Question Crash to Desktop problems when playing AAA titles with MSI RTX 3090 SUPRIM X

Feb 3, 2022
2
1
15
0
Hi everyone!

I bought an RTX 3090 SUPRIM X last year in april. For the first 4 month I had no problems with my system at all, but I only took a sneak peak at some triple A titles, as I didn't have enough time. Mostly I played some multiplayer games like Valheim, 7 days to die or Minecraft and didn't experience any problems. End of September I finally had some days of vacation and wanted to play some MS Flight Simulator. After an awesome experience for 1 1/2 hours, I had my first CTD/Crash to Desktop. I didn't suspect anything, started the game again and was able to play for 2 hours, then again: CTD. No error message (neither windows nor driver), no BSOD, nothing in the eventlog. Frustrated to the max I wanted to try something else and started Cyberpunk 2077. Truly beautiful, but 45 minutes into the game an BAM, Crash to Desktop again. So now I was suspecting a heat issue and installed MSI Afterburner with Rivatuner statistics server to monitor everything while playing. But everything seemed to be within tolerance (will list the temps below). So I asked my good friend google and found plenty of threads and questions about similar problems (not with the same GPU though). So I planned to check on that problem again in my next vacation (last christmas), but I ran into the exact same issues again. Every new(ish) AAA title I own would crash after a random time (20 minutes up to3 hours) without any error message (except those "oops, I crashed" messaged from the games themselves). It will happen at Ultra settings or at low settings, with or without ray tracing and/or DLSS. So every weekend in January, I tested like I have never tested a system before :D This is my regular setup:

Mainboard: ASUS ROG Strix X570-E Gaming
CPU: Ryzen 9 5900X
RAM: G.Skill Trident Z Neo 3600MHz C16
GPU: MSI RTX 3090 SUPRIM X 24G
PSU: BeQuiet! Dark Power 12 1000W

Except for D.O.C.P., nothing is overclocked.

Temperatures:
Idle
  • CPU: 35°C
  • GPU (Core): 37°C
  • GPU (Hot Spot): 46°C
  • GPU (Junction): 44°C
While playing (temps stay like this after about 30 minutes)
  • CPU: 55°C
  • GPU (Core): 75°C
  • GPU (Hot Spot): 85°C
  • GPU (Junction): 88-90°C
After everything I found out until this point, those temps seemed to be quite good and nothing to worry about. As mentioned before, it sometimes takes 3 hours until it crashes. No matter which GPU hungry title - it's like pressing Alt+F4.

My testing included:

RAM tested with memtest86 (in windows with 12 instances for 10 hours, no errors; with bootstick several times 4/4 with zero errors) and exchanged the modules (3200MHz, 3600MHz, 4000MHz) and tested with auto setting instead of D.O.C.P.
CPU tested with Prime95 and the CPU-Z stress-test (8 1/2 hours overnight, no errors) and exchanged it (R7 5700G, R7 5800X, R9 5900X)
Mainboard tested with different hardware and checked all temps with HWInfo; exchanged the mainboard (All ASUS: B550-I, X570-E, X570-F)
PSU exchanged (BeQuiet Dark Power 12 1000W, ASUS Thor 1200W)
GPU tested with FurMark, 3DMark and F@H - it didn't crash even after 9 hours of FurMark; exchanged (RX5700 XT, RTX 3060 TI, RTX 3090) - of course the other two cards don't need as much juice as the 3090 (in case it's a power issue)

The only component always involved in the Crashes to Desktop was the RTX 3090.

I also tried a lot of stuff I read in forums regarding similar cases:

Uninstall any monitoring tool
Setting up a fresh windows installation
Underclocking (GPU)
Set PCIe 4.0 to 3.0
Checked all firmware/BIOS versions (usually keep everything up to date)
Install (and uninstall) NVIDIA GeForce Experience
Uninstall ASUS Armoury Crate
Uninstall Corsair iCue (should have done that a long time ago)
Logging with GPU-Z
Monitoring all temps with HWInfo64
Disabled hardware accelerated GPU scheduling in Windows
Set the Windows power plan to full performance (was balanced before as I heard Ryzen likes it)

Nothing helped, not even a bit.

Last weekend I stumbled across an Igor's Lab article regarding a CTD problem of 30 series GPUs. The error seemed to be exactly like mine. GPU intense games --> CTD. So I contacted MSI Support and explained the situation. They asked for BIOS and driver version. Both was up to date and the support then asked to use DDU and make a clean install of the drivers. It feels like I did that about 5000 times, in my regular windows installation and all the new ones for testing. So MSI Support couldn't help me and asked me to approach my vendor for RMA. I already applied for that, but I wanted to use the weekend for more testing. I am a bit sceptical, of how they will test my GPU. Stress tests won't show the error - will someone sit there and play a game for a couple of hours to see if the crash occurs? I don't know.

So, maybe one of you guys has a great idea or has had a similar/THE SAME problem? I am really clueless and starting to get frustrated. I have never owned a high end computer and I had to save for a long time to make this happen. Is this one of those life lessons everyone talks about?
I didn't post any screenshots yet, so please let me know if you need to see something specific.

Thank you so much in advance for your input!
Have a good one
 
Last edited:

Nine Layer Nige

Distinguished
Is it possible that you could be receiving a broken packet through your internet that might upset the game ?
Are you using Wifi or ethernet ?
I'm all outa ideas as it looks like you have been very comprehensive in your testing.
 
Feb 3, 2022
2
1
15
0
I got my hands on another B550 Board yesterday. The problem is no more! Though I really don't understand what's wrong with the other 3 Boards, as they all run fine with an RTX 3060 TI or RX 5700 XT. Maybe the 3090 is just too hungry...
 
Reactions: Nine Layer Nige

ASK THE COMMUNITY

TRENDING THREADS