Announcement

Collapse
No announcement yet.

E-450 graphics performance issues

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

  • E-450 graphics performance issues

    I have serious graphics performance issues with a new Lenovo X121e notebook that uses the E-450 APU.

    - GNOME shell runs very sluggish. Often the framerate seems to dip down to ~10fps, for instance when switching between workspaces. That's no better than my old POS Intel Atom GMA950 netbook!
    - I get only ~500fps in glxgears (not a benchmark, I know, but it still seems to be very slow).
    - The game TORCS is not very demanding AFAIK, but still runs pretty badly at < 30fps.
    - 2D performance is also noticeably worse than on the old Atom netbook (e.g. scrolling in a browser)

    OpenGL is definitely accelerated, according to glxinfo, and I'm seeing the same on Ubuntu 12.04 (stock or with updated kernel/mesa/ddx, doesn't matter) and Fedora 17 (Live USB).

    What the heck is going on?! The radeon drivers might not be as optimized as Intel's drivers, but this sort of performance is unacceptable. The GPU should be at least an order of magnitude faster than the old GMA950, but I see performance that is not actually better than this very old non-GPU.

  • #2
    I use the same laptop, but I don't use Gnome Shell. From what I remember, Gnome Shell's/Clutter's Sync-to-Vblank option is slowing stuff down with Catalyst (it was an issue on Ubuntu (Compiz), too, which got fixed with Ubuntu 12.04. So try turning that off first.

    Also, Catalyst's 2D performance is overall not *that* impressive *yet*, you can try the experimental ShadowPrimary acceleration.

    #edit: Just saw that you were talking about the open source radeon driver. Well, that should not happen. 2D is kickass with it for me.

    Comment


    • #3
      Well, I see pathologically bad graphics performance in many places. For instance, Ubuntu's login manager is so slow you can watch it redraw, no kidding, and Unity is quite slow overall, too. Scrolling in the dash application list is unusuably slow, it just crawls.

      I wonder if that is a hardware issue, or something specific to the E-450. Which model variant do you have, the old one with the E-350 or the new one with E-450?

      Comment


      • #4
        AFAIK E-450 default is low power profile. Did you try changing the power profile?

        http://www.x.org/wiki/RadeonFeature#...gement_Options

        Comment


        • #5
          I messed around with power_method and power_profile in various ways, but that does not seem to affect performance or power usage. According to /sys/kernel/debug/dri/0/radeon_pm_info the GPU is always hovering around with a very low engine clock (5000-15000 KHz), and the clock never exceeds 200 MHz (the "default clock"). The Radeon 6320 should go up to 508 MHz by default. Is power management and clock control completely broken or what is going on?

          Please, can someone else report what radeon_pm_info says for an E350/450?

          Comment


          • #6
            By the way, is the drm driver supposed to dump info about available power states in the kernel log (dmesg)? According to some older posts it did that, but I do not see anything like that, only a "power management enabled" message.

            Comment


            • #7
              That dumping was only via manually added printfs, IIRC, never in a released kernel.

              Comment


              • #8
                Originally posted by brent View Post
                I have serious graphics performance issues with a new Lenovo X121e notebook that uses the E-450 APU.

                - GNOME shell runs very sluggish. Often the framerate seems to dip down to ~10fps, for instance when switching between workspaces. That's no better than my old POS Intel Atom GMA950 netbook!
                - I get only ~500fps in glxgears (not a benchmark, I know, but it still seems to be very slow).
                - The game TORCS is not very demanding AFAIK, but still runs pretty badly at < 30fps.
                - 2D performance is also noticeably worse than on the old Atom netbook (e.g. scrolling in a browser)

                OpenGL is definitely accelerated, according to glxinfo, and I'm seeing the same on Ubuntu 12.04 (stock or with updated kernel/mesa/ddx, doesn't matter) and Fedora 17 (Live USB).

                What the heck is going on?! The radeon drivers might not be as optimized as Intel's drivers, but this sort of performance is unacceptable. The GPU should be at least an order of magnitude faster than the old GMA950, but I see performance that is not actually better than this very old non-GPU.
                Please post the output of 'glxinfo | grep OpenGL'

                Comment


                • #9
                  droidhacker, OpenGL is accelerated, if that is what you're getting at. I'm currently experimenting with an updated Mesa driver, but if anything that improved performance a little bit. Anyway, here's the output:

                  OpenGL vendor string: X.Org
                  OpenGL renderer string: Gallium 0.4 on AMD PALM
                  OpenGL version string: 2.1 Mesa 8.1-devel (git-6404095 precise-oibaf-ppa)
                  OpenGL shading language version string: 1.30
                  Apparently (verified by someone else with a desktop E350 mainboard), the kind of performance I'm experiencing is "normal" with current drivers. This is very disappointing. I carefully researched how well Brazos APUs work nowadays, and everyone assured me there are no major issues anymore.

                  Comment


                  • #10
                    Enable 2D tiling in your xorg.conf.
                    Option "ColorTiling2D" "True"
                    in the device section.

                    Comment


                    • #11
                      Why do you need a xorg.conf file these days for oss drivers? Wouldnt it be more logical when those options are default?

                      Comment


                      • #12
                        I already enabled 2D color tiling earlier, and it seems to help a little bit, but not much. Performance still is largely pathologically bad. For instance, a simple GLSL-based plasma effect I made a while ago runs with approximately similar performance to GMA950, again. It's a very simple fragment shader, yet a fullscreen window of it runs at just ~20 fps.

                        Is this really the current state of Mesa's APU support? What's holding back performance?

                        Comment


                        • #13
                          Originally posted by brent View Post
                          Is this really the current state of Mesa's APU support? What's holding back performance?
                          Performance is fluid on my APU systems. Try profiling that apps you are having problems with and see where the bottlenecks are. The open source driver is not as fast as the closed driver at the moment. If maximum 3D performance is what you need, you may want to use the closed source driver in the interim.

                          Comment


                          • #14
                            fglrx offers great OpenGL support and performance, and power management is also much better (~1W less at idle!), but literally everything else is worse or completely broken. I don't care that much for the driver being closed-source, but features like suspend to ram simply must work reliably. OTOH, I don't need superb OpenGL support/performance, but it should be *acceptable* at least.

                            Performance is fluid on my APU systems.
                            Can you be more specific?

                            agd5f, by the way, do you know if power management works correctly on Brazos APUs? As I said before, radeon_pm_info spits out some rather nonsensical values. Is the GPU actually clocked correctly up to its full frequency? I did some simple tests and observed no difference between low/mid/high profiles or dynpm. What power management features are missing to make power consumption as low as with fgrlx?
                            Last edited by brent; 07-07-2012, 11:21 AM.

                            Comment


                            • #15
                              I enabled some DRM debugging, and this is what I see for *every* call to r600_get_dynpm_state:

                              [25233.444967] [drm:r600_pm_get_dynpm_state], Requested: e: 17369 m: 0 p: 1
                              The clock unit is 10 KHz. Does that mean the engine clock is essentially fixed to ~173 MHz for some reason? I'll try to find out what's in the power state tables. If the debug output is right, the GPU is always running at basically 1/3 of the specified clock, not even taking the "turbo" mode into consideration.

                              Comment

                              Working...
                              X