• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

Windows Auto Super Resolution Limited to Copilot+ PCs with Snapdragon X Elite SoCs Only—Not x86

btarunr

Editor & Senior Moderator
Staff member
Joined
Oct 9, 2007
Messages
47,670 (7.43/day)
Location
Dublin, Ireland
System Name RBMK-1000
Processor AMD Ryzen 7 5700G
Motherboard Gigabyte B550 AORUS Elite V2
Cooling DeepCool Gammax L240 V2
Memory 2x 16GB DDR4-3200
Video Card(s) Galax RTX 4070 Ti EX
Storage Samsung 990 1TB
Display(s) BenQ 1440p 60 Hz 27-inch
Case Corsair Carbide 100R
Audio Device(s) ASUS SupremeFX S1220A
Power Supply Cooler Master MWE Gold 650W
Mouse ASUS ROG Strix Impact
Keyboard Gamdias Hermes E2
Software Windows 11 Pro
Microsoft Auto Super Resolution (ASR), the standardized game super-resolution based performance enhancement, is initially only being offered to Copilot+ AI PCs powered by Qualcomm Snapdragon X Elite processors, says the Microsoft FAQ for Copilot+ AI PCs. "At initial launch, this feature will be exclusive to Copilot+ PCs equipped with a Qualcomm Snapdragon X Elite processor and a curated set of games that can be found here, a third-party open-source site that Microsoft has contributed compatibility data to," the FAQ answer reads, in response to the question "What is automatic super resolution?"

The way we understand this, Microsoft ASR will be launched initially only Windows Arm devices, specifically those powered by the Snapdragon X Elite SoC. The Snapdragon X Plus is excluded; but more importantly, all x86-64 platforms (Intel or AMD) are excluded from the initial rollout. This doesn't mean that ASR won't make it to x86, it just will at a later date. Copilot+ devices based on the Snapdragon X Elite tend to have a high degree of design collaboration between the OEM and Microsoft, and Redmond wants to use them as tech demonstrators, mostly since these platforms lack the usual super resolution tech such as AMD FSR, Intel XeSS, or NVIDIA DLSS. It's important to note that ASR is a super-resolution tech that's not meant to be confused with DirectSR, the API it's based on. Development of DirectSR for Windows PCs on x86-64 platforms continues.



View at TechPowerUp Main Site | Source
 
Fake differentiation between Windows versions and now product categories. The Microsoft way indeed...

At this rate, I wouldn't be surprised if somebody found out that Windows 11 is actually already in Windows 7!
 
Fake differentiation between Windows versions and now product categories. The Microsoft way indeed...
IMO, this has little to do with differentiating between versions, more to do with Microsoft trying to pull an Apple and move as many components of their platforms in house, which would be easier to do with ARM than x86, and in which case the real joke would be on Qualcomm. I see another Imagination Tech drama on the horizon.

The tech will be ported over to x86 eventually.
 
The point of CoPilot+ is the integrated NPU. Snapdragon X is the first and in the 1st party Surface devices but Microsoft did also mention working with Intel and AMD on this starting with Lunar Lake and Strix Point - key seems to be an integrated NPU.
 
Wasn't directSR using DLSS, FSR and XESS based on hardware? If this ASR is based on directSR what does it do?
 
So do you think MS will add "AI support hardware" requirement to Windows Hardware requirements??!
 
The point of CoPilot+ is the integrated NPU. Snapdragon X is the first and in the 1st party Surface devices but Microsoft did also mention working with Intel and AMD on this starting with Lunar Lake and Strix Point - key seems to be an integrated NPU.
x86 machines are already capable of handling ML-based upscaling workloads, as evident in whatever crap PC GPU vendors make. Requiring an embedded NPU is an artificial requirement.

Argument could be made that GPU AI-accelerators aren't standardized, but that's not an issue considering MS already has a long established system to force standardization over vendors (DirectX), which primarily addresses the market ML upscalers prioritize in the first place, unlike this copilot crap.

Also keep in mind the NPU is only relevant here because the graphics processing is done in the SoC itself. That would not be the case with x86 gaming, and bouncing data from the GPU back to the promissed CPU's NPU is obviously out of the question.
 
Oh man what a hilarious and blatant attempt to extract more money from unsuspecting n00b customers by offering next to nothing.

I think AI wrote this marketing strategy, its so silly, its unreal. Theoretically it should work, but its like the human factor is ignored entirely. Do they really think we're this stupid? Anyone can visit a web page and get AI. Why would anyone want a dedicated service or button for it on their OS that needs separate updating? Whole crowds have been trained for over a decade that everything is into the cloud now. And MS goes for... a button on your local Kb. Or fake differentiation 'because NPU' so others like Intel can sell more hardware that barely progressed over the same last decade.
 
That means it’s not ready and hardly anyone will be using it in the foreseeable future.
 
Oh man what a hilarious and blatant attempt to extract more money from unsuspecting n00b customers by offering next to nothing.

