Primary GPU Crossfire Tempearture Problem

Toggle sidebar Toggle sidebar
S

SiuKingBon

Prominent
Jun 9, 2017
7
0
510
So I have two R9 290's in Crossfire and each have their own H75 watercoolers.
Under full load my first card hits up to 97C while the second card barely touches 60C.

I tried plugging my monitors into the second card to see if that would make it the "primary" GPU and the temperatures are better this way but the computer sometimes crashes unexpectedly and the framerate is clearly not being delivered.
I read that there might be an option in the BIOS to change which PCIe slot is the primary but couldn't find it on my Z77's BIOS.

Because of how the case I have (NZXT Switch 810) is setup, and the fact the the CPU is cooled with an H100i AIO, I don't have much room to get better airflow.

Anyone have any suggestions as to what I could try next? Is there a way to get the second PCIe slot to become the primary slot as that would probably solve the temperature problem?

MzLNTwC.png


 
Solution
Eximo
First off, check the pump and fan on your Primary GPU, I struggle to see how it could ever get that warm. The pump could have gotten stuck on an air bubble as well, so unscrew the radiator and take the GPU out and give a few rotations in a couple of different directions.

Typically you wouldn't be able to saturate the incoming air to the point where it could cause real temperature problems, but if the hardware checks out you could reverse the case flow.

Switch your GPUs to be intake and have the front panel be the exhaust, that way the radiators are only sucking cool air.
Sort by date Sort by votes
S

SiuKingBon

Prominent
Jun 9, 2017
7
0
510


The power limit hasn't changed on the GPU's. I've had them set to 0% this entire time. I'll give them a swap to see if its isolated to the one card though and see if anything changes.
 
Upvote 0 Downvote
Eximo

Eximo

Titan
Ambassador
May 27, 2010
25,856
4,189
139,340
First off, check the pump and fan on your Primary GPU, I struggle to see how it could ever get that warm. The pump could have gotten stuck on an air bubble as well, so unscrew the radiator and take the GPU out and give a few rotations in a couple of different directions.

Typically you wouldn't be able to saturate the incoming air to the point where it could cause real temperature problems, but if the hardware checks out you could reverse the case flow.

Switch your GPUs to be intake and have the front panel be the exhaust, that way the radiators are only sucking cool air.
 
Upvote 0 Downvote
Solution
S

SiuKingBon

Prominent
Jun 9, 2017
7
0
510




So I've tried out a few things and swapped the GPUs about. Tried changing the GPUs so that the top one in pci0 goes with gravity and pushes out the bottom radiator. The second GPU needs to push up towards the top radiator. Opposite of what I have in the original photo there.
The temps are now swapped, the GPU which is pushing downwards (pci0) is now around 50 underload and the one that pushes upwards is 90 underload. Something about having to push the liquid upwards is causing an issue it seems. Don't see any way around that though.
FfIagqe.jpg



EDIT:
Gave the radiators/pumps a few rotations and changed the direction so the top radiator is an intake, seems like the temperatures of both GPUs stay below 60 now. For some other unknown reason, the PC crashes under load now but increasing the power limit by 10% seems to have resolved that, not really sure what caused this to start happening now. Gonna have to keep running a few more tests but the temps look good now at least.
 
Upvote 0 Downvote
S

SiuKingBon

Prominent
Jun 9, 2017
7
0
510
EDIT:
Gave the radiators/pumps a few rotations and changed the direction so the top radiator is an intake, seems like the temperatures of both GPUs stay below 60 now. For some other unknown reason, the PC crashes under load now but increasing the power limit by 10% seems to have resolved that, not really sure what caused this to start happening now. Gonna have to keep running a few more tests but the temps look good now at least.

EDIT2 UPDATE:
Late update but maybe it'll clear some things up about this particular case. I ended up replacing both the H75s and the temperatures in both came back down to normal. The only problem now is that the PC crashes under load and it is happening due to the fact that during all this testing I must've damaged one of the GPUs. The system works fine with one of the GPUs but the other GPU now used either alone or in Crossfire causes the system to crash.

I might reflow the card in a last ditch effort to save it but at least we know that the problem is isolated to that card in particular.
 
Upvote 0 Downvote
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom