Announcement

Collapse
No announcement yet.

Mesa 19.2.3 Released With Many Fixes While Waiting For Mesa 19.3 In A Few Weeks

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

  • Mesa 19.2.3 Released With Many Fixes While Waiting For Mesa 19.3 In A Few Weeks

    Phoronix: Mesa 19.2.3 Released With Many Fixes While Waiting For Mesa 19.3 In A Few Weeks

    Mesa 19.3 is expected to be out around the end of November as the next feature update to Mesa3D, but in the interim Mesa 19.2.3 has been released as the newest bi-weekly bug-fix release...

    http://www.phoronix.com/scan.php?pag...9.2.3-Released

  • #2
    Anyone knows if Fedora 31 will ship Mesa 19.3?
    Some releases ago they decided to upgrade to a new Mesa major version during its life cycle.

    Comment


    • #3
      This new version is quite broken for me. Maps do not show up in games such as Civ VI and Imperator: Rome. I reverted to 19.2.2 and it immediately fixed the issue. Using Arch.

      Comment


      • #4
        Well, since my problem is under wine I'm not sure if it's Mesa or not. But just in case one of the mentors here can help I'm going to describe it anyway

        I'm running Manjaro Linux with a R9-390 GPU and a few months ago I noticed I couldn't run games at high settings anymore without weird random "sparkles" beginning to appear all over the screen. The sparkles persist even after exiting the game and Steam, and sometimes the entire systems locks up if I don't exit the wine game and steam fast enough. And once I've exited I have to reboot because a few sparkles will continue unless I do.

        At first I thought the problem was just No Man's Sky because it recently began to use Vulkan, but then I found the same problem with the original Crysis and a few other games. So long as the graphics settings for games are set to whatever their equivalent of "standard" is everything runs fine. But if I set them any higher the sparkles will appear. I spent a few days off and on when I first discovered it trying to figure out some common settings that caused it, but finally had to get back to work before I could.

        So does that sound familiar to anyone here, or has anyone heard of these symptoms before? Of course I also did a lot of searching on the Internet for a solution but was unsuccessful. Any help would be greatly appreciated.

        Comment


        • #5
          Originally posted by muncrief View Post
          Well, since my problem is under wine I'm not sure if it's Mesa or not. But just in case one of the mentors here can help I'm going to describe it anyway

          I'm running Manjaro Linux with a R9-390 GPU and a few months ago I noticed I couldn't run games at high settings anymore without weird random "sparkles" beginning to appear all over the screen. The sparkles persist even after exiting the game and Steam, and sometimes the entire systems locks up if I don't exit the wine game and steam fast enough. And once I've exited I have to reboot because a few sparkles will continue unless I do.

          At first I thought the problem was just No Man's Sky because it recently began to use Vulkan, but then I found the same problem with the original Crysis and a few other games. So long as the graphics settings for games are set to whatever their equivalent of "standard" is everything runs fine. But if I set them any higher the sparkles will appear. I spent a few days off and on when I first discovered it trying to figure out some common settings that caused it, but finally had to get back to work before I could.

          So does that sound familiar to anyone here, or has anyone heard of these symptoms before? Of course I also did a lot of searching on the Internet for a solution but was unsuccessful. Any help would be greatly appreciated.
          My old 260x acted similar. It would get random square blocks all over the screen until a reboot. AFAICT, it was due to high temps over prolonged periods of time and it effected both Windows and Linux (7, 8.1, 10, Arch, Manjaro, Suse, Debian...all of them).

          The first few years, it would take over an hour of gameplay for it to happen and that lead me to think it was game related....until I used MSI Afterburner and realized my GPU would operate at 93C and not thermal throttle down (which caused damage to my GPU). Linux was the same.

          As the years went on, it took less and less and less to get it to start glitching up once overheated and I was having to try more and more things to mitigate it from OC/UV to opening my case and using a box fan (seriously...93C to 84C...). It finally died in February after 7 years where it would get random square blocks when idling at the desktop at 36C combined with randomly locking up.

          They best ideas I have would be to check your GPU/PC for dust build up and air flow blockage, possibly lowering and undervolting your memory speeds a tiny bit (that really helped my R7 260x more than anything), and to undervolt the clock frequencies it as much as possible since that is what is working for my current RX 580 that gets the same high temp random square blocks and is directly why I mention undervolting RX 580s so damn much because it hasn't effected mine since March when I finally dialed in my undervolt+overclock settings.

          Undervolting the clock frequencies wasn't an option for my CIK enabled R7 260x on Linux at the time, but I have to assume it would have helped.

          Comment


          • #6
            Originally posted by skeevy420 View Post
            ....

            They best ideas I have would be to check your GPU/PC for dust build up and air flow blockage, possibly lowering and undervolting your memory speeds a tiny bit (that really helped my R7 260x more than anything), and to undervolt the clock frequencies it as much as possible since that is what is working for my current RX 580 that gets the same high temp random square blocks and is directly why I mention undervolting RX 580s so damn much because it hasn't effected mine since March when I finally dialed in my undervolt+overclock settings.

            Undervolting the clock frequencies wasn't an option for my CIK enabled R7 260x on Linux at the time, but I have to assume it would have helped.
            Oh wow, thanks for the reply. I hadn't even considered it might be a hardware issue. I have a dual boot system with Windows 10 so I'll install Steam and the games in W10 to see if it's a hardware problem. That would be a bummer though, I'd hate to have to go out and get a new GPU. I love my R9-390!

            UPDATE:
            Yikes!!! I tested my GPU with NMS on Windows 10 and it worked perfectly, reaching a max temp of 77C. So I found something to monitor the temp under Linux and discovered it went from around 50C to 90C within four minutes while playing NMS.

            And after some investigation I found there's been a GPU destroying bug in Linux since kernel 4.19 that causes the fans to remain at their lowest setting until the very end, in which case they rev to 100% right before the card dies. A good discussion about it is at https://bugzilla.kernel.org/show_bug.cgi?id=201539.

            So now I don't know what to do. The sensors command indeed shows the ridiculous temperature settings discussed at the above URL, mine being (crit = +104000.0°C, hyst = -273.1°C), but there doesn't seem to be any real solution. I'm really shocked.

            Oh well, at least I know now to be very, very, careful when running Linux or I might lose my $350 GPU. If I can't figure out a solution I'm going to have to go back to Windows 10, which would really be disappointing. I have no idea why the kernel developers are ignoring this devastating, hardware destroying, bug.
            Last edited by muncrief; 11-07-2019, 03:22 PM. Reason: Found out about a long time GPU destroying Linux kernel bug.

            Comment


            • #7
              Originally posted by muncrief View Post

              Oh wow, thanks for the reply. I hadn't even considered it might be a hardware issue. I have a dual boot system with Windows 10 so I'll install Steam and the games in W10 to see if it's a hardware problem. That would be a bummer though, I'd hate to have to go out and get a new GPU. I love my R9-390!
              Both of mine are/were MSI cards and it could simply be that MSI cards have crap thermals and/or unnecessarily high voltage or wattage settings due to their factory overclocks. There is a TON of headroom on my RX 580 which I assume is because I play my games at 1080p60 instead of 1440p FreeSync.

              The 260x, OTOH, didn't have that voltage/wattage headroom and needed clock speeds lowered; specifically, the memory speeds...they were really finicky and just bumping them up by 50mhz would cause the square blocks issue I was having on a non-compositing XFCE desktop.

              Comment


              • #8
                I have same issue as Med_

                I have a Vega 64 card. As it seems to be not just me I hope that big will be fixed soon. I'm not familiar with their bug tracker so I'll wait a bit until filling an issue.

                Comment

                Working...
                X