Announcement

Collapse
No announcement yet.

The Direction Of ATI Radeon Graphics In Ubuntu 11.04

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

  • The Direction Of ATI Radeon Graphics In Ubuntu 11.04

    Phoronix: The Direction Of ATI Radeon Graphics In Ubuntu 11.04

    With Ubuntu 11.04 arriving in a little more than a month, the key packages to be found in this "Natty Narwhal" release are nearly settled. For those concerned about the open-source ATI graphics stack, the packages to note are the Linux 2.6.38 kernel, Mesa 7.10.1, and xf86-video-ati 6.14.0. What does this mean for the conventional user? This article provides a brief look at the state of open-source ATI in Ubuntu 11.04.

    http://www.phoronix.com/vr.php?view=15762

  • #2
    Originally posted by Phoronix
    Right now there is no publicly released Catalyst driver that supports the Linux 2.6.38 kernel and X.Org Server 1.10 as used by Ubuntu 11.04, but this will come with the Catalyst 10.3 or 10.4 releases. If the necessary kernel/xorg-server support does not arrive in this month's Catalyst 10.3 release, AMD once again will be seeding Canonical with a pre-release of Catalyst 10.4 to provide the necessary binary driver support before the Ubuntu 11.04 Beta.
    Catalyst 10.x/11.x

    Comment


    • #3
      Originally posted by d2kx View Post
      Catalyst 10.x/11.x

      Hmm yes, we seem to be in 2011 now.
      Michael Larabel
      http://www.michaellarabel.com/

      Comment


      • #4
        the benchmark show how Catalyst show middle finger to us.

        Comment


        • #5
          "We are testing how fast the ati driver is but we are not testing how fast it really is, only how fast it is as default.."?

          I happen to have the HD 4670 and I have an Athlon II 240.
          I use Archlinux and git versions of mesa, xf86-video-ati and the drm-radeon-testing branch of the kernel, all build on 20110306.
          I have color tiling and page flipping enabled.

          I only run on 1680x1050 and I have these framerates:
          http://phoronix.com/forums/showthrea...849#post180849

          Does 1680x1050 to 1920x1200 cause 30% performance drop in Urban Terror or is it only color tiling and page flipping?

          I also don't understand the nexuiz result. On the "middle"-preset for the effects I get between 60 and 90 fps constantly.
          Just saying, the statement should maybe a bit weaker than "Ubuntu 10.10 does not produce a playable frame-rate for the Radeon HD 4670 graphics card on the open-source drivers". Especially the second "Shadow"-Option in the effects-pane seems to have a huge impact on performance (on ultra settings with = 2 fps, without = 20-40 fps).

          Comment


          • #6
            Mesa 7.10
            ## VGA ##
            AMD: X1950XTX, HD3870, HD5870
            Intel: GMA45, HD3000 (Core i5 2500K)

            Comment


            • #7
              Originally posted by fabioamd87 View Post
              the benchmark show how Catalyst show middle finger to us.
              The benchmark is flawed. He's running the the opensource drivers with SwapBuffersWait on, for maximum fps it must be turned off. It's like having a street car race where one car (radeon) stops for every red light while the other one (fglrx) just ignores them. Of course the car which ignores the red lights is going to win (or crash..).

              Not that Catalyst wouldn't have a higher fps anyway but this is not a fair comparison on is not showing the real performance improvements of the oss drivers.

              Comment


              • #8
                Originally posted by monraaf View Post
                The benchmark is flawed. He's running the the opensource drivers with SwapBuffersWait on, for maximum fps it must be turned off. It's like having a street car race where one car (radeon) stops for every red light while the other one (fglrx) just ignores them. Of course the car which ignores the red lights is going to win (or crash..).

                Not that Catalyst wouldn't have a higher fps anyway but this is not a fair comparison on is not showing the real performance improvements of the oss drivers.
                At the same time then is the Catalyst driver results not fair because Catalyst AI is not manually set to the highest level by default, etc. Testing at the defaults is done for a reason. It's the developers that choose the defaults.
                Michael Larabel
                http://www.michaellarabel.com/

                Comment


                • #9
                  That is a decent argument. It's the default behaviour that people will get without tweaking, but one has to admit that the default Ubuntu behaviour is really stupid (Compiz always on).

