Announcement

Collapse
No announcement yet.

AMD Radeon HD 7970 On Linux

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

  • AMD Radeon HD 7970 On Linux

    Phoronix: AMD Radeon HD 7970 On Linux

    You've may have heard or seen that AMD introduced the Radeon HD 7970 graphics card this morning as their first product built in their "Southern Islands" family and is based on their new GPU architecture, but how well does it work under Linux for the open-source and closed-source AMD Catalyst Linux drivers?..

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

  • #2
    Graphic card information and programming are revealed six months prior to launch, according to anandtech. Does than mean, that we could already be having working opensource driver? Why it is not the fact?

    Comment


    • #3
      Anyone who's interested about this new architecture should check out the Anandtech article (not the benchmark one, the one previous to that). The article suggests that this is really the biggest architectural change at AMD since the R100 with the move away from VLIW to a more compute friendly design.
      Now, that is interesting for a few reasons.
      One, the OpenCL LLVM based compiler they released isn't too useful with the new design (but with the new design it seems as though the compiler becomes less important anyhow, so the compiler release was still a good thing).
      Two, and for me, the BIG ONE: hybrid video decode/encode. Apparently they've "copied" intel's ASIC for video decode/encode (though, it seems as though that was what UVD always was, but maybe UVD had more to do DRM?), but they're including a "hybrid" mode which offloads most of the work to the compute units (CU) and they leave some heavily serialised work to the UVD (again, I'm assuming DRM is included in that work). Thus is seems we might finally have some sort of video parity on the open source side with this new architecture (assuming, of course, they provide some help to driver developers to get the video decode/encode into a similar state as catalyst -- right now I can't imagine why they wouldn't as this doesn't seem like particularly valuable knowledge (ignoring any DRM/UVD issues), but perhaps Bridgeman can enlighten us?).

      Comment


      • #4
        Originally posted by crazycheese View Post
        Graphic card information and programming are revealed six months prior to launch, according to anandtech. Does than mean, that we could already be having working opensource driver? Why it is not the fact?
        From the anandtech article I would imagine the problem is just how different this new architecture is from everything before (on the AMD side at least). For at least this release, we need to have some patience.

        Comment


        • #5
          Originally posted by liam View Post
          From the anandtech article I would imagine the problem is just how different this new architecture is from everything before (on the AMD side at least). For at least this release, we need to have some patience.
          I donīt need. But if there would be no need for wait, I would be having amd instead.
          From hardware point of view, it looks like quite innovative(in terms of change from vliw), as well as polished on time arch; unlike bulldozer though, which was shame.

          If AMD takes their zero-core technology to APU and CPU, especially with bulldozer, that would be impressive addition.

          Comment


          • #6
            Originally posted by crazycheese View Post
            Graphic card information and programming are revealed six months prior to launch, according to anandtech. Does than mean, that we could already be having working opensource driver? Why it is not the fact?
            I've never heard or seen that statement (internally or externally), and the only search engine hit I could find was this thread. Can you post a link to the statement ?

            Comment


            • #7
              Originally posted by liam View Post
              One, the OpenCL LLVM based compiler they released isn't too useful with the new design
              Actually quite the opposite.. the plan is for the new compiler code to go into the Gallium3D driver and process GCN graphics shaders (after adding GCN instruction generation, of course) along with compute for all architectures, while pre-GCN graphics shaders will go through the existing r600g compiler/translator.

              Originally posted by liam View Post
              this doesn't seem like particularly valuable knowledge (ignoring any DRM/UVD issues), but perhaps Bridgeman can enlighten us?).
              Agree 100%. The problem is that you can't ignore the DRM/UVD issues.

              Comment


              • #8
                Originally posted by bridgman View Post
                I've never heard or seen that statement (internally or externally), and the only search engine hit I could find was this thread. Can you post a link to the statement ?
                Okay, this already means either I misunderstood anandtech, or anandtech misunderstood amd. Here, grep for "six".

                Comment


                • #9
                  I read most of that article but didn't see the reference, and searching for "six" on the first page didn't give any matches. Was it supposed to be on the first page, or one of the later ones ?

                  There are a lot of references to leaks of product information going back six months or so, but that's all I could find.

                  Comment


                  • #10
                    The 6 month reference

                    @Bridgman and others looking for the 6 month cycle reference...I believe it's here:

                    http://www.anandtech.com/show/5261/a...-7970-review/2

                    "As a result both NVIDIA and AMD have begun revealing their architectures to developers roughly six months before the first products launch. This is very similar to how CPU launches are handled, where the basic principles of an architecture are publically disclosed months in advance".

                    (the second page of the Anandtech review)
                    ---

                    Comment


                    • #11
                      Originally posted by bridgman View Post
                      I read most of that article but didn't see the reference, and searching for "six" on the first page didn't give any matches. Was it supposed to be on the first page, or one of the later ones ?

                      There are a lot of references to leaks of product information going back six months or so, but that's all I could find.
                      Oh, sorry, didnīt I took openbenchmarking wrapper all-long. Its this one.

                      Comment


                      • #12
                        Ahh, that part. Sorry, I did see that but kind of "filtered it out" in my head because I was looking for open source references.

                        That text is referring to events like AFDS and others, where we expose architectural information to give application and toolchain developers a heads-up. The information we provide at that point doesn't include anything like the level of detail required for driver development, however.

                        Thanks !

                        Comment


                        • #13
                          May you say some details about UVD4 and PowerTune support in open source driver now? Or there is still embargo for such information?

                          Comment


                          • #14
                            I haven't looked much at UVD in the latest GPUs but I thought they were still UVD 3. Can't really talk much about UVD at this point because I don't know what we are going to be able to release, although we are going to take another look in the new year.

                            I don't think we have looked at PowerTune specifically; my guess is that the first chance to do that will be a few months from now. There are some more fundamental power management improvements I would like to get out first, and they will probably be a pre-requisite for PowerTune anyways.

                            Comment


                            • #15
                              Originally posted by bridgman View Post
                              Actually quite the opposite.. the plan is for the new compiler code to go into the Gallium3D driver and process GCN graphics shaders (after adding GCN instruction generation, of course) along with compute for all architectures, while pre-GCN graphics shaders will go through the existing r600g compiler/translator.



                              Agree 100%. The problem is that you can't ignore the DRM/UVD issues.

                              Thanks for responding.
                              I'm not sure I follow you. It SEEMS as though you are saying you will use the open sourced compiler for the old r600 while the new diver will switch to the new one. I apologize if I've misunderstood.
                              Lastly regarding drm I was attempting to give the impression that since hybrid mode will do most of the work through the CUs we could ignore the small bit that the uvd handles and thus not use uvd at all thus ignoring drm

                              Happy holidays!

                              Liam

                              Comment

                              Working...
                              X