Image: Intel

A leaked Intel document has revealed that there will be over 50 Arc desktop and mobile GPU designs from manufacturers that include Acer, ASUS, Dell, GIGABYTE, HP, Lenovo, Samsung, and MSI.

The document also includes a list of partners that will support its AI-driven upscaling technology, XeSS, the third following NVIDIA’s DLSS and AMD’s FidelityFX Super Resolution. XeSS, which is expected to also work on GeForce and Radeon cards, had been demoed in a video for EXOR Studios’ The Riftbreaker and some images from Hitman 3.

Famous studios such as 505 Games, Codemasters, Ubisoft, Fishlabs, and Kojima Productions will adopt XeSS. Death Stranding Directors Cut is mentioned in the release.

The document also discusses Intel Deep Link, a technology that aims for increased performance when the GPUs are combined with an Intel CPU. Blackmagic Design’s DaVinci Resolve will be one of the first programs to support Deep Link Hyper Encode, enabling simultaneous encoding of video streams.

Image: Intel (via VideoCardz)

Intel had announced that its Arc Alchemist graphics cards will be available in Q1 2022.

Source: VideoCardz

Don’t Miss Out on More FPS Review Content!

Our weekly newsletter includes a recap of our reviews and a run down of the most popular tech news that we published.

Peter Brosdahl

As a child of the 70’s I was part of the many who became enthralled by the video arcade invasion of the 1980’s. Saving money from various odd jobs I purchased my first computer from a friend of my...

Join the Conversation

16 Comments

  1. Wow, that’s quite a large number of SKU’s.

    Sometimes it is better to keep the product lineup simple. Especially when you are just starting out with something new…

  2. [QUOTE=”Dogsofjune, post: 46191, member: 168″]
    I wonder if any of them will offer decent hardware acceleration for Plex and such…
    [/QUOTE]
    I’m just hoping that they don’t limit their latest transcoders to higher-end specs like Nvidia regularly does!

  3. [QUOTE=”Zarathustra, post: 46186, member: 203″]
    Sometimes it is better to keep the product lineup simple.
    [/QUOTE]
    Definitely not a part Intel’s strategy these days. Although NVIDIA isn’t much better. Sure their lineup will be under 50 but they’ve established a pattern of getting a bunch out ever since Maxwell days. Google the Wiki for the Maxwell releases and you’ll come across a chart that’s just plane insanity.

  4. [QUOTE=”Peter_Brosdahl, post: 46217, member: 87″]
    Definitely not a part Intel’s strategy these days. Although NVIDIA isn’t much better. Sure their lineup will be under 50 but they’ve established a pattern of getting a bunch out ever since Maxwell days. Google the Wiki for the Maxwell releases and you’ll come across a chart that’s just plane insanity.
    [/QUOTE]
    Yeah. I can see needing to make some budgetary tiers, and a difference between mobile and desktop (kinda), and even having room for a mid-life refresh of some sort. Maybe even mining and/or Pro and/or datacenter variants. But even with all that, I still find it hard to fathom 50 SKUs, and Intel isn’t attacking all of those markets with Xe/Alchemist/whatever name it has today.

    I think [USER=215]@Grimlakin[/USER] may be on to something: this might be Intel hedging their manufacturing. These cards aren’t even on an internal fab – these are being made at TSMC, but there’s still a lot that goes into a wafer apart from just the process that determines yields/binning/etc. And Intel hasn’t been a TSMC customer for terribly long to build up a rapport.

    This convoluted and it just confuses everyone. And we still haven’t seen a real physical card run anything, and all the leaked info to date keeps talking about the lower tier stuff… I really want this to work, but in my bones I just don’t see this thing doing what Intel wants it to do, nevermind what I want it to do. I think what Intel isn’t saying (or showing) is speaking a lot louder than what they are – especially if they are still trying to target a Q1’22 release.

    Holy crap, we are there already! Sure, there’s still a few weeks to be inside that window, but… the clock is ticking.

  5. [QUOTE=”Peter_Brosdahl, post: 46217, member: 87″]
    Google the Wiki for the Maxwell releases
    [/QUOTE]
    Hmm just out of curiosity:

    [QUOTE]
    The first Maxwell-based products were the GeForce GTX 750 and the GeForce GTX 750 Ti. Both were released on February 18, 2014, both with the chip code number GM107. Earlier GeForce 700 series GPUs had used [URL=’https://en.wikipedia.org/wiki/Kepler_(microarchitecture)’]Kepler[/URL] chips with the code numbers G[I]K[/I]1xx. First-generation Maxwell GPUs (code numbers GM10x) are also used in the GeForce 800M series and the Quadro Kxxx series. A second generation of Maxwell-based products was introduced on September 18, 2014 with the GeForce GTX 970 and GeForce GTX 980, followed by the GeForce GTX 960 on January 22, 2015, the GeForce GTX Titan X on March 17, 2015, and the GeForce GTX 980 Ti on June 1, 2015. These GPUs have GM[I]2[/I]0x chip code numbers.
    [/QUOTE]

    By my count, I see 5 different dies that were fabbed (GM107, 108, 200, 204, and 206). That doesn’t sound so bad.

