GPU BSOD

[PopeSixtusV]

Hi all,

 

I've got a Zotac Trinity 3080 - all out of box standard, no overclocking - which has some very weird thermals. Hwinfo reports the overall GPU temp and the mem junction to be pretty acceptable, but the hotspot temps get up to 100-105C under what I would consider to be a fairly mild load. I've done some digging and learned this generation of Zotac cards was sort of known for bad thermals all around, albeit more often bad thermal pads resulting in high mem junction temps. At any rate, that's obviously an issue and I plan to fix it; but it's only relatively recently come to my attention how bad it was due to another issue. I've been having intermittent but persistent BSODs and even just crashing straight to black with no BSOD. The only errors I get from it in event viewer are just the kernel power event ID 41, and another error simply telling me it failed to generate a log file for the crash. It got so bad that it even killed my PSU, which initially led me to believe it was simply a power issue, but then it kept happening after replacing the PSU (with a higher wattage version, just in case that was the issue).

 

So it feels like a bit of a dumb question, but is the GPU thermal issue likely to be the root of the BSOD problem? Would it really be that severe as to even kill the PSU? Even with the ridiculous hot spot temps, I never noticed any ridiculous power spikes from it under load; but maybe just too many sudden crashes while I was trying to diagnose the problem is what killed it?

 

Other system specs:

R9 5900x

kraken x53 aio

rog strix b550-i

2x16gb ripjaws 3600 ddr4

EVGA 850GM sfx psu (old one was corsair SF750)

currently in Hyte Y60 case - used to be in a LianLi TU150 but I swapped cases hoping the larger case might help the thermals, plus I just liked the Hyte a lot

 

[191x7]

Do you have the latest vBIOS on the card?

https://www.zotac.com/us/files/download/by_product?p_nid=956598&driver_type=All&os=All

 

 

[SavageNeo]

1 hour ago, PopeSixtusV said:

but is the GPU thermal issue likely to be the root of the BSOD problem?

no. that is hardware issue somewhere else, usually RAM or Video card error, but could be many things.

 

Thermal issue is usualy a blackscreen without pc shutting down, or the program crashing.

1 hour ago, PopeSixtusV said:

Would it really be that severe as to even kill the PSU?

no. sf750w is high quality, and 3080 killing it is ultra rare, that makes me think that other factors might have been in the play, liek dirty power etc.

 

hot spot temps are good up to 110C, so if they are not hitting that, do not fix it.

 

update motherboard bios and Vbios

 

you could try with only 1 stick of ram without XMP on, and try if that works.