Game crashes when using an old, stable card that has been swapped out for a new one.
Game crashes when using an old, stable card that has been swapped out for a new one.
Hello! Recently I've been having trouble with an OC'd (overclocked) GPU in games. My computer specs are EVGA GTX 1070 FTW, i5 6600K running at 4.5GHz, 16GB of RAM, and the driver is version 452.06. Last month I had to reinstall my operating system, which was Windows 10 x64 Home edition. Before that, everything worked perfectly. I used OC presets from MSI Afterburner with core speeds set to +150MHz, power at 112%, and temp allowed at 88C. I never needed to increase voltage because the preset was already right for my system. With this setup, games like RDR2 and Witcher 3 ran smoothly with great frame rates when I used the OC presets. After reinstalling Windows 10, I stopped using those games and didn't do any overclocking on my GPU.
Recently, I wanted to play RDR2 again and decided to use the same preset from before in MSI Afterburner. But now, whenever I launch any game with an OC preset—even just increasing the core frequency slightly, like 20-50MHz plus—after loading into a game and starting gameplay, it crashes with errors related to the GPU. RDR2: after loading, there were lots of visual glitches, and in 1-2 seconds I crashed completely with ERR_GFX_STATE error. Launched World of Warcraft using the same preset, it crashed right away after loading with "Error 3D-accelerator." This problem happens for both driver versions (452.06 and 451.67) and even if HAGS is on or off in Windows 10. Has anyone else had this exact issue? Can someone explain what's causing it? Any help with fixing the problem would be appreciated! Thank you so much in advance!!!
But without turning OC manually, the graphics card works great with no errors, no freezes, and smooth frame rates. However, if you boost the core frequency to 50MHz or higher, it starts crashing. Before I even reinstalled the operating system, this never happened before... My GTX 1070 runs at 1976Mhz without any manual tweaking when it's cold (I'm from Russia 😊), and it gets even faster on its own. The GPU didn't overheat more than 70-72 degrees Celsius; maybe once in Furmark it hit around 76 before the fans kicked in, but that was years ago... How do I know if it's actually dying? I haven't tried newer drivers yet, so I'm going to try them today after work. Please don't panic about a GPU failing just because this happened.
It is quite likely that your character's graphics card isn't working right anymore because of all the tweaking, and it won't stay steady unless you use the default settings.
I'm gonna try installing the old DDU and running my graphics card to check if things are better. Nvidia said that new drivers might cause this problem and even said it could happen with some specific models like HAGS.
Well, now that HAGS is on, things look better. After setting DDU and OC to preset plus 100 MHz with a power of 112% and a temp around 88 degrees Celsius, everything works fine without glitches, stuttering, or bad artifacts. I see more FPS than before. If I push the clock speed higher to +130 MHz, some artifacts start appearing. At +140 MHz there are lots of artifacts even if I try to lower voltage via Afterburner (it only goes up to 2% above max, like 1.062V) and at the default EVGA OC setting too. With a stable preset clock of +100 MHz, the GPU works fine on 2088 MHz because the base EVGA OC is around 1978 MHz. The temperature stays under 70 degrees Celsius (usually between 66 and 68). I could try going to +120 MHz but don't want to stress the GPU too much. I'm not sure why, but I think before my OS reinstall even at +150 MHz, with memory heat around 70-72 degrees Celsius, the GPU didn't exceed 2050 MHz automatically and just dropped back to 50 MHz on its own.