Announcement

Collapse
No announcement yet.

AMD catalyst 10.9 has been released

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

  • #46
    Problem is ATI has no solution. Closed driver is flawed and with newer hardware, not smooth...at least, according to many Linux owners and the open source driver is not yet ready. So, what do you do with your card? Run Windows?

    I read that the HD 5xxx cards are obtaining more progress with closed drivers but tough to know what this means, exactly.

    The open drivers are an ideal but a WIP always or at least, for a long period of time but people want to use their computers today, not months or half a year from now.

    Even if you contribute, it still can be months before fixes and features are working smooth.

    If one can deal with 'what is not yet working,' that's great for them but to expect everyone to have the same attitude/mentality/patience is not practical. The closed source driver should at least be comparable to the situation in Windows.

    Comment


    • #47
      Originally posted by Fluggo View Post

      Originally posted by erpe View Post
      Does Heroes of Newerth work again with evergreen cards?
      Can't test right now.
      No it does not

      Bridgman, comment please?

      5 releases now. Catalyst crashes inside it's OGL 4 driver.

      http://ati.cchtml.com/show_bug.cgi?id=1824

      Comment


      • #48
        I'm having a couple of issues and am wondering if anyone else is having them with 10.9 as well.

        System Config
        3 x 24" for 5760x1200
        2 via DVI, 1 via DP to DVI converter.
        Ubuntu 10.04 up2date kernel, etc.

        (Issues have been present 10.7 through 10.9)

        Firstly, when starting playback of video it seams the driver does something to the Display Port which causes the active adapter to blank the screen for about two seconds.

        When I power on my system, as the BIOS adjusts the resolution between the different BIOS/config splash screens (before GRUB) the display port adaptor blacks the connected screen for a moment. I've taken a video of the boot process and have counted the amount of frames the screen's blanked for due to a resolution change.
        Once the machine is booted, when starting video playback the screen connected via the adaptor is blanked for the same amount of frames which leads me to think that fglrx is resetting something on the display port to cause this.

        Secondly, when resizing windows past a certain size the resize process goes from being smooth to lagging very badly. It can take two or three seconds between resize events once this starts. All other screen updates from any other apps cease waiting for the resize event to complete. Is the driver throwing data between the frame store and main RAM once a certain operation size is occurring?

        Comment


        • #49
          Originally posted by Jecos View Post
          Bridgman, comment please?

          5 releases now. Catalyst crashes inside it's OGL 4 driver.

          http://ati.cchtml.com/show_bug.cgi?id=1824
          The bug report basically says :

          - original problem reproduced in house early June (w. Quake Live)
          - originally reproduced problem fixed, Quake Live works in house
          - a few weeks ago AMD dev asked for feedback re: Quake Live, no response yet
          - dev also confirmed that HON failure was repro'ed after original fix so presumably separate problem

          so... if we are talking about HON the problem was repro'ed in house a few weeks ago.

          Is the HON problem specific to Evergreen ? Don't think I saw that in the bug ticket (no mention of HW at all IIRC).

          Comment


          • #50
            Originally posted by bridgman View Post
            The bug report basically says :

            - original problem reproduced in house early June (w. Quake Live)
            - originally reproduced problem fixed, Quake Live works in house
            - a few weeks ago AMD dev asked for feedback re: Quake Live, no response yet
            - dev also confirmed that HON failure was repro'ed after original fix so presumably separate problem

            so... if we are talking about HON the problem was repro'ed in house a few weeks ago.

            Is the HON problem specific to Evergreen ? Don't think I saw that in the bug ticket (no mention of HW at all IIRC).
            Just added that info

            Comment


            • #51
              Originally posted by Qaridarium View Post
              i build packages and upgrade the packages by dpkg
              So do I under Opensuse 11.2/64. Can anybody confirm that in that case - even though the different versions show wrong values - the driver is up to date?

              Regards,
              Martin

              Comment


              • #52
                Guess the community didn't dot the i's and cross the t's enough for the dev team to look at the problem until now. The external link clearly states evergreen and the driver version.... no one cares enough about a binary blob that's been griefing them with lackluster quality for years now. if they did care maybe they would of effectively gone through all the loops of reporting it so the dev team couldn't ignore it.

