Question Recent PC build crashing in all games to desktop

[DayRahn]

I recently built a new pc a few days ago and games crash in a seemingly random timeframe, they will crash to desktop, giving no message and instantly cutting off audio and visual feedback once the crash occurs, as if I was alt-f4ing. One of those blue screens I read the error for and it was “Hyper_Visualization Error” . I have XMP disabled, Rebar OFF. I have downloaded the latest nVidia game ready driver aswell. I’m very confused as I have never built a PC and used a laptop before, never really having any of these issues, and when I rarely did they would give me a clear readout of what went wrong and then have an easy fix. Does anyone know where to start? The GPU rarely goes to 80C and the CPU rarely over 45C, Windows Memory Diagnostics Tool came up with nothing aswell.



CPU: I7 12700K

CPU cooler: MSI Mag Core 360 v2

Motherboard: Asus TUFF Z690 Plus Wifi LGA1700 Bios ver 2802

Ram: Corsair Vengeance 5600Mhz

SSD/HDD: MSI Spatium 2TB PCIE.4

GPU: NVIDIA RTX 3080 FE (this was bought used, even with a -200mhz the games still crash at similar rates, though 3d benchmark is more stable)

PSU: EVGA SUPERNOVA 850W 80Plus Gold is brand new

Chassis: Corsair 4000D Airflow

OS: Windows 11

Monitor: Samsung Odyssey G5 27

 

[punkncat]

Is this only happening in games? For instance, if you ran Furmark on it, will it crash then as well?

If you take out the graphics card and run your desktop on the inegrated graphics does the PC remain stable? I mean, understanding that you can't run games like that, just curious.

Have you already gone in a triple checked every connection and each cable within the connections? Nothing loose? Nothing pushed backwards out of the connector?

Is Event Viewer showing you anything else as far as codes?

When you built this system, did you install all the base level drivers from the motherboards website? Are there any available BIOS updates that pertain to the system*?

*it would not be a particular good idea if this shutdown is truly random and even occurring on the desktop as a power cycle in the midst of a BIOS update can brick a board

 

[DayRahn]

Is this only happening in games? For instance, if you ran Furmark on it, will it crash then as well?

If you take out the graphics card and run your desktop on the inegrated graphics does the PC remain stable? I mean, understanding that you can't run games like that, just curious.

Have you already gone in a triple checked every connection and each cable within the connections? Nothing loose? Nothing pushed backwards out of the connector?

Is Event Viewer showing you anything else as far as codes?

When you built this system, did you install all the base level drivers from the motherboards website? Are there any available BIOS updates that pertain to the system*?

*it would not be a particular good idea if this shutdown is truly random and even occurring on the desktop as a power cycle in the midst of a BIOS update can brick a board

It happens in 3D Mark aswell

I haven’t tried to run things on the IGPU yet

The cables are all checked and properly inserted (could that cause a crash to desktop? The pc is perfectly stable immediately after I see the desktop, not even a flicker.)

I do not understand event viewer and dont know what to look for

All drivers and Bios updates are installed I believe, and device manager says I am running the latest firmware for all

The desktop has remained perfectly stable for hours at a time, even when using apps and watching streams on discord. I did stream one of the crashes by streaming my screen, while the game crashed, the stream for others to see remained perfectly stable and didn’t close when the game did.

 

[punkncat]

With nothing to really back it up...but your comment stating the graphics card was used. Do you have another known working card available to you?

If this is a purchased "too long and mine now" situation, it may not be a bad idea to go in and check the thermal paste and pads on the graphics card.

 

[DayRahn]

With nothing to really back it up...but your comment stating the graphics card was used. Do you have another known working card available to you?

If this is a purchased "too long and mine now" situation, it may not be a bad idea to go in and check the thermal paste and pads on the graphics card.

I can use a friends later today, but if it helps the gpu never passes 78 degrees and the memory maxed out at 104, but a crash later happened when it was a cooler 101.

 

[punkncat]

I really am not well versed in how hot a GPU should be getting, but over 100* anywhere on the card seems extreme to me. I am assuming you are discussing the VRAM being that hot. Those are the parts that would be relying on thermal pads for proper cooling. Could it be that the previous owner opened it up to clean and "re", doing so with the wrong thickness pads?

This is pure conjecture on my part, just thinking out loud at this point.

