Announcement

Collapse
No announcement yet.

RadeonSI Primitive Culling Lands In Mesa 19.2

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

  • #11
    Originally posted by ihatemichael

    Yes, same problem with xf86-video-amdgpu.
    which GPU and desktop environment you have? because honestly i don't have any issue either with POLARIS or Cape Verde and neither with TAHITI before my older 7950 died, i also have archlinux running on some kaveri and kabini systems without issues.

    I tested Plasma / Gnome and Deepin and all is dandy on my systems, i'm guessing there is something in your particular GPU/driver version/distro that trigger those issues

    Comment


    • #12
      Originally posted by ihatemichael

      Yes, same problem with xf86-video-amdgpu.

      xf86-video-amdgpu and modesetting (they both use GLAMOR), so it makes no difference if I use one or the other.
      Alright.

      So all I can think of is trying TearFree if you have not yet, or maybe trying a different compositor.
      Of course it might just be an issue with the exact hardware you have. If so maybe the clock is too low during desktop time for what you need, I;m only on 1080p so I don't require much power there.

      Comment


      • #13
        Originally posted by jrch2k8 View Post

        which GPU and desktop environment you have? because honestly i don't have any issue either with POLARIS or Cape Verde and neither with TAHITI before my older 7950 died, i also have archlinux running on some kaveri and kabini systems without issues.

        I tested Plasma / Gnome and Deepin and all is dandy on my systems, i'm guessing there is something in your particular GPU/driver version/distro that trigger those issues
        I'm wondering the same things or if that GPU is going bad since I haven't had any issues with Polaris.

        I would get artifacts on my 260x, Bonaire, whenever it got really hot (on Windows or Linux) and when it started going bad after 7 years it would act up, but when the temps were normal and it was new, it ran without problems. Don't know why, but that 260x would not thermal throttle until it hit 94C and that's probably what killed it.

        Comment


        • #14
          Originally posted by geearf View Post
          If so maybe the clock is too low during desktop time for what you need, I;m only on 1080p so I don't require much power there.
          The 1080p part doesn't matter if it's one of those weird power management bugs. Simply forcing it to high should be enough to confirm or deny if that's the case. I dealt with it for a few months a few years back with my 260x.

          For me, whenever my cards have acted like that it was due to thermal issues so I'm curious if the card is running hot or not.

          Comment


          • #15
            Originally posted by skeevy420 View Post

            The 1080p part doesn't matter if it's one of those weird power management bugs. Simply forcing it to high should be enough to confirm or deny if that's the case. I dealt with it for a few months a few years back with my 260x.

            For me, whenever my cards have acted like that it was due to thermal issues so I'm curious if the card is running hot or not.
            Well I think I read somewhere people complaining that clocks were not high enough to drive a 4k monitor or maybe 2 monitors not sure anymore, so I thought maybe the size of the output could be related to the issue.

            Comment


            • #16
              Originally posted by ihatemichael

              I get frequent artifacts (corruption, invisible objects, etc) with the modesetting driver.
              What version do you use?
              Version 19.0.0 was broken and i had to change "AccelMethod" to "EXA" until i upgraded to 19.0.1
              Do you have a more detailed description or bug rapport?
              What graphics card is it that's that buggy?

              Comment


              • #17
                Originally posted by geearf View Post

                Well I think I read somewhere people complaining that clocks were not high enough to drive a 4k monitor or maybe 2 monitors not sure anymore, so I thought maybe the size of the output could be related to the issue.
                It effected my 260x for a little while and I only ever ran it at 1080p. It could be as simple as as the GPUs get more lower end, they're more likely to be effected by the DPM issues as the resolution drops.

                EDIT: Mine happened to have an easy, hacky fix for the few months it did effect me -- made a script that set it to high and then to auto.
                Last edited by skeevy420; 05-20-2019, 04:31 PM.

                Comment


                • #18
                  Originally posted by ihatemichael

                  You can talk all day about these things but the fact is that RadeonSI+GLAMOR continues being buggy. Intel (i965+modesetting+glamor) gives me zero issues.
                  That's fine, just not sure why you're talking about unrelated topics then.

                  Comment


                  • #19
                    Such complaints are useless without a proper description how to reproduce.

                    Comment


                    • #20
                      a RX480 here and also zero problems with glamor and nodesetting on slackware-current
                      maybe a hardware specific problem or a distro problem, some bad X11, glamor or whatever lib version?

                      Comment

                      Working...
                      X