Jump to content

Welcome to ExtremeHW

Welcome to ExtremeHW, register to take part in our community, don't worry this is a simple FREE process that requires minimal information for you to signup.

 

Registered users can: 

  • Start new topics and reply to others.
  • Show off your PC using our Rig Creator feature.
  • Subscribe to topics and forums to get updates.
  • Get your own profile page to customize.
  • Send personal messages to other members.
  • Take advantage of site exclusive features.
  • Upgrade to Premium to unlock additional sites features.
IGNORED

Windows 11 24H2: AI Super Resolution Technology, Works Across All PCs With AI NPUs


bonami2

Recommended Posts

Social Media Manager
1.5k 845
Quote

According to a post by @PhantomofEarth, it looks like Microsoft will be offering its own AI SR or Super Resolution technology within the upcoming Windows 11 24H2 update. In the description, Automatic Super Resolution will utilize AI to make supported games play more smoothly with enhanced details.

WCCFTECH.COM

Microsoft is preparing its own AI-powered Super Resolution technology which will be available in Windows 11 24H2.

 

Edited by UltraMega
Made title shorter
  • Thanks 1

Owned

 Share

MOTHERBOARD: MSI MPG Z790i EDGE
CPU: Intel 13900k + Top Mounted 280mm Aio
RAM: 2x24gb Gskill 6400 cl36-48-48 1.4v
PSU: Cooler Master V850 SFX Gold White Edition
GPU: UHD ULTRA EXTREME BANANA GRAPHIC
MONITOR: [Monitor] LG CX48 OLED [VR] Samsung HMD Odyssey Plus OLED + Meta Quest 2 120hz
CASE: CoolerMaster NR200P White Mini ITX
SSD/NVME: 2TB Intel 660p 1tb sn850 1tb sn770
Full Rig Info

Owned

 Share

CPU: Asus Strix G15 AE 6800m 5900hx 32gb ram 1440p
RAM: MSI GT60 Dominator 870m 4800MQ
GPU: Alienware M11x R2 i7 640um Nvidia 335m 8gb Ram
MONITOR: Lenovo X270 1080p i7 7600u 16gb ram
SSD/NVME: Acer Chromebook 11.6
Full Rig Info

Owned

 Share

CPU: Ryzen 5560u
MOTHERBOARD: Beelink SER5 Mini PC Box
RAM: 2x32gb Sodimm
CASE: Jonsbo N1 Mini ITX
HDD: 8TB + 4TB HDD + 2 x Intel DC S3500 800GB
Full Rig Info
Link to comment
Share on other sites

This might be the first real feature of Windows 11 over 10!

 

I generally don't use upscaling or frame generation if I don't need it, but it may be a way to keep older games relevent on new hardware.

 

I'm curious how AI is actually learning this stuff.  If it doesn't know right from wrong, good from bad, how does it know what improvements are desired and what isn't?  It must be an interesting algorithim if they plan to impliment it on thousands of games at once.

 

There's some talk about different processing units in the article, but it's something I didn't understand.  Can anyone explain it better?  The only thing I know is, Nvidia GPUs are good with AI, AMD and Intel, not so much. 

 

I've always wondered what is physically different with AI compute as opposed to rasterization.  The fundamental concept still leads to a 1 or 0, so why is it hardware that's different and not software?  Maybe we'll start to see software/firmware bridge the gap that hardware was doing.

  • Agreed 1
Link to comment
Share on other sites

I was wondering if something like this would take off at this level vs the Nvidia and AMD offerings. If these features become more universal within the OS itself long-term and become just as good or better than what we currently get, that will be interesting to see how DLSS and FSR continue and what they offer on-top of this.

null

Showcase

 Share

CPU: AMD Ryzen 9 5900X
GPU: Nvidia RTX 3080 Ti Founders Edition
RAM: G.Skill Trident Z Neo 32GB DDR4-3600 (@ 3733 CL14)
MOTHERBOARD: ASUS Crosshair VIII Dark Hero
SSD/NVME: x2 Samsung 970 Evo Plus 2TB
SSD/NVME 2: Crucial MX500 1TB
PSU: be Quiet! Straight Power 12 1500W
MONITOR: LG 42" C4 OLED
Full Rig Info

null

Owned

 Share

CPU: E8400, i5-650, i7-870, i7-960, i5-2400, i7-4790k, i9-10900k, i3-13100, i9-13900ks
GPU: many
RAM: Corsair 32GB DDR3-2400 | Oloy Blade 16GB DDR4-3600 | Crucial 16GB DDR5-5600
MOTHERBOARD: ASUS P7P55 WS SC | ASUS Z97 Deluxe | EVGA Z490 Dark | EVGA Z790 Dark Kingpin
SSD/NVME: Samsung 870 Evo 1TB | Inland 1TB Gen 4
PSU: Seasonic Focus GX 1000W
CASE: Cooler Master MasterFrame 700 - bench mode
OPERATING SYSTEM: Windows 10 LTSC
Full Rig Info

Owned

 Share

CPU: M1 Pro
RAM: 32GB
SSD/NVME: 1TB
OPERATING SYSTEM: MacOS Sonoma
CASE: Space Grey
Full Rig Info
Link to comment
Share on other sites

  • UltraMega changed the title to Windows 11 24H2: AI Super Resolution Technology, Works Across All PCs With AI NPUs

This is an interesting concept...

