Announcement

Collapse
No announcement yet.

18-Way GPU Linux Benchmarks, Including The Radeon RX 460 & RX 470 On Open-Source

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

  • 18-Way GPU Linux Benchmarks, Including The Radeon RX 460 & RX 470 On Open-Source

    Phoronix: 18-Way GPU Linux Benchmarks, Including The Radeon RX 460 & RX 470 On Open-Source

    Yesterday I published early open-source benchmarks of the Radeon RX 470 while today is a full 18-way graphics card comparison including the newly-launched Radeon RX 460 and Radeon RX 470 graphics cards alongside the RX 480 Polaris graphics card. All of the AMD graphics cards tested for this article were running the very latest open-source driver stack on the Linux 4.8 kernel and Mesa 12.1-dev Git.

    http://www.phoronix.com/vr.php?view=23421

  • #2
    Michael

    For RX 460 range cards 4k is useless because this gpus are too weak

    Maybe can add more 1080p tests and some vulkan test



    Comment


    • #3
      Nice. Any chance of getting Unigine Heaven back into the rotation ? AFAIK developer focus is more on higher GL levels these days so Heaven gives a more accurate picture than Valley of where performance is going.

      It wasn't clear what resolution Metro: Last Light Redux was running at. I figured I would just go back to earlier tests and compare performance/resolution to figure it out, but apparently the resolution didn't show in the earlier test results either
      Last edited by bridgman; 08-10-2016, 12:28 PM.

      Comment


      • #4
        There are several typos RX 470 vs 460 in the article.

        Comment


        • #5
          Originally posted by puleglot View Post
          There are several typos RX 470 vs 460 in the article.
          "The RX 460 testing is from the factory-overclocked SAPPHIRE NITRO+ Radeon RX 470."

          "The R9 290 was back into its regressed state with Tesseract and the RX 470 continued performing around the same speed as the R7 260X."

          Comment


          • #6
            Originally posted by bridgman View Post
            Nice. Any chance of getting Unigine Heaven back into the rotation ? AFAIK developer focus is more on higher GL levels these days so Heaven gives a more accurate picture than Valley of where performance is going.

            It wasn't clear what resolution Metro: Last Light Redux was running at. I figured I would just go back to earlier tests and compare performance/resolution to figure it out, but apparently the resolution didn't show in the earlier test results either
            Sure I can add Heaven back to the queue. Basically was just doing Valley since it was newer but I have no problems running both.

            Metro is always run at whatever resolution the reported system resolution was, in this case 3840 x 2160. Metro Last Light / 2033 Redux don't offer any CLI controls for resolution handling so basically stuck running it at whatever is the native resolution.
            Michael Larabel
            http://www.michaellarabel.com/

            Comment


            • #7
              Maybe I have that backwards then... I thought Heaven was newer. Hold on, will check...

              Nope, looks like you are right and Heaven did come out first. I guess I just thought Heaven was newer because it seems to use more advanced graphics tech, but maybe it's just tesselation in Heaven vs. more detailed textures in Valley.

              So yeah, I guess running both would be best and we'll try to figure out why Heaven runs faster than Valley. Thanks !
              Last edited by bridgman; 08-10-2016, 01:31 PM.

              Comment


              • #8
                Thanks for the article.

                Firstly I have to reiterate that whilst it's interesting to see just how far AMD's open source efforts have come, and that in many cases they are competitive with proprietary drivers, I think it's disingenuous to compare open source AMD to closed source NVIDIA. And this is doubly so when it comes to performance per Watt, as this is somewhere the open source drivers are behind the proprietary drivers. I understand it's because of the custom 16.10 system you have to downgrade.

                Secondly (and I can see that Metro has issues - in the future maybe you could use the AMD virtualised framebuffer that you wrote about recently to get 1080p results here) 4K or even 1440p on the 460 is simply not what the card is aimed at, and you cannot assume the result scales sensibly down from 4K. It's great that the card can do 4K in some low-end Linux games, but in the end the card is being pushed as a 1080p Medium card, which for $109 is understandable. I look forward to the next article containing far more suitable resolutions.

                In addition, would it be possible to split the graphs, e.g., have the 460/950/960 and similar level cards have their own graph (at 1080p Medium) and have the 470/480/970/980/1060 have their own 1080p/1440p Ultra graph? I understand it's extra work, but it would mean that results were segregated better by product class.

                "At first it looked the RX 460 could be competitive with the GTX 950 under Linux, but that ended up being isolated to just a few of the OpenGL Linux games, while in the end it tended to trail behind in raw performance for the generation-old GTX 950."

                Yeah, but you should emphasise again that this was pitting open source versus proprietary. Now there's a strong chance that AMD's proprietary results won't be that much better (or may be worse in select situations) but at least you'd have some evidence to back up the statement.

                There's also the worry that in a month or three, the 1050 will arrive, without firmware blobs - at leat the 950 can run open source drivers. Have you any idea when Nvidia will release the Pascal blobs for Nouveau?

                And hopefully some more Linux games will get auto-benchmarking support built-in.

                Comment


                • #9
                  Well, I did compare against on of RX 480 benchmarks: http://openbenchmarking.org/result/1...KH32&obr_sor=y and it turns out that the benchmarks can't be compared against different PCs (well quite similar, i5 6600 non-K, 8 GB RAM, SSD). And the 4K resolution benchmarks are rather incomparable when not having a 4K display.

                  - In Tesseract R9 270 is better than R9 Fury.
                  - In Unigine Heaven on fglrx (Ubuntu 15.10) it's bit below R9 270x, but after upgrade to 16.04 it looses performance (doesn't seems to use amdgpu).
                  - Xonotic on 16.04 is better than RX 480 Pro, where as on 15.10 fglrx is below R9 290
                  - In GPUTest triangle it's better than GTX 1080.... can you skip 4K resolutions as likely on 1080p display those benchmarks don't work properly

                  BTW I never was able to run any Steam benchmark. Aside of that it would be good to mark which Steam benchmark can be run without buying anything.

                  Oh well, Windows benchmarks should be better and more meaningful as few games for which I need a dedicated GPU work on Windows only currently (like Witcher 3, Skyrim, Fallout 4...)

                  Comment


                  • #10
                    4K results are only interesting for the GTX 1080 and maybe 1070, it makes no sense to test everything in 4K. At most 2% Steam users have 4K monitors according to the Steam hardware survey.

                    Comment

                    Working...
                    X