Random crashes during gaming

Cazington · November 10, 2020

I'm experiencing seemingly completely random crashes when I play games. It doesn't seem to matter what game I play. They happen anywhere from a few minutes after starting a game to several hours of otherwise perfect performance. I've had these issues for a few months and have done intermittent troubleshooting. Because it's been several months I might forget some of the things I've tried. But first, the system:

OS: Windows 10 Home (64-bit)

GPU: Powercolor AMD Radeon RX 5700 XT Red Devil (RMA'd once)

CPU: Intel i7-8700k 3.70GHz

Mobo: MSI Z370 Gaming Pro Carbon AC

RAM: Corsair Vengance DDR4 3000MHz 2x8GB

Storage: M.2 Samsung 960 Evo 250GB (boot drive)

Samsung 850 Evo 500GB

Crucial MX500 2TB

Seagate Constellation ES.3 7200 RPM 2TB

PSU: Corsair RM850 (850W)

So like I said, sometimes I can play for hours and hours with not a single sign of problems. Other times it happens within minutes. There's no clear tell when it's about to happen, it just goes to a gray or black screen (seemingly depending on the monitor used). No stuttering audio or video, just crashes out of the blue. There's also no audio when the gray screen is up. The weird thing is that prolonged stress doesn't seem to cause any issues, I've stressed the rig with 3D Mark, AIDA64 and FurMark without any signs of instability. It has never crashed during browsing, watching videos, idling or what have you. Only when I play games and only seemingly random.

Here's a smorgasbord of what I've tried:

I've run MemTest64 with no errors showing up. I've tested the drives and they're all healthy. I've reinstalled windows. I've reseated the GPU and the RAM, I've checked the thermal paste on the CPU. The thermals in general are all normal, the GPU never goes beyond 80, usually hovering around 65-75C in games. The CPU rarely goes above 60-65C. I have two monitors, I've tested to run on only one monitor. I've made sure all cables are connected properly. I've even run two power cables to my GPU instead of using the splitter in case it had something to do with overloading the capacity of the cable. I've tried with using HDMI instead of DP (don't know why that would do anything but computers are stupid sometimes). I've tried undervolting the GPU. I've also tried overclocking. Flashed the BIOS. Oh yeah, drivers? I've done them all, I've reverted to older drivers that some users said had better stability and I've also used the most up to date drivers ofc. And some drivers in between for good measure. Used DDU in safe mode every time I did driver stuff. None of this solved the issue or had any noticeable impact.

I think I got everything I've tried. At this point I feel fairly certain it isn't software related.

My current working hypothesis is that the card trips the overcurrent protection of the PSU. I know this is an issue of the new Nvidia 3000 series. Though I haven't found anyone saying that happens with AMD's 5000 series. However due to it happening randomly and seemingly not because of stress I can't really think of anything else being the culprit. Sadly I do not have a better PSU than my current one to test this hypothesis on.

I included a pic of what the monitor looks like when it crashes, literally just a gray screen (black on the second monitor). I'm no expert but I couldn't find anything in event viewer past "windows shut down unexpectedly". Before I go spending money on a new PSU I'd like to hear from people who know more than I so here I am. If you got any thoughts they would be most appreciated as I am at my wits end. And if I have to bite the bullet and buy a new PSU what should I keep in mind? My current one is a pretty decent one.

*Edit: Forgot to mention, I RMA'd the 5700 XT once for the same issue and got a completely new one.

Edited November 10, 2020 by Cazington