Announcement

Collapse
No announcement yet.

Radeon Linux Benchmarks: Catalyst 15.3 Beta vs. Linux 4.0 + Mesa 10.6-devel

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

  • Radeon Linux Benchmarks: Catalyst 15.3 Beta vs. Linux 4.0 + Mesa 10.6-devel

    Phoronix: Radeon Linux Benchmarks: Catalyst 15.3 Beta vs. Linux 4.0 + Mesa 10.6-devel

    Before ending out March, here's some new OpenGL Linux benchmarks comparing the closed-source Catalyst 15.3 Beta driver against the Linux 4.0 development kernel with Mesa 10.6 Git for the freshest open-source graphics driver code.

    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
    I wonder why Unigine benchmarks are omitted, those are where Radeon annihilates Catalyst with up 400% performance difference.

    Comment


    • #3
      Use a better PPA

      That's NOT the freshest OSS driver code, not if it's still using LLVM 3.6.

      Big changes are present in 3.7.

      Comment


      • #4
        Originally posted by smitty3268 View Post
        That's NOT the freshest OSS driver code, not if it's still using LLVM 3.6.

        Big changes are present in 3.7.
        But it seems LLVM 3.7 is not release yet.

        Comment


        • #5
          Is there any performance improvement from catalyst 14.12 to catalyst 15.3 beta?

          Comment


          • #6
            I was surprised how much of a difference there is. It used to be the open source was nearly as fast as Catalyst, to the point where installing Catalyst was pointless. Now I'm looking at my machines wondering if I should install Catalyst 15.3 Beta. Either there's a huge regression in the drivers or AMD's Catalyst 15.3 Beta is getting serious. Even the 6870 which is usually faster with open source seems to fall apart here.

            Comment


            • #7
              Originally posted by humbug View Post
              Is there any performance improvement from catalyst 14.12 to catalyst 15.3 beta?
              good question

              Comment


              • #8
                Originally posted by Pecisk View Post
                But it seems LLVM 3.7 is not release yet.
                Neither are the Mesa and kernel versions he is testing.

                I just think it's kind of silly to use git versions of those but not LLVM - either go all-git, or all-released. A frankenstein mix of the two is weird, and claiming in the article that it's the latest code is just plain misleading.

                Comment


                • #9
                  Originally posted by smitty3268 View Post
                  Neither are the Mesa and kernel versions he is testing.

                  I just think it's kind of silly to use git versions of those but not LLVM - either go all-git, or all-released. A frankenstein mix of the two is weird, and claiming in the article that it's the latest code is just plain misleading.
                  Good point. I would like to see newest LLVM improvements too.

                  Comment


                  • #10
                    Originally posted by Dukenukemx View Post
                    I was surprised how much of a difference there is. It used to be the open source was nearly as fast as Catalyst, to the point where installing Catalyst was pointless. Now I'm looking at my machines wondering if I should install Catalyst 15.3 Beta. Either there's a huge regression in the drivers or AMD's Catalyst 15.3 Beta is getting serious. Even the 6870 which is usually faster with open source seems to fall apart here.
                    Well Catalyst improved yes, but there is just slight perf regress in radeonsi driver it seems:

                    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


                    But looking at xonotic results, r600 seems to have big regression -40% for 6870! If is true, somone who uses r600 should check and eventually bisect that.
                    Last edited by dungeon; 31 March 2015, 03:31 PM.

                    Comment

                    Working...
                    X