#Routine and Inevitable game crash using dual NVIDIA 2070 Supers

10 messages · Page 1 of 1 (latest)

ivory mason
#

When in a game world, the game will run fine and then hit a brick wall and crash to desktop. Error codes are consistently Line 198 or Line 684. Crashes have happened during Olympus starter mission, livewire, predator (the only ones we have done so far). Crashes have happened at all times of day and in all weathers regardless of player choice of activity. Being motionless in base, walking, crafting, or fighting does not seem to dictate a crash. Sometimes the crash happens after a minute or two. The longest I have played was one 6-8 session with 3 missions before it crashed. I am typically playing steam friends only but my friend is not always connected when it happens.

To troubleshoot, I have tried:

  • Playing in DX11, DX12, medium settings on DX11 and DX12, epic settings on DX11 and DX12. I have reinstalled GPU driver. I have reinstalled GPU driver fresh. I have verified steam files. I have enured updated C++ redistributables for x64 and x68. I have tried regedit'ing GPU response time to be at least 60 sec. I have even disabled one of the GPUs. I have also used MSI afterburner to ensure GPUs are synchronized. Windows is also fully up to date. They also don't seem to be stock overclocked. This accounts for every troubleshooting method I have read, aside from reinstalling the game (though again, steam verified integrity of files). Regardless of any updates or different graphics settings, the Line 198 or Line 684 crashes seem inevitable.

I have no other games that are crashing with GPU-related errors. Icarus is unmodded and played through Steam.

Intel(R) Core(TM) i7-7800X CPU @ 3.50GHz 3.50 GHz
44.0 GB RAM
Windows 10 Pro, 64 bit
Dual NIVIDIA 2070 supers

pulsar river
#

D3Device lost is an indication that your GPU became inaccessible for some reason. common causes are power, thermal, or hardware problems.

ivory mason
#

It's got a 1K watt power supply so it should be good. I am going to try taking everything apart for a deep cleaning because I do see one is running hotter than the other. It would be slightly embarrassing if dust was the issue but that would be the cheapest problem to fix. oO

pulsar river
#

assuming power is good just because it's got a 1kW PSU isn't valid. I mean no offense by that, but 1) power supply performance does degrade over time, sometimes significantly, 2) a lot of people end up attaching a lot of power-hungry peripherals to a system, enough that it can affect performance and 3) you said explicitly you're running two GPUs. It could be something where kicking off both GPUs and the fans needed to cool them, while other stuff is running, is just enough to dip you below stable operation.

#

I'm not saying that's what it is definitely, but probably 60-70% of the time that someone comes in here with device lost errors, it ends up being a power issue and resolved by a new PSU.

ivory mason
#

I was mistaken. I gave it a Thermaltake 1200 watt PSU. I cleaned everything. We will see how it goes. I noticed 1 back fan is out but I will have to watch it powered on with the panel off before I know it is just that fan.

ivory mason
#

So... the lower graphics card wasn't bothering to use its fans. MSI Afterburner is now forcing it to use its fans ~ which it can without a problem. But I guess MSI Afterburner has to stay open for this fan v temp curve to stay in affect?

ivory mason
#

@pulsar river I'm going to say thermal from your first post was the main issue. We can consider this closed. Thank you.

ivory mason
#

Update. Cleaning and using MSI Afterburner to ensure fans keep heat under control seem to help a lot. However, after a few hours of play on a redo on the first livewire, I got another Line 684. MSI Afterburner was up the entire time on screen 1. The second card was a little hotter but not abnormally so and the fans were still respecting the heat vs fan speed curve. I saw no power weirdness either.

robust zephyr