                  Still, it would be nice to benchmark what the drivers can actually do at this moment, especially if we're talking about trivial changes. The way it is right now, you have to add 60% to every number you publish for open source drivers, and this can't be the point of a benchmark.

                  What about a "default" benchmark (no tweaks) and a "best" benchmark (minor one-liner tweaks that everybody uses anyway)?

                  Comment


                  • #10
                    Originally posted by monraaf View Post
                    The benchmark is flawed. He's running the the opensource drivers with SwapBuffersWait on, for maximum fps it must be turned off. It's like having a street car race where one car (radeon) stops for every red light while the other one (fglrx) just ignores them. Of course the car which ignores the red lights is going to win (or crash..).

                    Not that Catalyst wouldn't have a higher fps anyway but this is not a fair comparison on is not showing the real performance improvements of the oss drivers.
                    The benchmark is good. Those are the defaults. This benchmark is only about the default drivers on Ubuntu. Users don't know stuff about swapbuffers and whatnot. They simply use the stuff at the recommended defaults supplied by Canonical.

                    Comment


                    • #11
                      Originally posted by pingufunkybeat View Post
                      That is a decent argument. It's the default behaviour that people will get without tweaking, but one has to admit that the default Ubuntu behaviour is really stupid (Compiz always on).

                      Still, it would be nice to benchmark what the drivers can actually do at this moment, especially if we're talking about trivial changes. The way it is right now, you have to add 60% to every number you publish for open source drivers, and this can't be the point of a benchmark.

                      What about a "default" benchmark (no tweaks) and a "best" benchmark (minor one-liner tweaks that everybody uses anyway)?
                      That 'everybody uses'? I don't even think 10% of Ubuntu users will even go ahead and manually add swapbufferswait and other tweaks, etc.

                      Like I've done swapbufferswait disabling in that January article when page flipping was the focus of the article.

                      In terms of 'best' performance, as I and Matthew have said many times now whenever someone complains about defaults being used: I will gladly do tweaked benchmarks vs. defaults, etc, IF there is a publicly available document by the project that details for end-users what they recommend for optimal performance/features/whatever. Yet no one has come forth to do such. Even a Wiki page for what they recommend as settings for enthusiasts.
                      Michael Larabel
                      http://www.michaellarabel.com/

                      Comment


                      • #12
                        Originally posted by Michael View Post
                        That 'everybody uses'? I don't even think 10% of Ubuntu users will even go ahead and manually add swapbufferswait and other tweaks, etc.
                        If you told them how to, and what it brings, I'm sure that they would.

                        Comment


                        • #13
                          I mean, like I said: testing defaults is legitimate. Nobody faults you for that.

                          But dynamic powersaving is not default, and you need to echo things to /sys as root to turn it on, and you did test that. And gallium is not default at all (on r600, and definitely not on Intel), and that gets tested.

                          Just like with VDPAU, which is not default on anything, and needs to be enabled first.

                          These are the things that people are interested in. Default is fine, but it's nice to also see progress, and if you're going to test the frame rates, it is really useful to tell the Ubuntu user (who might not know better) that Compiz is cutting his frame rate in half. He will be interested, and he will have a better gaming experience.

                          Comment


                          • #14
                            Originally posted by Michael View Post
                            At the same time then is the Catalyst driver results not fair because Catalyst AI is not manually set to the highest level by default, etc. Testing at the defaults is done for a reason. It's the developers that choose the defaults.
                            You should do multiple ones. A set with the defaults and a set with optimizations. It has more value to show the performance of optimally configured software in comparison to the performance of software with poorly configured defaults than only the performance of software at stock settings. The former is useful. The latter is worthless for any user technical enough to care about the benchmarks.

                            Comment


                            • #15
                              Originally posted by Shining Arcanine View Post
                              You should do multiple ones. A set with the defaults and a set with optimizations. It has more value to show the performance of optimally configured software in comparison to the performance of software with poorly configured defaults than only the performance of software at stock settings. The former is useful. The latter is worthless for any user technical enough to care about the benchmarks.
                              It's not economical to do multiple ones each and every time.
                              Michael Larabel
                              http://www.michaellarabel.com/

                              Comment

                              Working...
                              X