Page 2 of 5 FirstFirst 1234 ... LastLast
Results 11 to 20 of 41

Thread: R600 Gallium3D Loses To AMD Catalyst Legacy Driver On Windows 8

  1. #11
    Join Date
    Jul 2013
    Posts
    204

    Default

    Was this the stock AMD WDDM 1.1 drivers, 13.1 legacy or 13.4 beta legacy?

    I found that WDDM, 13.1 and 13.4 are all terrible when it comes to coming. My gaming performance with those drivers on an HD4200 IGP was horrid compared to latest drivers on Windows 7.

  2. #12
    Join Date
    Feb 2008
    Location
    Linuxland
    Posts
    5,051

    Default

    Nice work Michael, you got all the combinations I wanted to see.

  3. #13
    Join Date
    Feb 2008
    Location
    Santiago, Chile
    Posts
    240

    Default

    The news here is: Catalyst is losing more and more of its edge against the free driver. Last year we were speaking about 1/6th of the performance across the board, now we're talking about 1/2, 1/3.

    Now please, Mesa guys, boost that triangle figure! If we focus in GPUTest / Triangle we can have some serious optimizations.

  4. #14
    Join Date
    Sep 2011
    Posts
    184

    Default

    Quote Originally Posted by przemoli View Post
    Second that.

    For latency avrg is equal to avrg FPS... Here is some good article about good methodology for couting and comparing latency:
    http://techreport.com/review/21516/i...e-benchmarking

    Hope its easy enought to integrate into PTS.
    Maybe adding a Frame Time Variance to PTS would be a good thing then.

  5. #15
    Join Date
    Sep 2012
    Posts
    283

    Default

    Quote Originally Posted by log0 View Post
    AFAIK Blender is using GL_SELECT which is not hardware accelerated in mesa. There has been a mesa patch though I think.
    If what you say is true, then how come the problem started to happen like 6-8 months ago with a mesa update ? It sounds like a regression to me.

  6. #16
    Join Date
    Sep 2011
    Posts
    24

    Default

    Quote Originally Posted by TemplarGR View Post
    Useless comparison... Apples to oranges...
    What? No.

    You're comparing different operating systems with different drivers, yes. But that's the point... this is what the end user will see. The open source driver is the default/standard on Linux systems, whereas Catalyst is the only option for Windows. It makes sense to compare whether or not Radeon on Linux can compete with Catalyst on Windows, especially when using "legacy" hardware.

    I'm really happy with Radeon's progress lately. Still too slow for "real" usage, but it's really getting there. I don't even think of FGLRX anymore.

  7. #17
    Join Date
    Jun 2009
    Posts
    1,121

    Default

    Quote Originally Posted by Newfie View Post
    What? No.

    You're comparing different operating systems with different drivers, yes. But that's the point... this is what the end user will see. The open source driver is the default/standard on Linux systems, whereas Catalyst is the only option for Windows. It makes sense to compare whether or not Radeon on Linux can compete with Catalyst on Windows, especially when using "legacy" hardware.

    I'm really happy with Radeon's progress lately. Still too slow for "real" usage, but it's really getting there. I don't even think of FGLRX anymore.
    well i would not say too slow, since in most test is very close to 1/2 and well enough above 60 FPS, sure is still half speed but is progressing very nicely, well if you have passive cooled card you could be right but at least my office 4350 is fast enough for fast desktop

  8. #18
    Join Date
    Apr 2013
    Posts
    39

    Default

    Quote Originally Posted by Alejandro Nova View Post
    Now please, Mesa guys, boost that triangle figure! If we focus in GPUTest / Triangle we can have some serious optimizations.
    The result of Triangle benchmark is above 100 thousands on my 5750 with r600g and I don't think that 4870 renders single triangle 7 times slower. It seems something is wrong with Michael's results, maybe some problem with compositor/Ubuntu?

  9. #19
    Join Date
    Aug 2009
    Location
    Russe, Bulgaria
    Posts
    506

    Default

    Quote Originally Posted by log0 View Post
    AFAIK Blender is using GL_SELECT which is not hardware accelerated in mesa. There has been a mesa patch though I think.
    I confirm that.

  10. #20
    Join Date
    Aug 2009
    Location
    Russe, Bulgaria
    Posts
    506

    Default

    Quote Originally Posted by vadimg View Post
    The result of Triangle benchmark is above 100 thousands on my 5750 with r600g and I don't think that 4870 renders single triangle 7 times slower. It seems something is wrong with Michael's results, maybe some problem with compositor/Ubuntu?
    Vadim, did you expect better performance out of selected test-cases? Lately I had the impression we are *very* close to catalyst.

Posting Permissions

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