I think AI wrote this marketing strategy, its so silly, its unreal. Theoretically it should work, but its like the human factor is ignored entirely. Do they really think we're this stupid? Anyone can visit a web page and get AI. Why would anyone want a dedicated service or button for it on their OS that needs separate updating? Whole crowds have been trained for over a decade that everything is into the cloud now. And MS goes for... a button on your local Kb. Or fake differentiation 'because NPU' so others like Intel can sell more hardware that barely progressed over the same last decade.
I've seen an example of copilot being able to "see" what's happening in your game and giving you advice. Seems hard to do with a web browser. And it's not just Microsoft, Google and Apple are about to deploy the same kind of AI integration in their device. It's not just about a being "glorified web engine" anymore, but being able to comment on what the user is doing on its devices as a whole.
Gemini on Android is a new kind of assistant that uses generative AI to help you be more creative and productive. This experience, which is integrated into Android, is getting even better at understanding the context of what’s on your screen and what app you’re using.

Soon, you’ll be able to bring up Gemini's overlay on top of the app you're in to easily use Gemini in more ways. For example, you can drag and drop generated images into Gmail, Google Messages and other places, or tap “Ask this video” to find specific information in a YouTube video. If you have Gemini Advanced, you’ll also have the option to “Ask this PDF” to quickly get answers without having to scroll through multiple pages. This update will roll out to hundreds of millions of devices over the next few months.

And we’ll continue to improve Gemini to give you more dynamic suggestions related to what’s on your screen.
Apple in Talks With Google to Bring Gemini AI Features to iPhone - MacRumors
 
Oh man what a hilarious and blatant attempt to extract more money from unsuspecting n00b customers by offering next to nothing.

I think AI wrote this marketing strategy, its so silly, its unreal. Theoretically it should work, but its like the human factor is ignored entirely. Do they really think we're this stupid? Anyone can visit a web page and get AI. Why would anyone want a dedicated service or button for it on their OS that needs separate updating? Whole crowds have been trained for over a decade that everything is into the cloud now. And MS goes for... a button on your local Kb. Or fake differentiation 'because NPU' so others like Intel can sell more hardware that barely progressed over the same last decade.

It's all about cramming what essentially amounts to AI "snake oil" down peoples throats. From what I've seen so far, the use cases where an LLM can do something useful, are very very limited. Chat bots, programming aid, video and picture creation and manipulation, translation.

The reason is they've got nothing else, there's little to no reason to upgrade from hardware, software, in the last 5 or even 10 years for most or nearly all use cases.

So they're trying to create one, because sales collapsing is the alternative.
 
I’m ready for Valve to step up and take over the reigns of PC gaming. Steam OS should get APIs and standards for games going forward. They should make a super sampling standard. To get started, a bundle called ‘Three’ should be released exclusive to Steam OS:

Half-life 3
Portal 3
Day of defeat 3
Left for Dead 3
Counterstrike 3
DOTA 3
Team Fortress 3
 
They should make a super sampling standard.
DirectSR is the standard API for "Super Resolution". Common API for games and other applications to use it, manufacturers can deal with the implementation side.
 
On a scale of not interested, dont care levels of news, this fits the zero Fs to give catageory.
 
Last edited:
I'm starting to get very lost on all the AI on desktop OSes thing. Like, the super resolution thing sounds cool and all but at the same time they want AI remembering all you do on your PC which I really don't want. At this point all I want to know if they are going to spaghetti everything in the same package or if it will be separate pieces inside the OS and someone will come with an O&O Shutup AI to only have what you really want. Because with SteamOS and all the advancements in gaming in Linux I'm really considering going full Linux and fight with Wine at this point.
 
It's all about cramming what essentially amounts to AI "snake oil" down peoples throats. From what I've seen so far, the use cases where an LLM can do something useful, are very very limited. Chat bots, programming aid, video and picture creation and manipulation, translation.

The reason is they've got nothing else, there's little to no reason to upgrade from hardware, software, in the last 5 or even 10 years for most or nearly all use cases.

So they're trying to create one, because sales collapsing is the alternative.
This, this this this. AI is another "big paradigm" to drive silicon, software, and general computer equipment development and sales just like "the cloud" was back in the late '00s/'10s. AI is a godsend for the industry as a whole because it gives a reason for all these jobs to exist and people to continue to work and get paid and for the market as a whole to expand. As stupid as it is, from an economic growth standpoint it's extremely beneficial, even if the real world use case isn't there. This is why these companies are pursuing it with reckless abandon. It's a Tennessee Valley Authority for the electronics industry, the CHIPS act is a perfect example of this. A bunch of jobs need filled to prop this shit up, silicon design, marketing, software development, sales, roadmapping, project management, etc. It's a way to justify inflated budgets and extreme spending to get this stuff up and running and makes economic development and employment reports look very good.

1716385217273.png


Not saying I agree with any of it, just giving an explanation for why this is being crammed down our throats with a mallet. There's big, big bucks to be had all around in developing the hell out of this stuff and the tech market has been looking for a reason to grow for a while now. Every company that's even tangentially related to the computer industry sees this as an opportunity for explosive growth.
 
Gemini can suck my balls, it was disabled as soon as it was done updating. Now there is a "Capture user actions and errors" for Google products on the phone too that is disabled, no mention of who all gets the data, and what they consider an error.

