GPU goes from 30 degrees to 60 degrees Celsius

EternalRest180

Honorable
Nov 29, 2013
14
0
10,510
Graphics card goes from 30 to 60ish within 30 seconds of starting any video game and crashes giving me a random solid color screen and the other screen goes white. However, it is the only thing that gets anywhere over 40 degrees. I have a Nvidia GTX 780 GPU(idling at 32 degrees celsius), Intel i7-4770k CPU (liquid cooled and idling at 39 degrees celsius), I have 750w power supply, and I have a ASROCK z87 extreme MoBo.

The Gpu isn't dusty and neither is most of my machine. This just started happening yesterday while playing World of Warcraft after owning the computer since December of 2013. I will say that I was manually throttling the GPU fan. When it got to about 57 degrees celsius and I throttled the fan to about 70% the screen would just flicker random solid color strips across the screen. If I throttle the GPU fan to 100% it seems like it works like normal. Anyone have any ideas?

Update #1 - I just started up Overwatch to check my GPU's temp again and it only made it to 42 degrees celsius before it started flickering and wanting to crash.
 
Solution
As I first said, unless the temp reading is wrong, it doesn't seem like it's the problem. The thermal paste idea is only if it heats up too quickly.

I really don't know what's happening to make it crash, my only suggestion is to increase the voltage slightly. There's a good chance this won't help, in which case the GPU may be on the way out.....sometimes they simply fail.
It can't be an overheating issue, unless the temp reading is wrong.
I was going to suggest replacing the thermal paste and checking the cooler is seated properly, as this may be the cause of the rapid temperature rise.

The flickering and wanting to crash can be caused by too much heat and also too lower voltage for it's clock speed. Did you play around with the clock or the voltage at all? Increasing the voltage may help stabilize the GPU but it will also make it hotter.
 


I used MSi to decrease the memory and core clocks by 25 but it didn't stabilize the system at all. Not really sure how to mess with the voltage. Also are you saying to check the thermal paste of the graphics card? I touched the circuit board of the GPU when got up to 60 degrees and it was pretty hot.
 
If the GPU heats up in a matter of seconds, it usually means the heat isn't transferring to the cooler as quickly as it should. This is the result of poor contact between the GPU and cooler, so new paste and reseating the cooler would help.
If you haven't changed the voltage, it still should be good. I've never used MSI to change the clocks or voltage (always used other programs that come with the card), but it should be a simple job to do both.
 
The GPU isnt even making it to 45 degrees anymore with crashing the computer. Im going to get more thermal paste to add to the gpu. I also did a clean uninstall and reinstall of the nvidia drivers. Does it seem like its still poor thermal paste? I looks like it still has a decent amount of paste on it.
 
As I first said, unless the temp reading is wrong, it doesn't seem like it's the problem. The thermal paste idea is only if it heats up too quickly.

I really don't know what's happening to make it crash, my only suggestion is to increase the voltage slightly. There's a good chance this won't help, in which case the GPU may be on the way out.....sometimes they simply fail.
 
Solution