Announcement

Collapse
No announcement yet.

Intel Formally Announces Iris Xe MAX Graphics, Deep Link

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Intel Formally Announces Iris Xe MAX Graphics, Deep Link

    Phoronix: Intel Formally Announces Iris Xe MAX Graphics, Deep Link

    Laptop vendors recently disclosed "Xe MAX" graphics as discrete Intel graphics set to appear within laptops in the coming weeks. That announcement was a bit unexpected and Intel did not brief the media in advance while today -- in an unusual announcement for a Saturday (Intel says it's timed for system availability, seemingly first in China) -- the company is formally announcing Iris Xe MAX.

    http://www.phoronix.com/vr.php?view=29666

  • #2
    Would be nice to have one of these as a little standalone PCI-E card with passive cooling and SR-IOV support.
    Plop it in and have AV1 dec, fully working Wayland and all that as a companion to nvidia.

    Comment


    • #3
      It is interesting that they are going hard at video encode. That is some thing that is important to me but is still relatively a niche need over all. I will be curious to see how this works in the real world. They are comparing this to 2080 rather than 3080 but neither NVidia or AMD have really talked up their hardware encoding yet on the new generation boards so we don't know if this is just carried over from the previous generation or if they have brought any improvements in IQ or encode speed. More competition is good but I think they are going to get steam rolled by AMD.

      Comment


      • #4
        So Deep Link is more or less AMD Crossfire or at least the fabled version of Crossfire that was supposed to tie together an AMD APU's iGPU to an AMD dGPU.

        Interesting stuff...but this is a classic Intel paper launch which is why there was no heads up. It's Intel's Marketing Dept going " HEY..HEY...LOOK...WE'RE STILL INOVATING. WE'RE STILL RELEVANT !! < what...no...we're not commenting on 10nm much less our roadmap to 7nm. no..we're not commenting either on having to sell off our entire NAND division...nor our modem division...nor our mobile CPU division...nor our Edison project. and no...launching GPUs designed by half of AMDs graphic engineering department doesn't mean Larrabee/Knights Landing was a abysmal failure. yes...we did shut down that department as well...but...>
        Last edited by Jumbotron; 31 October 2020, 09:48 PM.

        Comment


        • #5
          Originally posted by Jumbotron View Post
          So Deep Link is more or less AMD Crossfire or at least the fabled version of Crossfire that was supposed to tie together an AMD APU's iGPU to an AMD dGPU.

          Interesting stuff...but this is a classic Intel paper launch which is why there was no heads up. It's Intel's Marketing Dept going " HEY..HEY...LOOK...WE'RE STILL INOVATING. WE'RE STILL RELEVANT !! < what...no...we're not commenting on 10nm much less our roadmap to 7nm. no..we're not commenting either on having to sell off our entire NAND division...nor our modem division...nor our mobile CPU division...nor our Edison project. and no...launching GPUs designed by half of AMDs graphic engineering department doesn't mean Larrabee/Knights Landing was a abysmal failure. yes...we did shut down that department as well...but...
          They may not have 7nm yet, but at least they have 4:4:4 encoding and a stable open-source driver.

          Comment


          • #6
            Originally posted by tildearrow View Post

            They may not have 7nm yet, but at least they have 4:4:4 encoding and a stable open-source driver.
            Exactly, at this point:
            Intel has good OEMs relationship, the best open source drivers and 'meh' hardware (for the price);
            NVIDIA has good OEMs relationship, shit drivers (open and closed) and good hardware (but not for the price);
            AMD doesn't have good OEMs relationship (AMD CPU+GPU laptops are still shit at this point), but has acceptable drivers and good hardware (for the price);

            In the long run [for Linux], NVIDIA is a no-go (except if you really need CUDA), Intel can have a good place if they release mid-range GPUs and AMD can be the king if they improve their drivers...

            Comment


            • #7
              Originally posted by Jumbotron View Post
              So Deep Link is more or less AMD Crossfire or at least the fabled version of Crossfire that was supposed to tie together an AMD APU's iGPU to an AMD dGPU.
              It allows you to use them both for some compute or media conversion workloads. For graphics they use the software stack to determine whether a game should run on the iGPU or the dGPU (because the iGPU in Tiger Lake is faster at some games). Sounds like it may function similar to Optimus under the covers. Anandtech has a deeper dive on the distinctions.

              Comment


              • #8
                Originally posted by tildearrow View Post

                They may not have 7nm yet, but at least they have 4:4:4 encoding and a stable open-source driver.
                I think on linux their driver support and a cheap price point could be a winning combination, particularly for people buying AMD cpus that don't have an IGP on them and don't need a high end graphics card performance. There's apparently going to be a desktop version of DG1 for sale, although it's unclear whether it will be easy for users to buy themselves or just OEM.

                It's a lot tougher to make that argument in the Windows market since they don't have an advantage in their driver support there.
                Last edited by smitty3268; 31 October 2020, 05:51 PM.

                Comment


                • #9
                  Originally posted by nxij View Post
                  Would be nice to have one of these as a little standalone PCI-E card with passive cooling and SR-IOV support.
                  Plop it in and have AV1 dec, fully working Wayland and all that as a companion to nvidia.
                  Hopefully vGPU or whatever the GVT name for it is. Intel is the only GPU vendor that supports that kind of virtualization setup where you can split the resources in a flexible manner across VMs and not require SR-IOV (which I think differs still in comparison?).

                  Comment


                  • #10
                    I can't fault Intel for going after this low end market, even though I wish the company ill in general. Nvidia's offerings are thin on the low end, and AMD's seem to be thinner, and OEM only.

                    It does seem that such laptops should ideally use 'KF' version CPUs with disabled integrated graphics, unless Intel can harness both at once. I'm not clear if 'we can use GPU and CPU on the same tasks' also means 'we can use GPU and CPU and CPU-integrated-GPU all on the same task.' The latter sounds more like a mismatched Multi-GPU issue of the kind that Vulkan allows solving, but no one does outside that Ashes of the Singularity example.

                    Intel marketing are making a bad situation worse with the way they are acting. Overhyping in some areas, trying to hide things in others. But as I said I don't wish them well, so... keep up the bad work, boys?

                    Comment

                    Working...
                    X