Announcement

Collapse
No announcement yet.

Linux 3.12 Brings Big AMD Radeon Improvements

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

  • #16
    What happened to all of the performance improvements in 3.11? There was a similar article a couple months ago showing how 3.11 stomped 3.10, and this article shows it doing worse.

    Comment


    • #17
      Originally posted by cbgoding View Post
      What happened to all of the performance improvements in 3.11? There was a similar article a couple months ago showing how 3.11 stomped 3.10, and this article shows it doing worse.
      That was as a result of re-clocking the GPU properly with the DPM changes. They had to be enabled manually in 3.11 but were going to be enabled by default in 3.12, AFAIK. So these improvements could simply be a result of DPM being turned on.

      A comparison against 3.11 with and without DPM would be useful to know if it's specifically a change in 3.12 or not.

      Comment


      • #18
        Originally posted by tom.higgy View Post
        That was as a result of re-clocking the GPU properly with the DPM changes. They had to be enabled manually in 3.11 but were going to be enabled by default in 3.12, AFAIK. So these improvements could simply be a result of DPM being turned on.

        A comparison against 3.11 with and without DPM would be useful to know if it's specifically a change in 3.12 or not.
        3.12 does NOT turn DPM on by default. There were way too many changes in this kernel to consider the feature stable yet, but maybe 3.13 will do so. Also, the 4870 card michael is testing here is unlikely to get any performance benefit from it anyway - it's likely clocked high by default, and would just save power with DPM. It's the newer cards that tend to be clocked low by default and get big speedups - Radeon SI cards, Cayman, and the APU parts. And likely mobile as well.

        Comment


        • #19
          So, DPM is enabled by default in kernel 3.12? Its still not clear, because it says "Radeon HD 8000 "Sea Islands" Dynamic Power Management (DPM)", so, I think is the same support older gpu have since kernel 3.11, and that would mean its still not by default?

          Comment


          • #20
            compiling mesa from git

            hi - sorry for being kind of offtopic... but I am in the mood to experiment. I have want to compile Mesa from git to see oss-radeon-driver improvements. Iīm already on kernel 3.11.3 but mesa is 9.2.1.

            Is it enough to compile only MESA or do i also need DRM and radeon-DDX and there is also somethin caled DRI... so my problem is I donīt know which howto I should follow at best:

            http://dri.freedesktop.org/wiki/Building/

            http://www.x.org/wiki/radeonBuildHowTo/

            or even http://www.reddit.com/r/linuxquestio...n_ubuntu_1210/ and adapting it to my opensuse 13.1 RC1

            any hint much appreciated!

            Comment


            • #21
              Originally posted by Nille View Post
              And please with some contemporary games. No one care this games.
              They are still useful as a base benchmark. And while I agree throwing more games into the mix would be better, please stop being up yourself. Seriously.

              Comment


              • #22
                As for the performance gains:

                And it just keeps rolling along...

                Comment


                • #23
                  Originally posted by agd5f View Post
                  The only asic that commit affects is r6xx chips. R7xx and newer are not affected since they use the async dma rings for bo moves and have since kernel 3.8 or so. The only time the new cpdma code would come into effect for newer chips is when you run the kernel bo tests. TBH, I don't know of any particular changes that would have had much impact on performance.
                  Alex, Michael mentioned some "a major ring handling clean-up" is it possible that the old 'dirty' code was inefficient and that the clean-up gave it a more direct, faster codepath ( or stopped hitting SLOW code paths). I could be totally wrong but if the blit work doesn't effect this test, thats the only other thing mentioned by Michael that stands out. Mesa 9.3 git also uses the SB backend by default, doesnt it? So does 3.12 include anything that a better shader compiler would benefit? Just spitballing idas here

                  Comment


                  • #24
                    Originally posted by Ericg View Post
                    Mesa 9.3 git also uses the SB backend by default, doesnt it?
                    It does now, yes. But (as I can see this as being a potential source of confusion), it should be noted that Michael has published earlier benchmarks utilizing 9.3 dev builds that came prior to the change which enabled sb by default.

                    Comment


                    • #25
                      Seeing as the results even surprised Alex, I'll run more tests today on a completely different system with various generations of AMD GPUs and try to get these results up tomorrow to see how those go.
                      Michael Larabel
                      http://www.michaellarabel.com/

                      Comment


                      • #26
                        Originally posted by Michael View Post
                        Seeing as the results even surprised Alex, I'll run more tests today on a completely different system with various generations of AMD GPUs and try to get these results up tomorrow to see how those go.
                        Maybe a git bisect too Michael? Just to narrow down the commit or at least a series of commits that could've effected this

                        Comment


                        • #27
                          Originally posted by Ericg View Post
                          Maybe a git bisect too Michael? Just to narrow down the commit or at least a series of commits that could've effected this
                          First I'll run another article or two out of it (to see just how rare or widespread the improvements are and for maximizing views...) before letting PTS bisect it
                          Michael Larabel
                          http://www.michaellarabel.com/

                          Comment


                          • #28
                            Originally posted by Michael View Post
                            Seeing as the results even surprised Alex, I'll run more tests today on a completely different system with various generations of AMD GPUs and try to get these results up tomorrow to see how those go.
                            Thanks, I and probably others here appreciate the time you put into Phoronix and am appreciative of you taking the time to run through this again. Not a lot of people would do something over like this, they'd probably publish it and call it quits.

                            Comment


                            • #29
                              Originally posted by Michael View Post
                              Seeing as the results even surprised Alex, I'll run more tests today on a completely different system with various generations of AMD GPUs and try to get these results up tomorrow to see how those go.
                              Thanks very much

                              Can we request from a E-350 to a A10-6800K ?

                              Comment


                              • #30
                                Originally posted by smitty3268 View Post
                                3.12 does NOT turn DPM on by default. There were way too many changes in this kernel to consider the feature stable yet, but maybe 3.13 will do so. Also, the 4870 card michael is testing here is unlikely to get any performance benefit from it anyway - it's likely clocked high by default, and would just save power with DPM. It's the newer cards that tend to be clocked low by default and get big speedups - Radeon SI cards, Cayman, and the APU parts. And likely mobile as well.
                                OK, so these improvements are for other reasons. Even better!

                                Comment

                                Working...
                                X