[SOLVED] GPU is overheating and not sure why

spacemanjupiter

Distinguished
Jan 8, 2009
52
0
18,530
I had a system with an 8700k, Noctua D15, 980 Ti Rog Strix, Aorus Gaming 5 mobo, 16 GB patriot viper ram, 850w PSU, and WIN10.

I simply took that hardware and put it in a different case that is very similar to the one it was in, but with a 750w PSU instead of 850. All the fans were set up the same, so very similar airflow. I never ever had any temp issues with the CPU and especially the 980 Ti Strix. The GPU always ran cool, never leaving the 60s or 70s under great loads, depending on the game, such as playing RDR2 for 4 hours at a time.

Suddenly, after moving all this into a different, but similar case, and a different PSU, the GPU is overheating very quickly and easily. Playing RDR2 or COD MW 2019 for just a few minutes puts it up to 80+ pretty quick and even shut the PC once. Other games will run fine and not over heat the GPU, and stay in the 60s or 70s. It's only with certain games.

Any suggestions on why this might be happening? I was considering that it may be a driver issue and may try older drivers this evening. Temps for everything seem fine except for the GPU. The GPU fans seem to be working as designed. Case fans are all working.
 
Solution
Can't share what these 2 chassis are?

Any suggestions on why this might be happening?
1)Paste coincidentally reached its limit when the switch was made.
2)A gpu fan(s) coincidentally broke when the switch was made. [Yes, I listed it anyway.]
3)Misuse of gpu anti sag brackets, such that it's putting too much leverage on one end of the gpu and screwing with the cooler mount; cold plate doesn't mate as well with the die as it once did.
4)The gpu sags more in the new chassis than it did in the old one. This can cause the cooler cold plate to not mate as well with the die and leave air pockets in between.
5)The new chassis isn't as airflow friendly. You didn't share what the 2 were... kinda pullin' at straws on this one...
6)Using a...

Phaaze88

Titan
Ambassador
Can't share what these 2 chassis are?

Any suggestions on why this might be happening?
1)Paste coincidentally reached its limit when the switch was made.
2)A gpu fan(s) coincidentally broke when the switch was made. [Yes, I listed it anyway.]
3)Misuse of gpu anti sag brackets, such that it's putting too much leverage on one end of the gpu and screwing with the cooler mount; cold plate doesn't mate as well with the die as it once did.
4)The gpu sags more in the new chassis than it did in the old one. This can cause the cooler cold plate to not mate as well with the die and leave air pockets in between.
5)The new chassis isn't as airflow friendly. You didn't share what the 2 were... kinda pullin' at straws on this one...
6)Using a riser card, but the gap between the gpu's fans and the solid side panel is narrower in the new chassis. [Again, don't know what what they are, so relying on guesses.]
7)External psu cables 'pulling' too much on the gpu, causing sag.

8)A really wild idea: gpu wasn't working correctly this whole time, until the swap happened, after which it's now 'fixed'. The gpu runs 80C+ in certain titles, but not all?
That makes me want to retract a number of the earlier points...
 
Solution

spacemanjupiter

Distinguished
Jan 8, 2009
52
0
18,530
Found the culprit. I took the GPU apart and replaced the dry as dirt thermal paste. Easiest thing I've ever taken apart. Now even RDR2 over hours of play won't push it over 73c. Most other games are well below that under load.