 

Having an upscaler that can run on any NPU is interesting because as far as I know an NPU doesn't have to be part of a GPU or CPU, it can just be a PCIe m.2 card... which technically means one could be added to a PS5 or Xbox Series console via the storage expansion slot. It also means other weird mis-matches would be possible. I guess in theory someone with an older PC that doesn't support an upscaling options could throw an NPU into a PCIe slot and use Microsoft's upscaling. It also means any modern CPU with an NPU will be able to get a lot more mileage out of the integrated graphics. 

 

 

I asked CoPilot to make a list of NPU capable hardware:

Spoiler

As of now, several consumer-level hardware components incorporate NPUs (Neural Processing Units) to enhance AI-related tasks. Here's a list of some notable ones:


- Intel Core Ultra Processors:
    - The Intel Core Ultra processors (formerly codenamed Meteor Lake) feature built-in NPUs.
    - These NPUs are widely utilized by various computer companies in their latest Ultrabook and Notebook laptops.


- AMD Ryzen 8040-Series Laptop Processors:
    - The AMD Ryzen 8040-series laptop processors also integrate NPUs.
    - These processors enhance AI capabilities and contribute to improved performance in AI workloads.


- Qualcomm Windows PCs:
    - While not as common, Qualcomm's Windows PCs were among the first to include NPUs.
    - Arm processors used in most smartphones have included machine-learning acceleration for several years, and this technology has extended to some Windows laptops.


- Apple M-Series Chips for Macs:
    - Apple's M-series chips for Macs (including the M1 and later versions) incorporate NPUs.
    - These chips have been designed to accelerate AI tasks, including neural network inference.


- Future Trends:
    - A recent report predicts that 60% of PCs sold will be AI-capable by 2027.
    - As technology evolves, we can expect more consumer-level hardware to integrate NPUs, further enhancing AI experiences.

 

 

Ironically enough, CoPilot forgot to mention Nvidia GPUs which have NPUs via the tensor cores. AMD 7000 series GPUs do not have an NPU, which means they might not be able to run Microsoft's upscaler. When I looked this up, apparently AMD GPUs should be supported via a fall back method, but it might not have the same quality level as an NPU based method. 

Spoiler

Microsoft Super Resolution (MSR), also known as DirectML Super Resolution, is an AI-based upscaling technique developed by Microsoft. It aims to enhance the resolution and visual quality of images and videos. However, as of my last knowledge update (which was current until 2021), MSR primarily relies on DirectML, which is a framework for machine learning acceleration on Windows devices.
Here's the situation regarding MSR and AMD 7000 series GPUs:
- DirectML Compatibility:
    - DirectML is compatible with a wide range of hardware, including both NVIDIA and AMD GPUs.
    - However, the specific capabilities and performance of DirectML on different GPUs can vary based on their architecture, driver support, and other factors.
- AMD 7000 Series GPUs:
    - The AMD Radeon RX 7000 series (such as the RX 7700 XT, RX 7600 XT, etc.) are based on the RDNA 3 architecture.
    - While these GPUs are powerful for gaming and general graphics tasks, they do not inherently include dedicated NPUs (Neural Processing Units) like some other AMD products.
    - MSR primarily relies on the DirectML framework, which leverages the GPU's compute capabilities for AI tasks.
- MSR and AMD 7000 Series GPUs:
    - MSR should technically run on AMD 7000 series GPUs, but the performance and efficiency may not be optimal compared to GPUs specifically designed for AI workloads.
    - For the best experience with MSR, consider using GPUs that have better AI acceleration capabilities, such as those with dedicated NPUs or specialized AI hardware.

 

 

It will definitely be interesting to see how this all plays out and what kind of weird hardware configs it will allow for. As far as I know, there are no NPU only devices available on the consumer market yet, but I think they are coming. I recently saw a video about an NPU device that looked like an NVME drive and went in the m.2 slot.

 

 

Edited by UltraMega
  • Respect 1

null

Owned

 Share

CPU: 5800x
MOTHERBOARD: ASUS TUF Gaming B550-Plus
RAM: 32GB 3600mhz CL16
GPU: 7900XT
SOUNDCARD: Sound Blaster Z 5.1 home theater
MONITOR: 4K 65 inch TV
Full Rig Info
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...

Important Information

This Website may place and access certain Cookies on your computer. ExtremeHW uses Cookies to improve your experience of using the Website and to improve our range of products and services. ExtremeHW has carefully chosen these Cookies and has taken steps to ensure that your privacy is protected and respected at all times. All Cookies used by this Website are used in accordance with current UK and EU Cookie Law. For more information please see our Privacy Policy