[SOLVED] Games crash / Video driver crash

[Guille001]

Hi there, I have been having this problem for the last 5 days and need help.

I recently bought an M.2 SSD and installed it in my PC with no problem. I also removed the old thermal paste and put some new, cleaned all the fans and the heatsink, etc...I installed Windows 10 from a bootable pendrive and downloaded all the latest updates, drivers and my usual programs. I also updated my BIOS to the latest version. The only thing I kept is some files from my old computer, mostly photos and documents.

Everything worked fine before changing from my old SSD and HDDs to the new one, but now after just a few seconds of opening whatever game I get this error: ''Video driver crashed and was reset. Make sure your drivers are up to date. Exiting...''. Sometimes it even shuts down my computer with no error, the game just freezes for a few seconds, then everything goes black for some more seconds and it just shuts down. I fixed this by increasing the page file size, now it doesn't shut down anymore, just goes black for a few seconds and then I get the error and the game closes. Some other games, for example, Fallout 4, just close after around 15 seconds of playing, no error or anything.

This is my PC specs:

AMD Ryzen 7 2700X (with AMD Wraith Prism)
ASUS PRIME X370-PRO
ASUS ROG STRIX GTX 1080 8GB
2 x 8GB Crucial Ballistix Tactical 3000Mhz
SSD M.2 ADATA 1TB XPG Spectrix S40G RGB
SeaSonic M12ii 620w Evo
NZXT Phantom 410

I have a good PC so it's not because it isn't capable of running them properly. I have tried everything I could think of and every solution I found on the internet: disabled the auto overclock settings for the CPU on the BIOS, set my RAM speed to 2133Mhz instead of 3000Mhz, downgraded to my older BIOS version, made several new Windows installations after formatting, downgraded my Realtek audio driver to the version I was using before, tried both Game Ready and Studio Nvidia drivers, run sfc / scannow and DISM with no errors found, ran a memtest with no errors found, even swapped the RAM sticks from the slots in the motherboard...I also checked my M.2 SSD firmware but it's already the latest version. I am running Windows 10 Pro ver. 1909 with the latest updates (I have checked several times), this isn't a heat issue because when games crash temperature is just above idle. I also tried changing the TdrDelay on the registry but that didn't work either. I was looking through some older questions and found people with similar problems. Some of them fixed it by replacing the PSU. I changed mine for a brand new EVGA 650 GD and had the same problem, so it has to be something else :/

I don't have idea what may be causing this. Any help is super appreciated! Thanks in advance!

 

[JohnBonhamsGhost]

try running DDU(Display Driver Uninstaller) and remove all traces of your current graphics drivers and settings. restart and install only the latest graphics driver(Standard not DCH) directly from Nvidia, Advanced Driver Search | NVIDIA.
also if you have any type of malware/anti-virus software running, disable it completely.

what exactly are your average temps while idle and at full load for both the CPU and the GPU? are you following them as you try and run the games or just checking them immediately after they crash?

 

[Guille001]

I just finished doing what you said and the problem was not solved. I downloaded the latest Game Ready driver (also Standard not DCH) and used DDU to uninstall it.

I do not have any anti-virus except for Windows Defender, should I disable it?

Idle temps:
CPU - Around 40°C
GPU - Around 45°C

Full load:
CPU - Around 80°C
GPU - Around 70°C

I used MSI Afterburner to check the temps real time and they were both around 60°C when the crash occurred. Crash occurs just a few seconds after playing, so I don't think it can be related to high temps.

 

[JohnBonhamsGhost]

does it only crash when running games? have you tried running CPU or GPU stress tests individually?

not GPU driver issues so there's a possibility it is CPU related.
there's a chance a pin may have bent when you just cleaned & reinstalled. or a chance that it is just not fastened on correctly at the moment.
would remove and carefully check with strong light & magnifying glass to check for even slightly bent pins. if fixable or if nothing noticeable; carefully clean and apply small amount of thermal paste again and reinstall.