                Fglrx couldn't die any slower a death now can it?

                Comment


                • #53
                  Originally posted by smartysmart34 View Post
                  So do I under Opensuse 11.2/64. Can anybody confirm that in that case - even though the different versions show wrong values - the driver is up to date?

                  Regards,
                  Martin
                  the driver is up to date and to proove this is easy.

                  10.8 do have litle blackwindows bug 10.9 do not have black windows bug.

                  10.8 can handle the standart 32 bit flash in 10.04 kubuntu and the 10.9 goes down if you switch from fullscreen to windows mode only strg+alt+f1 and after that strg+alt+f7 brings the screen back.

                  every FGLRX in the past do have an 'bug-'signatur''' i can tell you the version number if i use the driver and watch the bugs and i can tell you some beta driver numbers to in that way ;-)

                  every fglrx is an individual!

                  Comment


                  • #54
                    catalyst 10.9 still has a terrible slow blur effect:
                    http://www.phoronix.com/forums/showt...736#post141736

                    Comment


                    • #55
                      Originally posted by Qaridarium View Post
                      10.8 can handle the standart 32 bit flash in 10.04 kubuntu and the 10.9 goes down if you switch from fullscreen to windows mode only strg+alt+f1 and after that strg+alt+f7 brings the screen back.
                      Are you telling me, that if flashplayer / the screen hangs I can go to the text-console by pressing strg*alt+F1?
                      On my 64-Bit Opensuse 11.2 with desktopeffects/compiz enabled this caused the system to hang under 10.8 and even before.

                      I would be glad as hell, if this finally worked again. With 32-Bit suse all was fine. After changing to 64-Bit the system hung up (but only with 3d-Desktop (compiz) enabled).

                      Comment


                      • #56
                        Originally posted by smartysmart34 View Post
                        Are you telling me, that if flashplayer / the screen hangs I can go to the text-console by pressing strg*alt+F1?
                        On my 64-Bit Opensuse 11.2 with desktopeffects/compiz enabled this caused the system to hang under 10.8 and even before.

                        I would be glad as hell, if this finally worked again. With 32-Bit suse all was fine. After changing to 64-Bit the system hung up (but only with 3d-Desktop (compiz) enabled).
                        Failing to delete the /etc/ati folder often leads to wrong version number showing in CCC.

                        To get the right version numbers in CCC:

                        1. Uninstall old driver.
                        2. delete /etc/ati folder
                        3. install new driver

                        Comment


                        • #57
                          Shouldn't the rpm-package generated by the installer take care of this by itself?!?

                          Comment


                          • #58
                            Originally posted by Fluggo View Post
                            Failing to delete the /etc/ati folder often leads to wrong version number showing in CCC.

                            To get the right version numbers in CCC:

                            1. Uninstall old driver.
                            2. delete /etc/ati folder
                            3. install new driver
                            It would seem that deleting the files in /etc/ati doesn't get it to report the correct version number. I'm assuming that deleting the directory as well shouldn't make much difference cause 10.9 has populated /etc/ati again but still reports old version number.

                            Comment


                            • #59
                              Originally posted by smartysmart34 View Post
                              Are you telling me, that if flashplayer / the screen hangs I can go to the text-console by pressing strg*alt+F1?
                              On my 64-Bit Opensuse 11.2 with desktopeffects/compiz enabled this caused the system to hang under 10.8 and even before.

                              I would be glad as hell, if this finally worked again. With 32-Bit suse all was fine. After changing to 64-Bit the system hung up (but only with 3d-Desktop (compiz) enabled).
                              there is another way to get a better flash feeling in 64bit just stop using the 32bit flash and get the 64bit flash

                              yes its not 'stable' but more stable than the 32bit flash in 64bit linux.

                              Comment


                              • #60
                                Originally posted by mugginz View Post
                                It would seem that deleting the files in /etc/ati doesn't get it to report the correct version number. I'm assuming that deleting the directory as well shouldn't make much difference cause 10.9 has populated /etc/ati again but still reports old version number.
                                only the 2D version number is right... this is a known bug just ignore it!

                                Comment

                                Working...
                                X