Shut down problems

Associate
Joined
19 Feb 2018
Posts
76
Hi all, having issue with my pc just restarting for no reason during gaming. Started at once a week ish but now at least once in a 2 hour session.
I’m pretty sure it’s not temp related, have them displayed in front of me and keep a regular eye on them. Cpu generally in low 70s gpu normally high 70s and h100i water maxes around 38.
Psu is around a year old so would like to think that’s ok but can’t be certain. Only thing I notice is when it reboots quite often the aio and lighting node are missing from Icue, not sure if that’s related.
Bios and drivers are all up to date.
So far it has only done it during warzone 2 but that’s mainly what I’m playing at the mo.
Anyone got any recommendations on what I can check?
Was thinking about leaving it on a stress test? Any recommendations on a program to stress all components at once?

Specs are
I9 10850k
Msi tomahawk board
Gigabyte 3080
H100i aio

Any help would be much appreciated
 
Build is prob around 2 years now. Nothing changed at all. Everything has been fine. Only thing I’ve noticed is recently before the shutting down started, it was slow to load start up programs. Ie Icue, discord. Could take a few mins.
Gpu has separate cables
 
As a rule of thumb, graphics cards can sink a peak current from the 12V rail of 2.5 times the TDP current, but it totally depends on the workload. If you have noticed that some games bring it to a higher peak temperature than other games you'll see what I mean. I think a 3080 is specified at about 320W TDP? So the effective peak power will be about 800W, in very short duration spikes. What with the CPU and other power users, that could push it over the limit of the PSU.

Of course it could be something else. What's in the SEL just prior to reboot?
 
Hmm ok, that could sound reasonable. I assume there is no way of monitoring this? I guess I could under clock the 3080 and see if that makes any difference. I have now limited the max fps in cod to 100 and it hasn’t as yet rebooted so that could be it.

It’s strange how it has only just started and been fine for ages. Could a psu degrade in a year?
 
GPU power is directly proportional to clock rate. By capping FPS you are reducing the "need" for the GPU to boost its clock. You could underclock to check if the GPU is tripping the PSU but it would be the harder method: The GPU has internal boost clock control with the two primary inputs being temperature and power. The latter input is the best way to limit boost clocks. Since you have a 3080 you can use nvidia-smi (part of the driver) to set a power cap. I believe the command is "nvidia-smi -pl"
 
Back
Top Bottom