Announcement

Collapse
No announcement yet.

AMD Catalyst 9.9 Released !

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

  • #31
    Originally posted by rbmorse View Post
    If you try to run bleeding edge software on ancient hardware at some point you're bound to discover that backward compatibility is more than an abstract concept.
    Yeah. That's right. Good point.
    Although, this is because of non open source bleeding edge software, but is it true for open source ones too?
    I really want to know, whether open source drivers which are written to work on ancient hardware, are still in living? are they distributed by linux vendors? Like for example drivers for cable network cards?
    Or are they thrown away? just a void optimism?

    Comment


    • #32
      One of the arguments against a static ABI is that it forces developers to update drivers that any ABI changes break. So most in-kernel drivers should be fine - outside of that though, and it's up to the driver devs to keep up to date with new software. But of course that's the beauty of open source - you can change it yourself if it doesn't work.

      Comment


      • #33
        Originally posted by mirv View Post
        One of the arguments against a static ABI is that it forces developers to update drivers that any ABI changes break.
        Information flow could potentially be better though even with current system. As it is, it's possible out-of-tree developers won't even know of the changes before the new kernel version is already out and end-users complain things don't work anymore. It's easy to say though that it should be better; I can't immediately think of any simple way to do it in the current system.

        Comment


        • #34
          Fix for hardware cursor?

          Originally posted by Technojunkie View Post
          Yup. 64-bit F11, Radeon HD 3200 integrated here. I was running 9.8 with the fix to allow hardware cursor previously, if that makes a difference. Too soon to comment on stability. Kernel 2.6.30.5-43.fc11.x86_64.
          Could you please send a pointer to this fix?

          Thanks,

          (sorry for late post)

          Comment


          • #35
            Just installed. 9.9 results in a visible regression in 2d performance under Compiz (for example when scrolling Opera or Firefox). The 9.10 beta also suffers from this.

            9.8 on the other hand, offers the best 2d performance I have ever seen from fglrx so far.

            Comment


            • #36
              fglrx 9-9 is really bad when you try to watch full hd with mplayer. using gl/gl2 then it stopps after a few seconds for a short time, raising cpu usage to 100% at this point. xv is faster but without vsync, so basically unwatchable. radeon oss driver can handle xv for full hd content just fine. why is fglrx so EXTREMELY bad with xv? nvidia with vdpau is the best with that but with a fast cpu at least xv is needed. opengl is no valid choice.
              Last edited by Kano; 09-17-2009, 03:45 PM.

              Comment


              • #37
                Anyone with Karmic care to test if the 9.10 beta has xv vsync?

                Originally posted by Alberto Milone
                Can you see if you can still reproduce the problem with the latest driver in Ubuntu Karmic, please?

                [Source]

                Comment


                • #38
                  It seem you believe in miracles.

                  Comment


                  • #39
                    Originally posted by Kano View Post
                    It seem you believe in miracles.
                    Heh, if radeon can do it fglrx can too. It's a matter of when, rather than if.

                    I'm not that optimistic for 9.10 myself - just forwarding what the Ubuntu devs are asking. I'd test this myself but latest Karmic daily doesn't boot on my system (grub2 "boot error" message with no indication as to the cause).

                    Comment


                    • #40
                      Been noticing this
                      Code:
                      [fglrx:fireglAsyncioIntEnableMsgHandler] *ERROR* interrupt source ff000066 is not supported on this hardware (return code = 1)
                      showing up with catalyst 9.9

                      looks like it's an outstanding one:
                      https://bugs.launchpad.net/fglrx/+bug/288620

                      [EDIT]
                      *surprise*
                      [/EDIT]

                      [EDIT2]
                      I seriously hope that what's left of AMD management is seriously considering poaching nVidia driver staff...
                      [/EDIT2]
                      Last edited by cutterjohn; 09-23-2009, 09:35 PM.

                      Comment


                      • #41
                        Also ran into another bug that according to launchpad is catalyst driver related. In this case the notebook thrashes severely on wake from sleep(to the point of being almost entirely responsive in one case).

                        The launchpad bug entry indicates that this is a regression with 9.9 as it, apparently, appeared with older drivers at times as well. One user stated that installing the karmic early release 9.10 solved the problem for him.

                        I just hope that this doesn't appear with the final 9.10 release as well.

                        launchpad bug entry:
                        https://bugs.launchpad.net/ubuntu/+bug/391628

                        [EDIT=much later, 10/12/09]
                        I'm going to chalk this one up to a kernel update problem or some other system lib as it appears to have disappeared for me now that I've gone through several update cycles, i.e. hasn't happened to me in quite some time now...
                        [/EDIT]
                        Last edited by cutterjohn; 10-12-2009, 11:29 AM.

                        Comment


                        • #42
                          Oh joy anothe rpost in a row by me, but I just noticed that the black screne corruption bug, when entering a reply on /. and scrolling ff using the arrow keys, it looked like a black portion of the entire message entry form appeared at the top of the screen while the entry form was still close to the bottom.

                          Could this be a frame buffer updating problem?
                          [EDIT2]
                          As an additional note, I have only observed these black rectangle anomalies in fully 2D mode, which is what I most often work in as I have to disable compiz/beryl for the notebook to reliably wake from sleep and be able to play video. (Video is not a big deal for me for the most part, but there are youtube items & previews that I look at on occassion as well as other things...

                          Also still Ubuntu 9.04 x86-64 plus current updates...
                          [/EDIT2]


                          [EDIT]
                          And on an almost entirely unrelated note...
                          [/EDIT]
                          (Also noticed what I presume to be more shader related problems with more games under windows(Vista SP3 32b). Oblivious which seemed to work fine when I briefly tested it long ago but now when I was testing a new mod and happened to cross the bridge in the city with the vampire count w/lots of those fire pot things triggered that atikmdag has stopped working bug. Also happened in a tavern when attempting to talk to some NPCs. Seems to be some sort of shader related weirdness going on with catalyst drivers under windows. Unfortunately I have NONE of these games installed under wine so cannot test them in that environment(no disk space ATM to install either, or extra disk space plus time in reality. Odd thing is that running furmark benches or just letting it run NEVER triggers that bug... although it does manage to set the top mark for GPU temps... I suppose that I should trundle over to AMD forums and see if anyone can comment there on this... It's really a remarkable bug in that these games all have GREAT FPS until something triggers the bug, then everything freezes until the driver recovers. Also tried morrowind + MGE + tons of mods for a while and nothing triggered it there, but it is an oldder game and I don't use many shaders in MGE...)

                          Anyways, the upshot being does anyone know of some sort of demo that I could run in linux AND windows that would maximize shader usage and possibly a second that would max shader usage plus VRAM usage? [EDIT] Preferably SM3 I think or anything that will exercise a maximal number of shader cores under load... [/EDIT]
                          Last edited by cutterjohn; 10-12-2009, 11:28 AM.

                          Comment

                          Working...
                          X