Chrome, Gmail, Messages, Youtube.

They are trying to use AI to datamine humans.
 
This, this this this. AI is another "big paradigm" to drive silicon, software, and general computer equipment development and sales just like "the cloud" was back in the late '00s/'10s. AI is a godsend for the industry as a whole because it gives a reason for all these jobs to exist and people to continue to work and get paid and for the market as a whole to expand. As stupid as it is, from an economic growth standpoint it's extremely beneficial, even if the real world use case isn't there. This is why these companies are pursuing it with reckless abandon. It's a Tennessee Valley Authority for the electronics industry, the CHIPS act is a perfect example of this. A bunch of jobs need filled to prop this shit up, silicon design, marketing, software development, sales, roadmapping, project management, etc. It's a way to justify inflated budgets and extreme spending to get this stuff up and running and makes economic development and employment reports look very good.

View attachment 348401

Not saying I agree with any of it, just giving an explanation for why this is being crammed down our throats with a mallet. There's big, big bucks to be had all around in developing the hell out of this stuff and the tech market has been looking for a reason to grow for a while now. Every company that's even tangentially related to the computer industry sees this as an opportunity for explosive growth.
As the whole, the use case for A.I (or ML) is real, for example all the latest professional cameras have A.I enhanced autofocus. Subject recognition, and eye autofocus on humans and animals. The future Canon R1 will even be able to automatically focus on the subject who's in control of the ball in sport events.

A.I is just in the "let's see what's stick" period. If the OS level A.I assistant don't get used, it's just going to meet the same fate as Cortana in a few years
 
I'm starting to get very lost on all the AI on desktop OSes thing. Like, the super resolution thing sounds cool and all but at the same time they want AI remembering all you do on your PC which I really don't want. At this point all I want to know if they are going to spaghetti everything in the same package or if it will be separate pieces inside the OS and someone will come with an O&O Shutup AI to only have what you really want. Because with SteamOS and all the advancements in gaming in Linux I'm really considering going full Linux and fight with Wine at this point.
It's like a classic Bond villain, they tell us what they're doing with their maniacal "good intentions," and then they overplay their hand. I can actually see why Apple is slow-playing this a bit more. It's a shame they are getting pressured, because usually they aren't the first to use any new tech, but rather wait until there's maybe a good purpose for it and try to implement it pretty well. Now they are in talks with AI sources, and they might be just as bad as the rest. What do customers want?
Gemini can suck my balls, it was disabled as soon as it was done updating. Now there is a "Capture user actions and errors" for Google products on the phone too that is disabled, no mention of who all gets the data, and what they consider an error.

Chrome, Gmail, Messages, Youtube.

They are trying to use AI to datamine humans.
"We noticed you aren't using our services enough, and we really don't like the way you're thinking. These are clearly errors, and you have been reported!"
As the whole, the use case for A.I (or ML) is real, for example all the latest professional cameras have A.I enhanced autofocus. Subject recognition, and eye autofocus on humans and animals. The future Canon R1 will even be able to automatically focus on the subject who's in control of the ball in sport events.

A.I is just in the "let's see what's stick" period. If the OS level A.I assistant don't get used, it's just going to meet the same fate as Cortana in a few years
No doubt there are good use cases. In the case of a digital camera, that's not as worrisome since cameras are still classically programmed with updates, and there's no telemetry back to the mothership. It's all based on the traditional model of design, test, release, and improve. AI on devices that have our entire lives on them is a completely different animal.
 
No doubt there are good use cases. In the case of a digital camera, that's not as worrisome since cameras are still classically programmed with updates, and there's no telemetry back to the mothership. It's all based on the traditional model of design, test, release, and improve. AI on devices that have our entire lives on them is a completely different animal.
LocalAI exists for this type of case. I'm all for AI as long as there is a way to limit it to local runtime and not report any telemetry outside of the network. Much like what the NPU major use case is on the photos and camera apps on smartphones since 2017 (and even before that with the GPUs).
 
It's all about cramming what essentially amounts to AI "snake oil" down peoples throats. From what I've seen so far, the use cases where an LLM can do something useful, are very very limited. Chat bots, programming aid, video and picture creation and manipulation, translation.

The reason is they've got nothing else, there's little to no reason to upgrade from hardware, software, in the last 5 or even 10 years for most or nearly all use cases.

So they're trying to create one, because sales collapsing is the alternative.
That's my view as well.

Remarkably similar to RT. And Crypto.
'We have this performance on tap, must find purpose'
As if this has ever resulted in a worthwhile endeavour...
 
I’m ready for Valve to step up and take over the reigns of PC gaming. Steam OS should get APIs and standards for games going forward. They should make a super sampling standard. To get started, a bundle called ‘Three’ should be released exclusive to Steam OS:

Half-life 3
Portal 3
Day of defeat 3
Left for Dead 3
Counterstrike 3
DOTA 3
Team Fortress 3

Oh my god, I want it so bad. With no depot leaks and zero marketing.

Would break internet backbones for at least 2 days, and completely reverse a gaming industry crash.
 
Back
Top