Question GPU Problem? Gaming PC glitches, freezes, and crashes

Sep 7, 2022
5
1
15
Hi, I recently bought a gaming PC and I'll include the specs below. After playing video games for about 2 hours with no issues, my gaming pc starts doing this weird freeze thing and I would have to shut down the entire PC and restart. I don't think it's a temp problem, but please point me towards the right way because it's a super annoying issue to deal with since I just got the PC! Thank you for any help!!

Pics of Freezes and Temps

Specs:
Zotac 3090 Trinity OC
AMD Ryzen 9 5950X
Noctua NH-D15 CPU Air Cooler
ASUS TUF Gaming X570-Plus
Corsair Vengeance LPX 32GB
Corsair RM850x 80+ PSU
WD_Black SN770 1TB M.2 SSD
Corsair 4000x RGB ATX
 
Was this purchased as a used machine or was this a legitimate "new" purchase with warranty, etc.?

Because you need to either return it if you can, or try to RMA the graphics card if you can't. You might try different display cables first, but that really looks like corruption/artifacting from a bad card or some part of a bad card to me.

Have you tried doing a clean install of the Nvidia drivers using the DDU?
 
  • Like
Reactions: vinvin2023
Sep 7, 2022
5
1
15
Was this purchased as a used machine or was this a legitimate "new" purchase with warranty, etc.?

Because you need to either return it if you can, or try to RMA the graphics card if you can't. You might try different display cables first, but that really looks like corruption/artifacting from a bad card or some part of a bad card to me.

Have you tried doing a clean install of the Nvidia drivers using the DDU?

I guess it would be considered as a used machine, and warranty is probably voided. I've tried hard resetting the PC, but haven't tried clean install of the drivers yet. I'll give that a go, thanks!

It's just super weird because I tried playing again today, and for 2 hours it works perfectly fine, but then it starts doing that weird freeze.
 
Usually if it works fine for a while and then after a while it starts freezing, it's because either something is overheating or something is getting hot and once warm something starts failing, like bad solder joint, VRAM problem, something.

I'd recommend you download HWinfo, install it, run it, choose "Sensors only" and uncheck "Summary" and then monitor your CPU core temps and graphics card temperatures.

What case do you have?

How many case fans and where are the located, exactly?

What orientation, in (intake) or out (exhaust) are they oriented as?

Might also try taking off the side panel and running it like that to see if you still freeze up after a while.
 
  • Like
Reactions: Roland Of Gilead
Hey there,

Make sure to update all drivers, including GPU, chipset etc. Use DDU to uninstall the GPU driver (in safe mode), and reinstall it fresh. Lets rule out a software issue first. Your chipset driver will include the most recent power plan for your CPU, so we can rule out any issues with that.

In terms of the artefacts, yes, they look odd. Could be from mining, but hard to know if it's mem related or the cores.

The 3090 is a strong GPU (Nice :) ). However, because it's so powerfull it has very high transient power spikes. At certain points while gaming the load on the GPU really whacks up, and so does the power. The 3090 can have power spikes in the region of 550w and sometimes a little higher (this just for the GPU). Whilst the spikes only last for a split second (20ms in fact) they can trip protections on your PSU. I'm not saying that's what's happened here (because the symptoms are different), but just something for you to note down the line. The RMX are very capable of dealing with transient power spikes in general, but if you experience any shut off's/restarts, then you might look at a 1000 RMX, which gives plenty of headroo for the pwer spikes.

Regarding the freezes, you mgiht look at testing your ram (outside of windows) with memtest86+. Run it for 4 passes, and as long as there are no errors, then we can rule out your ram too.

Finally, what bios are you running? See if there's an updated version, as this may solve some bugs or instability with your system.
 
Spikes don't usually cause that kind of artifacting. They cause other types of issues. But it's a possibility because we know that the RTX 3090 can pull well over 850w when it spikes so freezing is certainly a possibility and if those artifacts are only happening simultaneously with the freezing and aren't like that all the time, or even most of the time, then it's completely possible that the PSU is all or part of the problem especially if the card is pulling more power as it gets hotter.
 
Having or finding another PC that can support an RTX 3090 other than your primary system, is probably not going to be the easiest thing in the world for most people to do. Few people are going to have secondary systems with a capable enough power supply in it or be able to find somebody with a system that DOES have a capable enough power supply in it that is going to be willing to let you put a graphics card that might be questionable in there and put their own hardware at risk. But, if you do or you can, great.

Better would be finding a different graphics card to try in their own machine like you said first, as that is going to be a lot more feasible, however it may not tell you anything if it's unable to push the issue with whatever is causing the problem.

Still, they are viable options among other options.
 
  • Like
Reactions: Roland Of Gilead
Sep 7, 2022
5
1
15
Hi all, thank you so much for the responses! I've been super busy with work so i haven't check back on this forum. However, I used DDU and fresh installed nvidia drivers the other night and so far there hasn't been any freezing. The artifacts also happen/show right as the pc freezes, so during normal usage everything looks fine and temps are all good. I've been monitoring temps while playing and CPU gets around 70, GPU around 60, and VRAM around 80. I've also undervolted the GPU to potentially help out with the spikes. I'll continue to see if anything happens now after using DDU. Thanks so much again!!
 
Yeah, don't undervolt. That makes problems worse usually, not better. You seriously affect stability when you underclock/reduce voltage. Underclocking should only be undertaken by experienced overclockers in most cases, who know exactly what they are doing and why. I realize that there are tons of "personalities" and "experts" out there who tout the benefits of underclocking, and there CAN be benefits in SOME situations, but same as when you overclock, and maybe even MORE so, if you are going to do either of these things you need to do the VERY intensive in depth testing to ensure the system is still stable.

And that does not mean run some games or "well, I didn't get any blue screens". Especially with CPUs, but also to some extent with graphics cards, undervolting can cause very real stability issues. Micro errors aren't such a big deal with graphics cards because they aren't changing and saving data for the most part, but they can certainly cause problems in game or application usage. There are other considerations as well. As mentioned, if you can't run a piece of hardware at it's stock configuration then you are lacking another piece of the puzzle whether that is better cooling or a better/bigger power supply.
 
Sep 7, 2022
5
1
15
Hi again, I changed back from undervolting to default and for the past 2 hours it's been fine. GPU goes up to 70 and VRAM goes up to 90, but otherwise no freezes or instabilities. I'll update if I have any future freezes on the current settings since after fresh install of GPU drivers. Thanks again!
 
  • Like
Reactions: Roland Of Gilead