GTX 1080 artifacts and crashes

[m3ltman]

Hello,

My build:

-MSI Z370 Gaming Plus Mobo
-Intel core i7-8700k CPU
-MSI GeForce GTX1080 Gaming X GPU
-Samsung SSD 860 EVO 500GB
-WD Caviar Green 2Tb HDD
-Corsair Vengeance LPX 16GB (2x8GB) DDR4 DRAM 3000MHz C15
-Seasonic Focus Plus 750 Gold SSR-750FX 750W 80+Gold PSU

It’s a completely new build and first problem for me was installing NVIDIA drivers. Windows installed 398.36 version with it’s update, but when I tried to manually download latest version and do it via Nvidia installer (Uninstall device in device manager beforehand) it’s said that “This NVIDIA Graphics Driver is not compatible with this version of windows”. I did install every windows possible update beforehand of course, and still the same error. I fixed it by using DDU in safe mode. After that drivers installed without any issues.

I don’t think that for the first day or two I noticed or had any issues. I played a couple of matches in PUBG and may be spent in game hour or a little longer, everything was smooth and nice. Then, on weekend my taskbar started to flicker with different colors while I was adjusting different windows features, it lasted for a couple of minutes and disappeared. I installed MSI Afterburner and used riva tuner on-screen display in games (if it matters). Then, PUBG crashes starts to occur (exit to desktop with error message), I started to see artifacts and distorted textures before crashes. Via MSI Afterburner I tried to adjust Core Clock (OC and downclocking), I was very careful and don’t think that I could damage something (on core clock maximum that I achieved was gradually adjusting it to +120) and tried to increase voltage gradually max to +10, don’t think that I noticed any changes in performance or reduce the amount of crash, I just dropped settings to default. Tried to run Heaven benchmark, and I think that in the beginning it could even finish it, but now it’s just crashes (exit to desktop) in random places. Fans on GPU working properly. Also, I noticed that under load GPU starts to buzz (may be it was from the beginning and I’m just paranoid or wasn’t paying attention or played in headphones), I think that this is popular coil whine judging by sound though. I don’t think that I saw GPU temperature goes above 67 degrees also, in Heaven it crashes right on this temperature. CPU temperature doesn’t seem to go above 60 degrees either. Tried to play CS:GO for 20 minutes, it didn’t crash but artifacts and squares appeared and disappeared on screen periodically. I think I heard some strange noises from PSU during Heaven benchmark, but I’m not sure if it wasn’t my paranoia again.
Yesterday, I completely formatted my SSD and re-installed windows, the same error with drivers from the beginning. And heaven crashes in different places, but mostly in the begging of test, on 64-67 degrees, or even lesser. I updated BIOS, installed fresh chipset drivers. Checked PSU connectors, readjusted GPU on its slot (even tried second PCI-E on my mobo). This new windows downgrade driver version to 398.36 again during benchmark I think. And now artifacts periodically appears on Windows desktop screen. Nothing seems to work.

Sorry if this post is messy and inconsistent. I’m a non-native speaker

Any help will be much appreciated.

 

[Kashimi]

That sounds to me like a defective unit. You've done enough to warrant just getting the company who sent it to send you a new unit, or contacting MSI and doing a warranty exchange.

 

[SkyNetRising]

Any help will be much appreciated.

It's possible, you have damaged your new GPU with OC experiments.

Try downclocking your graphics card to Nvidia reference clocks (1607mhz base, 1733mhz boost). This corresponds to silent mode clocks of your graphics card. OC and Gaming clock settings probably are too aggressive for your particular unit.

 

[m3ltman]

That sounds to me like a defective unit. You've done enough to warrant just getting the company who sent it to send you a new unit, or contacting MSI and doing a warranty exchange.

Thanks for the reply! That's what I fear most, because I live in Russia and all components was bought from vendor in Germany, and this warranty procedure will cost me some money for sending it to them and of course it wouldn't be fast and easy. And I'm assuming they just can say, hey it works pretty fine for us and it's not the warranty case. So I need to be quite sure that it really defective. Do you think the only option left is to install it in another rig? Which I don't have btw 🙁

 

[m3ltman]

Any help will be much appreciated.

It's possible, you have damaged your new GPU with OC experiments.

Try downclocking your graphics card to Nvidia reference clocks (1607mhz base, 1733mhz boost). This corresponds to silent mode clocks of your graphics card. OC and Gaming clock settings probably are too aggressive for your particular unit.

But this experiments was so small by OC standards and from what I've read and seen on youtube, do you think it could damaged it? And artifacts and stuff became to occur even before that, and I've tried to downclock clock speed to even lower values, but it seems not to change anything.

 

[Kashimi]

That sounds to me like a defective unit. You've done enough to warrant just getting the company who sent it to send you a new unit, or contacting MSI and doing a warranty exchange.

Thanks for the reply! That's what I fear most, because I live in Russia and all components was bought from vendor in Germany, and this warranty procedure will cost me some money for sending it to them and of course it wouldn't be fast and easy. And I'm assuming they just can say, hey it works pretty fine for us and it's not the warranty case. So I need to be quite sure that it really defective. Do you think the only option left is to install it in another rig? Which I don't have btw 🙁

If you have a friend that has a PC then go about it that way, if not try contacting MSI directly and going through the warranty.