Question GPU crashing under load, monitors go black, have to force restart every time

[pajasso]

Hi!

This is a problem that appeared about one week ago and after a lot of different problem solving I'm now left clueless. In a nutshell, while gaming demanding games like Rust my monitors suddenly go black with "no signal", and I have to force a restart to be able to do anything on my computer. Most of the time when these crashes happen the sound from the game still keeps playing and I can still hear my friends and they can hear me even while the screen is black, but sometimes even the sound cuts out.

After reading multiple posts about similar problems I came to the conclusion that it must probably be my PSU, because the PSU was the only component I didn't renew when I built this PC. Now after receiving a new PSU, the problem still exists and now I really need some help.

Spoiler: Specs:

Asus PRIME Z390-P (long story, my go to "premium" mobo was DOA and I needed a new one quickly)
i9 9900K
Gigabyte RTX 2080 Super Gaming OC
Kingston HyperX FURY 16GB DDR4
Adata XPG SX8200 Pro 1tb
Corsair RM650/Corsair RM850x

All clocks are factory. I'm running a custom loop, idle temperatures are around 30C for both the GPU and the CPU. I like to keep my PC quiet since it's on my table, so the fans aren't really spinning that fast under load either, and the temps are around 50ish on both the GPU and CPU, while they reach 57-58C under extreme load. The only thing that really concerns me regarding the temperatures is the temperature of the GPU backplate, as it gets burning hot under heavy load, which should be normal, and since I don't own any thermometers I can't really tell how hot the backplate is.

I have tried different power outlets, different cables for the motherboard and for the GPU, I have reinstalled drivers, I have tested my RAM and I just can't seem to find the problem. I can play less demanding games (i.e. Hearthstone) for hours without any issues, but after about 10 minutes in Rust everything goes south. I've run a couple of benchmarks but they have yet to crash the GPU. As I already stated, this only started happening recently, and the PC has been able to perform under heavy load for almost 2 years now. One day I had Rust running while I was afk, and when I got back to the PC the screens were black with no signal.

I've got an Aqua Computer cuplex kryos NEXT with VISION as my CPU block, which makes monitoring temperatures easy. I've noticed that whenever the screens go black, the GPU temps on the VISION drop to 0C, which further leads me to believe something is wrong with the GPU? The GPU is still under warranty, but I've already been through one RMA process with Gigabyte (sadly after I got my GPU), and that's something I'm really not looking forward to. At this point I'm starting to run out of ideas, so if anyone could help me that would be great. If I've forgotten to include something please let me know, I'll try to get back to everyone as quick as possible.

 

[Lutfij]

Welcome to the forums, newcomer!

What BIOS version are you currently on for your motherboard? What OS version are you on(assuming you're on Windows 10)? Which PSU are you on? You've listed a 650W unit as well as an 850W unit. Which one is the new unit?

Are you on a custom watercooling loop for both the GPU and CPU?

You might want to see if uninstalling your GPU drivers using DDU, then reinstalling the latest drivers from Nvidia's support site, in an elevated command, i.e, Right click installer>Run as Administrator, changes your experience.

 

[pajasso]

Thank you!

I'm currently on BIOS version 2417 (06/03/2019) and I'm running Windows 10 Home. The old PSU was the RM 650, and I'm currently using the new RM 850x.

The custom loop is indeed for both the GPU and CPU. They are in the same loop, consisting of one 240mm rad and one 360mm rad.

I'll try uninstalling the drivers! I just did some heavy benchmarking and couldn't get the screens to black out, but as soon as I started Rust and was in game for about ~1 minute with the system already being warm from the benchmarking I ran into the issue.

Edit: Tried reinstalling all drivers with DDU, still got a black screen after a couple minutes

 

[pajasso]

I think I managed to catch a crash in a HWiNFO64 log while I was logging for about 60 seconds where the crash happened at about 46-50 seconds in. However I have no idea how to interpret this log, even when opening it in a log viewer I don't have a clue what to look for. It is obvious that some things like PCIe Link Speed drops from 8 to 0, but if anyone who has got more knowledge than me could interpret the log it would be great! Am I allowed to post the log .CSV file somewhere?