Announcement

Collapse
No announcement yet.

AMD Radeon R9 290 On Ubuntu 14.04 With Catalyst Can Beat Windows 8.1

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

  • AMD Radeon R9 290 On Ubuntu 14.04 With Catalyst Can Beat Windows 8.1

    Phoronix: AMD Radeon R9 290 On Ubuntu 14.04 With Catalyst Can Beat Windows 8.1

    The latest Linux graphics testing under the microscope at Phoronix is comparing the Ubuntu 14.04 LTS vs. Windows 8.1 performance with all available updates. Results from Intel, NVIDIA, and AMD hardware is coming up next week while today is a bit of a preview of the AMD numbers when using a Radeon R9 290 "Hawaii" graphics card. While the open-source AMD Hawaii support remains broken, with the Catalyst 14.4 driver on each operating system, the Linux Catalyst driver with the R9 290 graphics card can outperform Windows 8.1 Pro with some OpenGL games and benchmarks.

    http://www.phoronix.com/vr.php?view=20418

  • #2
    Hmm this is very interesting. I think this is the first time the linux catalyst drivers were doing better in every single test (though, there weren't that many tests). I might switch back to catalyst if AMD will look into wayland support.

    Comment


    • #3
      Probably they're shifting focus on OpenGL to Linux instead of focusing Direct3D and OpenGL on Windows and let OpenGL Linux starve. I think it should be the same as the codebase is supposedly shared.

      Comment


      • #4
        I would like to see a test of some source games to compare as well.

        Comment


        • #5
          Originally posted by Filiprino View Post
          Probably they're shifting focus on OpenGL to Linux instead of focusing Direct3D and OpenGL on Windows and let OpenGL Linux starve. I think it should be the same as the codebase is supposedly shared.
          Bridgeman stated something else, you are probably mistaken nvidia's share code over both platforms, amd's is not largely shared.

          Comment


          • #6
            Originally posted by tmpdir View Post
            Bridgeman stated something else, you are probably mistaken nvidia's share code over both platforms, amd's is not largely shared.
            Probably "less shared" is more correct than "largely not shared". Most of the OpenGL and OpenCL code is shared, however the parts which talk to the kernel driver (think libdrm/winsys in the open stack) are different.

            Comment


            • #7
              Won't Linux beat Windows every time (if all things were equal) because of WDDM?(since Vista) Shifting most of the kernel space drivers in to the user space to keep the video card from crashing the system

              Comment


              • #8
                Originally posted by vaudevillian View Post
                I would like to see a test of some source games to compare as well.
                One site says that Gallium3d+PTE+3.15+hyperz is kicking Catalysts anus in Steam games. My experience with Catalyst 14.4 has been slower then open source, in both Portal 2 and Wine. Catalyst for me has been stutter in fps and lower fps overall. This is on my Radeon HD 6370M. Though Catalyst for me does a better job at keeping my laptop cool.

                Comment


                • #9
                  Originally posted by Dukenukemx View Post
                  One site says that Gallium3d+PTE+3.15+hyperz is kicking Catalysts anus in Steam games.
                  If we tend to believe one Gentoo user, we are all badly wrong . That is severely young all git code, i expect "unexpected" there bugs too . But also keep in mind that is on Gentoo, so those parts of the graphics stack is heavely optimized for that one machine and that also can break somewhere . That is like overcloking, i can overclock my APU and beat Catalyst, and he can build his kernel with some patches which might also break Catalyst performance but improve radeon, so we can't really know that from one Gentoo user experiance .

                  http://www.linuxsystems.it/2014/05/r...t-steam-games/

                  And that needs to be tested and approved by other people, so there is no reason for everage Joe to be so happy right now .
                  Last edited by dungeon; 05-18-2014, 03:06 PM.

                  Comment


                  • #10
                    Originally posted by bridgman View Post
                    Probably "less shared" is more correct than "largely not shared". Most of the OpenGL and OpenCL code is shared, however the parts which talk to the kernel driver (think libdrm/winsys in the open stack) are different.
                    I said "not largely shared", "Largely not shared" feels like an completly different statement. The first could mean 65% is shared while the second probably means 35% is shared.

                    But I stand corrected, makes me look different at opengl test results and changes.

                    Comment


                    • #11
                      Originally posted by tmpdir View Post
                      I said "not largely shared", "Largely not shared" feels like an completly different statement. The first could mean 65% is shared while the second probably means 35% is shared.

                      But I stand corrected, makes me look different at opengl test results and changes.
                      Whoops, you're right. Didn't mean to reverse the words. That said, I don't like either phrase

                      I would translate "not largely shared" to something less than 50% (say 40-45) and "largely not shared" to something much less, maybe 20-25%. AFAIK the code sharing is still over 50%, although certainly lower than it used to be before MS moved the graphics memory manager and related code from the driver to the OS starting with Vista.

                      Comment


                      • #12
                        I have always wondered why it seemed that the free drivers were much closer in performance in FOSS games compared to commercial ones. It's starting to dawn on me that the proprietary drivers probably have tons of hacks and replacement-shaders for any commercial gaming/benchmark engine to artificially boost their FPS results. That is probably also part of their "secret sauce" that they absolutely cannot give up (ie. through open sourcing).

                        Comment


                        • #13
                          Originally posted by Ancurio View Post
                          I have always wondered why it seemed that the free drivers were much closer in performance in FOSS games compared to commercial ones. It's starting to dawn on me that the proprietary drivers probably have tons of hacks and replacement-shaders for any commercial gaming/benchmark engine to artificially boost their FPS results. That is probably also part of their "secret sauce" that they absolutely cannot give up (ie. through open sourcing).
                          Sir you forgot your tinfoil hat.

                          Comment


                          • #14
                            Originally posted by grndzro View Post
                            Sir you forgot your tinfoil hat.
                            Originally posted by http://richg42.blogspot.de/2014/05/the-truth-on-opengl-driver-quality.html
                            Historically, this vendor (Nvidia) will do things like internally replace entire shaders for key titles to make them perform better (sometimes much better). Most drivers probably do stuff like this occasionally, but this vendor will stop at nothing for performance. What does this mean to the PC game industry or graphics devs? It means you, as "Joe Graphics Developer", have little chance of achieving the same technical feats in your title (even if you use the exact same algorithms!) because you don't have an embedded vendor driver engineer working specifically on your title making sure the driver does exactly the right thing (using low-level optimized shaders) when your specific game or engine is running. It also means that, historically, some of the PC graphics legends you know about aren't quite as smart or capable as history paints them to be, because they had a lot of help.
                            This should be common knowledge, actually.

                            Comment


                            • #15
                              Originally posted by Ancurio View Post
                              I have always wondered why it seemed that the free drivers were much closer in performance in FOSS games compared to commercial ones. It's starting to dawn on me that the proprietary drivers probably have tons of hacks and replacement-shaders for any commercial gaming/benchmark engine to artificially boost their FPS results. That is probably also part of their "secret sauce" that they absolutely cannot give up (ie. through open sourcing).
                              Yep you are 100% right, those blob guys always play out of specs - but that is business .

                              Mantle? You can have it. Where? You know. But why? That is business .
                              Last edited by dungeon; 05-18-2014, 09:01 PM.

                              Comment

                              Working...
                              X