Announcement

Collapse
No announcement yet.

Is AMD's New 2D Acceleration Architecture Still Slow?

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

  • #31
    I don't have any graphics corruption in Firefox or any other app. I did however have a couple of spontaneous reboots, mostly triggered by switching or closing a Firefox tab. I don't know if this is a general driver issue or only triggered by the new 2d accel code paths. But I've disabled the new 2D accel code for now.

    On a positive note. The gstreamer-gl plugin from git now works correctly with this Catalyst release. So it's now possible to have tear-free video with gstreamer based video players and fglrx.

    Comment


    • #32
      I have some issues with firefox, and pretty much well anything gtk actually, notably when scrolling, but only on the secondary monitor (running 2 X servers, no xinerama or expanded desktop or anything). No black patches, but large areas of "blank" (grey, white, whatever is the background colour) until the window is refreshed (switching virtual desktops will fix it, until I scroll again).
      This is with E16, gentoo 64bit, hd3650, kernel 2.6.32 and xorg-server 1.7.6 (from memory, I'm tired and not at my own computer right now). No compositing enabled.

      Comment


      • #33
        For me it's an annoying bug. But you are an ati fanboy...

        Comment


        • #34
          Originally posted by Kano View Post
          You must really love ATI when that does not annoy you.
          i'm not gonna lie, i am a rather large AMD fan, but honestly, its not something that really affects my computer using habbits in anyway, so the source of the bug being from ati is not related to the fact that it doesn't really annoy me.

          i suppose i should qualify that "a second" was figurative, literally speaking, it is more of an instant flash from normal-black-normal.

          Comment


          • #35
            Originally posted by Kano View Post
            Did you really see a driver that works with KDE 4 correctly with the new accelleration active? I really think to disable it by default, because it only causes problems with gtk2 apps. Maybe it was tested with gnome (hint: ubuntu) only.
            yeah i think is still too buggy for everyday use, especially if you use kde4 and gtk apps (yeah same problem, i thougth it was only me yeah im not crazy).

            about the ubuntu only is pretty posible i concur.

            so far only the FOSS driver render perfectly kde4 99.99% of the time,mmm in the fglrx side i dont remember if ever existed such a driver.

            ohh, anyone noticed that enable full vsync in the control center hang the X and after the ovbious restart your 3d performance goes to hell?(NOT with glxgears, i mean comparing unigine tropics bench after and before)(i know with vsync i wont get 400 fps, duh!)(i mean 58~ before vsync ~32 after vsync)(yeah unigine tropics is nasty at full settings) (yeah my monitor is set at 85 not 32 or 30)

            anyway at least is nice to see that some QA is beggining to appear in the drivers, beside the ovbious this drivers screw less things that were working fine before than most previous driver release i remember, so +1 for the lonely guy at amd's basement

            another one, anyone noticed that many shaders operation in wine with fglrx are heavily corrupted? ill post some screen later to figure out if it is fglrx regresion or wine

            Comment


            • #36
              Pixbufs

              I would have liked to have seen a Pixbufs benchmark - that is where the catalyst driver was notably poor in previous tests, and helps to explain the real-world perception of improvement I think (e.g. things like scrolling).

              Comment

              Working...
              X