Contribute
Register

My GPU cannot sustain high load and shuts down due to overheat. Any suggestions?

Status
Not open for further replies.
Joined
Aug 2, 2012
Messages
169
Motherboard
Gigabyte X299 Designare EX ver1
CPU
i9-9960X
Graphics
RX Vega 64
Mac
  1. MacBook Air
  2. MacBook Pro
  3. Mac Pro
Mobile Phone
  1. iOS
Hello guys,
I'm using my Gigabyte - Radeon RX Vega 64 8 GB GPU.
My work frequently requires me to load my GPU, sometimes for prolonged period of time. However, leaving my GPU running at higher load, within a short period of time, gets the double GPU fans go into very high rpm mode, and within a few minutes, forces my Hack to reboot.
Until just recently, I didn't know what exactly makes my computer into force restart, and I discovered it by accident, as, after a normal restart, I couldn't start again, as soon as I see gray apple background and a white progress bar, forces to reboot again.
I opened the computer, and the GPU was hot as an oven.
I used my in-room AC, and directed the air right into the opened front panel. After a few minutes, I was able to proceed with my work. But, this is really bothersome. Any idea what I could try?

I considered buying a GPU-water-block, and drive the water tubes outside of the computer case, but I think this is silly. Maybe there's some setting that I could raise threshold of a GPU temp, to allow the unit to continue function? Would that even be safe?
 
How powerful is your power supply? The Vega cards can suck down a lot of power...

You can try adjusting the power play tables...
 
The Vega cards can suck down a lot of power
Hello Pastrychef, wow, you blew my mind with your answer.. seriously :) Because, while this makes a lot of sense, I've never thought about this. Now this makes me start thinking that it is possible that in my previous build I've had before the current one, I've had a similar issue, and I could not figure out the reason for it to shut down periodically. Maybe all I needed was to swap the PSU.

Anyhow, this is the unit that I bought, I thought I carefully calculated all the individual pieces' demands:
1630864053157.png
1630863351877.png

CPU+Motherboard+GPU


In any regard, I'm using 750W version, though this Antec design comes in 850w as well. Do you have any idea how much I need to allocate for my GPU in a high load scenario? I realize that just jumping to 850 might be a quick fix, but what if I need more power than even that?

I also partly relied on this article to double-check the accuracy of the GPU's consumption:
But I wonder if this was always my weak spot.

Please help if you have any recommendations,
 
Hello Pastrychef, wow, you blew my mind with your answer.. seriously :) Because, while this makes a lot of sense, I've never thought about this. Now this makes me start thinking that it is possible that in my previous build I've had before the current one, I've had a similar issue, and I could not figure out the reason for it to shut down periodically. Maybe all I needed was to swap the PSU.

Anyhow, this is the unit that I bought, I thought I carefully calculated all the individual pieces' demands:
View attachment 528343View attachment 528341
CPU+Motherboard+GPU


In any regard, I'm using 750W version, though this Antec design comes in 850w as well. Do you have any idea how much I need to allocate for my GPU in a high load scenario? I realize that just jumping to 850 might be a quick fix, but what if I need more power than even that?

I also partly relied on this article to double-check the accuracy of the GPU's consumption:
But I wonder if this was always my weak spot.

Please help if you have any recommendations,

750W should be fine. The most I ever saw my system pull was about 550W. This was with a Vega 56 (flashed to Vega 64 firmware) and heavily overclocked. Of course, this is assuming your power supply is performing up to spec...

The next thing check is if your Vega is running too hot and causing shutdowns.

Try using the VGTab tool to under clock and/or undervolt the Vega and see if it helps with stability.
 
Last edited:
Um, aside from the PSU, but have you considered re-pasting your GPU's thermal heatsinks?
Dried up used thermal paste on that is also likely to cause overheating.

As to helping with the cooling, I think adding larger extra case fans will also help in this scenario.
I have about 6 in my Z490 case - 4 x 120mm fans on the top and rear and 2 x 140mm at the
front and they keep my setup relatively cool. My RTX 3070 GPU rarely goes above 70 degrees C.
 
Status
Not open for further replies.
Back
Top