Page 1 of 6 123 ... LastLast
Results 1 to 10 of 59

Thread: AMD Radeon: Windows Catalyst vs. Linux Gallium3D vs. Linux Catalyst

  1. #1
    Join Date
    Jan 2007
    Posts
    14,369

    Default AMD Radeon: Windows Catalyst vs. Linux Gallium3D vs. Linux Catalyst

    Phoronix: AMD Radeon: Windows Catalyst vs. Linux Gallium3D vs. Linux Catalyst

    Our latest Windows vs. Linux benchmarks on Ubuntu 14.04 LTS are of three different AMD Radeon graphics cards from three different generations as we test the performance of Microsoft Windows 8.1 against Ubuntu 14.04 LTS, when using both the AMD Catalyst proprietary driver and the open-source R600/RadeonSI Gallium3D drivers.

    http://www.phoronix.com/vr.php?view=20076

  2. #2
    Join Date
    Nov 2013
    Posts
    45

    Default

    It would be really interesting to see if the performance delta between Linux and Windows on the R9 290 card has been reduced in the latest drivers.

  3. #3
    Join Date
    Mar 2013
    Location
    Porto
    Posts
    148

    Default not bad

    the r600 are in great shape, radeonsi need more work, but well done messa devs great job, next step?! put all devs working in opensource driver and merge the drivers

  4. #4
    Join Date
    Jan 2012
    Posts
    41

    Default

    Quote Originally Posted by Herem View Post
    It would be really interesting to see if the performance delta between Linux and Windows on the R9 290 card has been reduced in the latest drivers.
    I agree. It would be good to know if it is time to buy a new card

    Michael, how many please do you need to hear to do these benchmarks? (PLEEEEEEEEEEAAAASSSEEEE!!!!)

  5. #5
    Join Date
    Feb 2008
    Location
    Linuxland
    Posts
    4,994

    Default

    Quote Originally Posted by vein View Post
    I agree. It would be good to know if it is time to buy a new card

    Michael, how many please do you need to hear to do these benchmarks? (PLEEEEEEEEEEAAAASSSEEEE!!!!)
    If your "please" is written on a 50$ note

  6. #6
    Join Date
    Sep 2008
    Location
    Vilnius, Lithuania
    Posts
    2,525

    Default

    Not bad, the 6870 seems to be in a sort of a sweet spot where it's very close to Catalyst (and the Linux Catalyst seems to be better than the Windows one in some cases).

  7. #7
    Join Date
    Feb 2008
    Posts
    805

    Default

    What is wrong with triangle test on both intel and radeon drivers

    http://openbenchmarking.org/embed.ph...4b8a6c3eb3&p=1

    I guess that needs to be decent performer if someone wants later tesselation to play at decent speed

  8. #8
    Join Date
    Nov 2012
    Location
    France
    Posts
    541

    Default

    Quote Originally Posted by rikkinho View Post
    the r600 are in great shape, radeonsi need more work, but well done messa devs great job, next step?! put all devs working in opensource driver and merge the drivers
    It doesn't work like that.

    Quote Originally Posted by dungeon View Post
    What is wrong with triangle test on both intel and radeon drivers

    http://openbenchmarking.org/embed.ph...4b8a6c3eb3&p=1

    I guess that needs to be decent performer if someone wants later tesselation to play at decent speed
    It's not like tesselation is too essential, there are more important topics to spend time on currently.

  9. #9
    Join Date
    Jun 2010
    Location
    ฿ 16LDJ6Hrd1oN3nCoFL7BypHSEYL84ca1JR
    Posts
    1,023

    Default

    The open-source Radeon drivers can't yet handle GiMark or TessMark due to missing OpenGL/GLSL support.
    Sorry, not quite.

    tessmark really doesn't run, probably because it needs tesselation.

    But gimark seems to only need opengl 3.3 and does not use anything deprecated. So this works:
    Code:
    MESA_GL_VERSION_OVERRIDE=3.3 MESA_GLSL_VERSION_OVERRIDE=330 ./start_gimark_windowed_1024x640.sh
    Might be a good idea to ask the developer to use a core context for his benchmarks.

    (How much does a secret closed source "benchmark" mean anyway. GPUTest's numbers are always very strange compared to other benchmarks)

    edit:
    Code:
    2014:03:27@11:41:56(0000000016) < > [GpuMonitor plugin] GPU1 - Advanced Micro Devices, Inc. [AMD/ATI] Wimbledon XT [Radeon HD 7970M] (rev ff)
    
    2014:03:27@11:41:56(0000000017) < > [GpuMonitor plugin] GPU1 - PCI codes: 0x1002-0x6800
    2014:03:27@11:41:56(0000000018) < > GPU monitoring thread started.
    2014:03:27@11:41:56(0000000019) < > Display size: width=1920, height=1080
    2014:03:27@11:41:57(0000000020) < > Window size: width=1920, height=1080, left=0, top=0
    2014:03:27@11:41:57(0000000021) < > GiMark - OpenGL renderer init OK.
    2014:03:27@11:41:57(0000000022) < > GiMark - VSYNC disabled (xxx_SwapInterval_xxx available).
    2014:03:27@11:41:57(0000000023) < > GiMark - OpenGL version detected: 3.3
    2014:03:27@11:41:57(0000000024) < > GiMark - # OpenGL extensions: 146
    2014:03:27@11:41:57(0000000025) < > GiMark - OpenGL - Renderer model: Gallium 0.4 on AMD PITCAIRN
    2014:03:27@11:41:57(0000000026) < > GiMark - OpenGL - Renderer vendor: X.Org
    2014:03:27@11:41:57(0000000027) < > GiMark - OpenGL - API version: 3.3 (Core Profile) Mesa 10.2.0-devel (git-b02bcea)
    2014:03:27@11:41:57(0000000028) < > GiMark - OpenGL - Shading language version: 3.30
    2014:03:27@11:41:57(0000000029) < > GiMark : init OK.
    2014:03:27@11:43:01(0000000030) < > [Benchmark_Score] - module: GiMark - Score: 906 points (1920x1080 fullscreen, duration:60000 ms).
    2014:03:27@11:43:01(0000000031) < > Exit from render thread
    2014:03:27@11:43:01(0000000032) < > # frames rendered: 906
    2014:03:27@11:43:01(0000000033) < > GpuTest 0.7.0^M
    http://www.geeks3d.com^M
    ^M
    Module: GiMark^M
    Score: 906 points (FPS: 15)^M
    ^M
    Settings:^M
    - 1920x1080 fullscreen^M
    - antialiasing: Off^M
    - duration: 60000 ms^M
    ^M
    Renderer:^M
    - Gallium 0.4 on AMD PITCAIRN^M
    - OpenGL: 3.3 (Core Profile) Mesa 10.2.0-devel (git-b02bcea)
    Last edited by ChrisXY; 03-27-2014 at 06:45 AM.

  10. #10
    Join Date
    Feb 2008
    Posts
    805

    Default

    Quote Originally Posted by Calinou View Post
    It's not like tesselation is too essential, there are more important topics to spend time on currently.
    Yep currently there is no point to even implement tesselation it will be very slow because if i guess this test shows we are 3-7 times slower, so 3-7 times triangle bound... and that always does metter later for tesselation will even more .

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •