2 Replies Latest reply on Jan 11, 2016 8:01 AM by pfannkuchen_gesicht

    GPU stuck in lower power state

    pfannkuchen_gesicht

      Hello,

      so recently a new issue appeared. Basically the GPU gets stuck in a low power state or drops to it while in game.

      What I mean by that is that the GPU suddenly clocks down or can't clock above 501MHz when playing games.

      Sometimes it's stuck from the beginning right when I started the game, sometimes it drops down to 501MHz while playing games.

      It's especially noticable because the games start stuttering.

      My PC specs:

      Intel i5-4690(4 x 3.5-3.9GHz)

      2 x 4GB G.Skill DDR3-1333 7-7-721

      ASUS Sabertooth Z97 MK II

      Club3D HD7970GHz RoyalAce(1100MHz core clock, 3GB 1500MHz VRAM)

      BenQ XL2730z(2560x1440@144Hz, connected via Mini-DP to DP cable)

       

      The issue happens in all games I play, namely Battlefield 4, Killing Floor 2, Quake, Counter-Strike: Global Offensive, etc.

      Temperatures are okayish, just recently cleaned the heat sinks. GPU usually 40°C idle and about 87°C while playing. CPU is at 32°C idle and 45°C while playing.

       

      I could be wrong, but I think this started with my new monitor. I previously had a Samsung SyncMaster P2450 1920x1080@60Hz which was connected via DVI.

       

      EDIT:

      I just changed the thermal grease and except for a few degrees celsius less there is no change in behaviour. Just tried it in BF4 again and maybe 10 minutes into the game it clocked down to 501MHz while the gpu was at around 86°C

        • Re: GPU stuck in lower power state
          amdmatt

          Have you tried a Clean Install?

           

          What AMD driver are you using?

            • Re: GPU stuck in lower power state
              pfannkuchen_gesicht

              Hello,

              yes, I tried a clean install and also a downgrade.

              I had the Crimson 15.12(.1?) with hotfix installed and now downgraded to 15.7.1 again.

              Anyway, the type of the issue changed, as it turned out my method of measuring the temperature was flawed.

              It does get stuck in that state because it overheats now, it reaches 100°C and then clocks down to 501MHz.

               

              This is really strange however because it all worked fine up until a few weeks ago. So what I did now was to clean the heatsinks again, which didn't help.

              My next step was to change the thermal paste. That didn't help either so I tried that again with only marginal improvements, but still overheats eventually.

              I really don't know what happened that it overheats now all of a sudden.

              The fans are spinning fine, thermal paste is good, ambient temperatures are as usual(about 20°C)...

               

              The only thing that changed really is the monitor and thus a higher resolution(1920x1080 -> 2560x1440), but I'm under the impression that such a change in screen resolution shouldn't overheat your graphics card.

               

              Is there anything at all I can do? Right now I'm considering to buy a new heatsink from Arctic, but I really would like not to if avoidable.

               

              Anyway,  what do I do about this topic now? Edit it?