 

[DayRahn]

I really am not well versed in how hot a GPU should be getting, but over 100* anywhere on the card seems extreme to me. I am assuming you are discussing the VRAM being that hot. Those are the parts that would be relying on thermal pads for proper cooling. Could it be that the previous owner opened it up to clean and "re", doing so with the wrong thickness pads?

This is pure conjecture on my part, just thinking out loud at this point.

I’m quite sure the previous owner grabbed this card out of a tomb and sent it directly to me. This card probably never got more than some spray air over a year ago. 3080’s do get very hot vram but I know it’s definitely not off the table that that could be the issue

 

[punkncat]

I did a quick search for "how hot can VRAM on a 3080 get". Supposedly they have a throttle limit at 110 with fail being 120. Supposedly Nvidia recommends up to 106 as normal. Many of the other posters in Reddit and whatnot are saying to keep it in the 80-90's max and when not achieving said, to put on "better pads".

Take with spoonful of salt based purely on source, but probably worth a read.

 

[DayRahn]

I did a quick search for "how hot can VRAM on a 3080 get". Supposedly they have a throttle limit at 110 with fail being 120. Supposedly Nvidia recommends up to 106 as normal. Many of the other posters in Reddit and whatnot are saying to keep it in the 80-90's max and when not achieving said, to put on "better pads".

Take with spoonful of salt based purely on source, but probably worth a read.

Is there a way I can have the VRAM work less hard to make it run cooler, that way if it still crashes I’ll know if it was the issue? I’d rather get another GPU and open one up and modify it.

 

[punkncat]

Unfortunately, I don't have any real hands-on experience undervolting and underclocking. It can be just as dangerous to the equipment as overclocking can be.

I also don't know much about mining rigs but am aware that in some cases the V-BIOS could be changed for better performance on the task at hand. Have you verified this is on the correct BIOS for 'gaming'?

 

[DayRahn]

Unfortunately, I don't have any real hands-on experience undervolting and underclocking. It can be just as dangerous to the equipment as overclocking can be.

I also don't know much about mining rigs but am aware that in some cases the V-BIOS could be changed for better performance on the task at hand. Have you verified this is on the correct BIOS for 'gaming'?

I can go and do that, I’m running a windows memory diagnostic currently

 

[punkncat]

For what it's worth, I purchased a 3070 FE just as stock came back in on the end of the mining craze and pretty much just before Nvidia quit production on the 30xx models. Mine started black screen errors, artifacts around windowed videos, strange color patterns. I contacted Nvidia about the warranty and they didn't even pretend like they were going to put up a fight they accepted it back so fast. I mean, it was GREAT customer service, but that skeptic in the back of my mind instantly considered there must be some ongoing thing with this generation of cards. I have nothing but a tingling spidey sense as proof, so...

 

[DayRahn]

For what it's worth, I purchased a 3070 FE just as stock came back in on the end of the mining craze and pretty much just before Nvidia quit production on the 30xx models. Mine started black screen errors, artifacts around windowed videos, strange color patterns. I contacted Nvidia about the warranty and they didn't even pretend like they were going to put up a fight they accepted it back so fast. I mean, it was GREAT customer service, but that skeptic in the back of my mind instantly considered there must be some ongoing thing with this generation of cards. I have nothing but a tingling spidey sense as proof, so...

I was thinking the same but I haven‘t had any visual artifacts in all the games I played, not even a screen flicker

 

[punkncat]

The strangest aspect with mine was that if I did a clean install of drivers it would almost always stop doing strange things for a day or two. That got to be less and less time to a degree that watching a video on YT ended up being like some action movie trailer outside of the playing video (the video itself played perfectly).

The card they sent back to me is working pretty good so far as I can tell. I have seen some very minor glitching but not frequent enough to say it was the replacement card.

 

[DayRahn]

The strangest aspect with mine was that if I did a clean install of drivers it would almost always stop doing strange things for a day or two. That got to be less and less time to a degree that watching a video on YT ended up being like some action movie trailer outside of the playing video (the video itself played perfectly).

The card they sent back to me is working pretty good so far as I can tell. I have seen some very minor glitching but not frequent enough to say it was the replacement card.

a game actually comes up with an error about vram allocation and not having enough, so we now know it is the gpu. The BIOS of it is an official NVIDIA one