Hello there, so i’ve had this problem for like a month now and i’m at a complete loss.
When i play a game, usually a graphics intensive game like CoD, For Honor or Fifa, after some time my PC screen goes black, GPU fans go full speed and sound cuts off. The only thing i can do is hold the power button to shut off.
What i already did:
Run a windows memory test (nothing found).
Do a DDU and clean GPU Drivers install (didn’t work).
Do a Bios update (didn’t work).
Take out the GPU, clean it, the cables and reseat it. (Somehow worked, pc could go longer without crashing but ultimately crashed).
Use GPUTweak 3 to lower GPU power to 85% (worked for a while but i had to lower graphics or it would crash, but now it crashes again).
Can someone please help me, idk what to do.
Hey there,
Check windows event log, usually there is some info.
Use some temperature monitoring software coretemp/gpuz make sure to use one that does create logs. Start it , reproduce the black screen and then after restart check the temperature logs. Maybe it is overheating problem.
Other ideas I have:
-
Create a bootable linux usb and run some cpu/gpu stress test on that see if its stable there. If it is stable you know its your current ( I presume ) Windows install is the problem.
-
Start Swapping components, depend where you live you can order new hardware and return it free of charge for testing purposes. I would start with the Power Supply, test if error still occurs. If it does try a new gpu, maybe you have an old one laying around or borrow one from a friend.
1 Like
Please specify your hardware.
First thought would be the PSU though.
1 Like
i have a corsair rm1000x, asus tuf 3080ti, asus rog strix b550e gaming, gskill trident z rgb 3600mhz 2x16, samsung 980 pro 2to nvme ssd and a ryzen 9 5900x
What case/chassis?
Also, is the firmware for the NVME SSD up to date? You should be able to check with Samsung Magician.
event viewer doesn’t show specific error except some metadata staging error for container, i already did some stress tests, heaven benchmark crashes approximately a minute into test, usually as gpu reaches 70c, but an OCCT test doesn’t crash even with gpu at over 72c.
sadly i can’t swap components as it’s the first pc i buiilt in 2021 so i don’t have spare parts nearby
2 Likes
i have a be quiet pure base 500dx
i built it in 2021 with brand new parts, worked perfectly fine until last month
Did you try a driver rollback? SOMETIMES that helps, otherwise, @Draaksward could be right about the PSU.
yup, did a driver rollback, then ddu then clean install and still no change, i honestly think it might be a psu issue too, seems to happen during a powerspike, i lowered my gpu voltage and power through gputweak 3 and no crash
PSU isn’t necessarily an easy swap (because you HAVE to use the specific brand and model cables) but it’s the least expensive option to buy another than anything else in the system if it’s been running fine for a few years already.
Last question that I don’t see asked yet, have you done any overclocking to anything at all?
nope, never overclocked, i agree tho, i’ll swap for an atx 3.0 psu so it is compatible with a 4080 if power supply isn’t the issue
1 Like
Maybe look at 3.1 instead, for even more future compatibility. I’ve personally never had a problem with SeaSonic.
Adding on, also buy from somewhere that will allow you to return it if the PSU isn’t the problem. You may have to pay a re-stocking fee, but that’s usually between 5-25% (depending on location) of the cost of the item, but at least you’ll have ruled that component out.
oh i’m buying from amazon france don’t worry, and i’ll keep the psu anyway thanks
1 Like
Is it plugged into a power socket or a power bar?
Also possibly sounds like overheating to me since it doesn’t happen on anything else that isn’t pushing the system from what I read.
Try everything on full bore fan speed and try again and monitor the heat from time to time.
If they are all okay, check the caps on the motherboard since that is quick and if good then the above first question I would need to know the answer
What does electrical engineering have to do for a socket or a power bar and even checking caps don’t take much at all.
72 is hot even within specs. 72c just think about that! Mine gets issues at 70 and above(3070) so thought to throw that out there just in case.
Let’s just try to help the one in trouble
I have a 3080 Ti myself, just like OP. 70-ish is absolutely normal under a full load within an ATX chassis. That is even accounting for a 28° Celsius ambient temperature. I literally just did testing for 3rd party thermal pads.
Just because YOU MIGHT happen to know about capacitors and other SMD parts absolutely does not mean that everyone else does.
Rule of thumb when trying to help anyone with anything; ask questions and assume ignorance.