    From those 5 dies, I see
    4 700-series SKUs (one was a minor refresh from a 100-series to 200-series die)
    5 800M series
    9 900 series
    8 Quadro series (might have missed some here, I just did a quick count)
    and I won’t count the Jetson/Tegra stuff since that’s APU-level, and doesn’t appear to have been much of it.

    So that’s… 26? Now, each of those SKUs likely had various models from OEMs to support various trim levels and overclocks, which would make the list dizzying… but that’s different than the manufacturer making the die, which is what Intel is talking about doing with their 50+ SKUs on Alchemist.

    Yeah, a long cry from 50.

  6. [QUOTE=”Brian_B, post: 46229, member: 96″]
    Hmm just out of curiosity:

    By my count, I see 5 different dies that were fabbed (GM107, 108, 200, 204, and 206). That doesn’t sound so bad.

    From those 5 dies, I see
    4 700-series SKUs (one was a minor refresh from a 100-series to 200-series die)
    5 800M series
    9 900 series
    8 Quadro series (might have missed some here, I just did a quick count)
    and I won’t count the Jetson/Tegra stuff since that’s APU-level, and doesn’t appear to have been much of it.

    So that’s… 26? Now, each of those SKUs likely had various models from OEMs to support various trim levels and overclocks, which would make the list dizzying… but that’s different than the manufacturer making the die, which is what Intel is talking about doing with their 50+ SKUs on Alchemist.

    Yeah, a long cry from 50.
    [/QUOTE]
    My mistake, I mixed up the name. I meant Pascal. I remembered it was the 10 series but screwed up the name. Here you go. Here’s a niftly link of charts for ongoing NVIDIA releases too. It even includes the mobile variants.
    [URL unfurl=”true”]https://en.wikipedia.org/wiki/List_of_Nvidia_graphics_processing_units#GeForce_10_series_2[/URL]
    [ATTACH type=”full”]1394[/ATTACH]

  7. [QUOTE=”Brian_B, post: 46229, member: 96″]
    Hmm just out of curiosity:

    By my count, I see 5 different dies that were fabbed (GM107, 108, 200, 204, and 206). That doesn’t sound so bad.

    From those 5 dies, I see
    4 700-series SKUs (one was a minor refresh from a 100-series to 200-series die)
    5 800M series
    9 900 series
    8 Quadro series (might have missed some here, I just did a quick count)
    and I won’t count the Jetson/Tegra stuff since that’s APU-level, and doesn’t appear to have been much of it.

    So that’s… 26? Now, each of those SKUs likely had various models from OEMs to support various trim levels and overclocks, which would make the list dizzying… but that’s different than the manufacturer making the die, which is what Intel is talking about doing with their 50+ SKUs on Alchemist.

    Yeah, a long cry from 50.
    [/QUOTE]

    51 according to this [URL]https://www.techpowerup.com/gpu-specs/?architecture=Maxwell&sort=generation[/URL]

  8. [QUOTE=”Denpepe, post: 46235, member: 284″]
    51 according to this [URL]https://www.techpowerup.com/gpu-specs/?architecture=Maxwell&sort=generation[/URL]
    [/QUOTE]
    My new-old catchphrase for the year may end up becoming “But wait! There’s more!”

  9. [QUOTE=”Denpepe, post: 46235, member: 284″]
    51 according to this [URL]https://www.techpowerup.com/gpu-specs/?architecture=Maxwell&sort=generation[/URL]
    [/QUOTE]
    Yeah, looks like a lot of A varients and Quadro Mobile that I missed on my count

  10. [QUOTE=”Brian_B, post: 46248, member: 96″]
    Yeah, looks like a lot of A varients and Quadro Mobile that I missed on my count
    [/QUOTE]
    Yep, there was a whole other mobile chart I didn’t include but also came up on my searches. The desktop ones just got on my nerves after a while. It just seemed like every other month NV was releasing a new variant for 2-3 years. Back then things were on the shelves most of the time(until right before Turing) so I always felt sorry for any consumer trying to navigate that nightmare without understanding it all, especially with all the hype-marketing gibberish that goes on every box.

  11. [URL unfurl=”true”]https://www.tomshardware.com/news/intel-arc-gpu-launch-window-changed-on-website[/URL]

    Imagine that.

  12. Thanks! If Tsing doesn’t post it, I’ll prep it in the morning. Gotta admit I haven’t seen anyone else notice this yet on other sites yet but who knows by tomorrow morning.

  13. [QUOTE=”Burticus, post: 46329, member: 297″]
    Throw enough food at the wall and something has to stick, right?
    [/QUOTE]
    [B]let them eat cake[/B]

Leave a comment