 

[Guille001]

I will try running Furmark and Prime95 and see what happens.

 

[Guille001]

I did almost one hour of testing with Prime95 and nothing happened, no errors, shutdowns or anything. Temp reached 94°C on CPU. I ran Furmark for around 20 minutes and also no errors or anything. Max GPU temps of 64°C.

I will take the CPU out and take a look carefully at the pins. I made it last time when I cleaned the PC and all the pins were fine so I don't think that's the problem, but just in case...

The problem happens with games and intensive graphical programs like Unreal Engine.

 

[Guille001]

So, I took out the CPU and everything was fine. None of the pins is bent and it was correctly fastened. I cleaned the old thermal paste with alcohol and built everything again. I also cleaned the ram contact pins with alcohol and made sure there was no dust in the slots. Unfortunately, the error persists. What's most interesting for me it's that it didn't crash with Furmark but it crashes with games, kinda strange :/

 

[JohnBonhamsGhost]

if it's an available option; try the GPU in a different system and see how it performs.

i've dealt with similar issues in the past that turned out to be internal system settings. the most recent was RAM voltage being set too low. this had resulted in the system having almost identical issue. every test passed and daily tasks had no problem but as games would start to fill up the RAM at a certain point it would just crash. usually fairly early into playing.
maybe the board's default setting(s) is off?
have you manually set the RAM voltage, timings, speed to the manufacturer's specs. or have an AMP profile to set?

if this doesn't pertain or doesn't help, try also manually setting all other components to their manufacturer stated specs.

 

[Guille001]

Unfortunately I don't have any other PC to test the GPU on. I don't understand: ''maybe the board's default setting(s) is off?'', which settings do you mean? I didn't change anything except for some BIOS settings (fast boot and something else...). Also, none of my hardware is overclocked so everything should be on stock specs.

 

[Guille001]

So, I kept thinking about what you said and checked the RAM voltage on the BIOS just out of curiosity. Turns out it was at 1.2V while my memory is supposed to run at 1.35V, I increased it to 1.35V but it did not solve the problem. Then I slowly increased it up to 1.45V but it did not help either, although it seems like crashes are now softer (FPS drop and then it crashes, the drop in FPS wasn't happening before). This seems like progress, although I don't know much about timings and voltages so I just stopped increasing the voltage in case it may cause damage or something if I continue. I also tried enabling the XMP profile and setting the speed to 3000Mhz but the error persists.

I always had problems with this specific RAM kit and this motherboard and I'm intrigued to know if it was because the voltage was always too low...If I did set the memory speed at 3000Mhz it would always cause instability during heavy load, so I just kept it at 2133Mhz and everything seemed to work fine.

As I said I don't know much about this stuff, so if you can help me about what to do next I will super appreciate it!

 

[JohnBonhamsGhost]

a motherboard's BIOS will have default settings for CPU, RAM, storage, etc. a lot of the time they are different than the manufacturer specs and have to be input manually to function correctly.

also check in the BIOS and see what the voltage settings are for the CPU. if they match what the packaging or online product page states, that should be fine.

i don't have access to your specific BIOS so i can't really help with what page or what particular wording they use for these settings. try checking online searches about BIOS settings for your particular board. there will usually be a detailed explanation or a YouTube video showing what exactly to do.

I just stopped increasing the voltage in case it may cause damage or something if I continue. I also tried enabling the XMP profile and setting the speed to 3000Mhz but the error persists.

yeah, you can cause damage increasing it too high. the stock manufacturer setting should be fine at the default speed.
strange there would be an "XMP" setting, that is an Intel option. AMD has always used "AMP(AMD Memory Profile)" that i've seen. maybe they just were confusing too many builders and decided to use Intel's wording now.

 

[Guille001]

CPU voltage stays between 1.2V and 1.4V, for a Ryzen 7 2700X this is normal as far as I know.

I will check the BIOS and make sure everything is correct, especially voltages.

