PowerPC.

Yes, PowerPC is a risc CPU.

Yay for the future of having consoles that will be running smartphone games, and since they release the same games on PC, yay for a future of smartphone games on PC ...great argument.

If Microsoft wanted to spur games to natively support Windows on ARM, it would be a strategic move they could make. I'm not saying it will happen, but it it would be a game changer (excuse the pun) if it did.

How would that be a game changer?!
Windows on arm would take more resources and games would run much worse on the same hardware.
The xbox at least uses a super slim OS, I don't know if it's still based on windows but it used to be.

 

[bit_user]

Stop talking about how windows/MS looks at things.

I don't understand this comment. You asked what changed and I answered it.

Customers don't want to emulate something that they can run fine natively.

That's why I said business customers should be an easy win. They mostly run stuff like browsers and MS Office, which are all native.

MS failed 6 years ago running windows natively and you argue that they will do better emulating it... ?!

How do you figure they "failed"?

ARM is getting better year by year. Back in 2016, their cores were a lot more narrow, and they didn't have the special "X" versions or SVE. In the server market, they've shown they're already competitive with x86.

Oh, so MS is after the super small market share of people that use chromeOS pc/laptops...makes much more sense than MS trying to get into the absolutely huge android market.

No, go back and read what I said: it isn't very interesting to MS

The authors of this article are completely out of touch and think that the annual increase that happened once from apple fans buying apple products during release is going to be an ongoing trend.

Well, don't take it out on me.

it's a good sign for apple, but we all knew this already, apple has their audience in a death grip, it was the same when they switched from mc68000 to powerPC and then when they switched from powerPC to x86.
Those changes didn't change anything for the overall market share of these CPUs.

What you can't deny is that their ARM-based Macs are good. It proves the ARM ISA is up to the task of handling the daily and diverse computing workloads of Apple's users.

Yay for the future of having consoles that will be running smartphone games, and since they release the same games on PC, yay for a future of smartphone games on PC ...great argument.

I didn't say an ARM-based XBox would run smartphone games. My point was just that the gaming industry and games developers now have a lot more experience with ARM than they ever had with PowerPC. It wouldn't be such a big transition for them.

How would that be a game changer?!

Because you'd have loads of Direct3D games optimized for ARM, that would very easily port to Windows on ARM.

Windows on arm would take more resources and games would run much worse on the same hardware.

You mean as compared to XBox? If so, it's not a disparity that doesn't already exist between XBox and x86 PCs.

 

[chickenballs]

with everything soldered the only way to upgrade is to throw it away and buy a new one which is not "sustainable"

 

[JamesJones44]

My point is that I'd be pretty shocked if either Intel or AMD introduced a completely new ISA. Especially after the black eye Intel got from IA64.

I totally agree with that. What they should do and will do are very different.

 

[Sleepy_Hollowed]

Oh, Lenovo already released their Snapdragon based ARM laptop, it's quite decent, and runs Windows quite well, now I'm on the lookout for anyone running Linux on it, which is what I REALLY care about.

https://www.digitaltrends.com/computing/thinkpad-x13s-hands-on-new-specs-price-photos/