Announcement

Collapse
No announcement yet.

Benchmarks Of NVIDIA's New Linux GPU Driver

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

  • Benchmarks Of NVIDIA's New Linux GPU Driver

    Phoronix: Benchmarks Of NVIDIA's New Linux GPU Driver

    For seeing whether NVIDIA's new 319.12 Beta Linux graphics driver that introduces many new features has any impact on the OpenGL performance for many common Linux games, here are some Phoronix test results when comparing the NVIDIA 304.xx, 310.xx, 313.xx, and 319.xx graphics driver series on GeForce GTX 550 Ti "Fermi" and GeForce GTX 680 "Kepler" graphics cards...

    Phoronix, Linux Hardware Reviews, Linux hardware benchmarks, Linux server benchmarks, Linux benchmarking, Desktop Linux, Linux performance, Open Source graphics, Linux How To, Ubuntu benchmarks, Ubuntu hardware, Phoronix Test Suite

  • #2
    This gave me a benchmark idea: how about benchmarking various CPUs' game performance?

    Comment


    • #3
      I realy don't know what you want with those games. The main speed diff can only be seen when a game is used that can use the multitheading features introduced by the 310+ drivers like TF2. A fast card can show an impressive speed boost with that:



      I should have saved the csv file but did not notice it a this day. That would have got all settings stored what i used. Basically AA was off and the rest was on.
      Last edited by Kano; 11 April 2013, 04:03 AM.

      Comment


      • #4
        @Michael

        Will you perform power usage tests for almost-optimus vs intel+nouveau? (Do you have hardware?)

        Comment


        • #5
          Originally posted by Kano View Post
          I realy don't know what you want with those games. The main speed diff can only be seen when a game is used that can use the multitheading features introduced by the 310+ drivers like TF2. A fast card can show an impressive speed boost with that:



          I should have saved the csv file but did not notice it a this day. That would have got all settings stored what i used. Basically AA was off and the rest was on.
          I haven't been following the benchmark articles that closely, has there been any articles comparisons using before/after the multithreading features added in 310.x?

          Comment


          • #6
            Maybe the slower speed is because of this: https://devtalk.nvidia.com/default/t...19-12-drivers/ ?
            In just 30 minutes of use, my syslog file was 120MB big.

            Comment


            • #7
              And TF2, CSS when ?

              please test Team Fortress 2(it's FREE), Killing Floor, CSS and other major Steam Linux Games !

              Comment


              • #8
                Originally posted by Milos_SD View Post
                Maybe the slower speed is because of this: https://devtalk.nvidia.com/default/t...19-12-drivers/ ?
                In just 30 minutes of use, my syslog file was 120MB big.
                I'm sure that would have an impact. But also, this latest driver seemed to be more feature additions than bug fixes and optimizations, so a performance drop is to be expected.

                I would be interested to see the power efficiency effects of Optimus though.

                Comment


                • #9
                  Originally posted by mum1989 View Post
                  please test Team Fortress 2(it's FREE), Killing Floor, CSS and other major Steam Linux Games !
                  Michael already mentioned that he can't benchmark them as they don't return framerate in any way that the PTS could automatically collect and store just yet. So that's something that Valve needs to fix first.

                  Comment


                  • #10
                    You get even more stats in cvs files in the program dir when you run timedemos than pts could collect when you benchmark source engine games. I also benchmarked Killing Floor, with support for sed on Win (or perl) i could parse everything but that game has rendering errors with Nvidia. All notes about KF:

                    I recorded some simple demos (could add others or replace em if needed) and made a little script to select em. It always uses the default res (can only parse 1 monitor) and writes that into the default config. In the comments you see what to change if you want to use your standard settings instead. I used 1 level with nvidia rendering bugs, 1 normal speed one, 2 with speed problems.

                    Comment

                    Working...
                    X