I will let the memory at 1.35V then, and it wasn't XMP nor AMP actually, it's called D.O.C.P. on Asus boards, which is the same as XMP but with a different name so companies don't have to pay Intel royalties

I didn't see any voltage option for the GPU, is it possible that for some reason it automatically set itself to a lower value when I upgraded my PC? If it is possible to change it, I would like to increase it a little bit just to see what happens.

 

[scout_03]

video driver is related to gpu check if you have latest one in system after reinstall of os for the ram do memtest on each one at the time and dont go over 1.35 on voltages for them .

 

[Guille001]

I have the latest driver for my GPU, I did run memtest for my memory and everything was fine. Should I run it again now that I have enabled XMP (D.O.C.P.) and increased the voltage?

Also, all the other voltages on the BIOS seem normal to me:
https://drive.google.com/open?id=1Lfjq9UC-wcax4Hcnww6TYLy_gzjqdfh1

20200213_143756.jpg

drive.google.com

 

[JohnBonhamsGhost]

I will let the memory at 1.35V then, and it wasn't XMP nor AMP actually, it's called D.O.C.P. on Asus boards, which is the same as XMP but with a different name so companies don't have to pay Intel royalties

all of my ASUS Intel boards do use actual XMP profiles.
just to be safe, make sure this D.O.C.P. profile matches the same values(timing, etc) that Crucial states.

I didn't see any voltage option for the GPU, is it possible that for some reason it automatically set itself to a lower value when I upgraded my PC? If it is possible to change it, I would like to increase it a little bit just to see what happens.

this should be controlled through your GPU software; MSI Afterburner, ASUS GPU Tweak, etc. all of your GPU over/underclocking needs should be available there.

Also, all the other voltages on the BIOS seem normal to me:

you would have to compare them to the manufacturer stated specs for each.

 

[Guille001]

I fixed it! What I did was to start tweaking the settings in MSI Afterburner until it didn't crash anymore. I started modifying the GPU core voltage, I increased it to 100% but this alone didn't fix it. Then I reduced the clock to the minimum (-400Mhz) and it completely fixed the problem, although with a serious performance drop of course. So, I started to slowly increase the core clock until it crashed, and then I reduced it a bit from there. Right now I'm running it at -98Mhz and core voltage at 100% and it didn't crash after a few minutes, whereas normally it would have crashed only after a few seconds. I searched a bit and it seems like keeping the core voltage at 100% is completely safe with this GPUs (the actual increase in voltage is supposed to be very low).

This is not an ideal fix but at least my PC is working again

Any ideas about why this started happening now? My GPU was never overclocked nor underclocked. I would like to know if there's a way to make it work at stock speeds again.

 

[JohnBonhamsGhost]

if it's a factory overclocked version, it may just need a bit more than 100% to run stable. if stock Nvidia pre-set speeds, ??. though needing the boost now wouldn't make sense that it ran fine previously without.

i don't really see M.2 drives interfering with GPUs much. some M.2 slots will limit the PCIe slot's speed when used, but that shouldn't result in what you've been dealing with.

maybe you've knocked the GPU slightly askew when you installed the drive? could be a slight misallignment with the PCIe contacts. try removing and reinstalling making sure it is a tight fit into the PCIe slot and doesn't sag too much.

 

[Guille001]

Unfortunately, I think my GPU died just after some hours of use. It started with small artifacts which became bigger over time, it wasn't completely stable and I had to restart 3 times because everything would just freeze. Video signal abruptly ended shortly after that and the fans started working at 100% for no reason. Then I smelled something burning, so I immediately turned it off. I will contact Asus because it is still within warranty. I think the GPU was just faulty and for some reason it started showing errors when I reinstalled windows. Anyways, thanks for the help!

 

[JohnBonhamsGhost]

I think the GPU was just faulty and for some reason it started showing errors when I reinstalled windows. Anyways, thanks for the help!

that's a good possibility.
always sucks to have to wait for RMA services. good luck with the process.