Render device has been lost

Hello,

I have an RTX 2070 (not manually overclocked) and I consistently experience this Error. This happens only with overwatch, the temperature is about right (around 75 C - normal for a blower) and the max VRAM utilised is around 1GB. Everything in game is set in low except texture quality which is in medium and reduce buffering which is on.

I checked the overall temperatures of my parts right after the error and everything seems fine. Also, my card doesn’t seem to be faulty since it is able to perform more intensive tasks without problems like this (tuning deep learning models and benchmarking). Also, I haven’t experienced problems like this in other games, including screen freezes every min and loss approximately of 1 second of frames.

This is a bug in the latest Nvidia Driver, download the hotfix 445.78 to fix

1 Like

Cheers. I’ll update the drivers right now and I’ll update tomorrow about the results.

updating driver to 445.78 DID NOT WORK for me. did clean install.

Always do a clean install…

Clean install is often not enough. You may want to remove previous copies with Display Driver Uninstaller first.

1 Like

Indeed, I still experience this issue after updating. I followed DDU steps and reinstalled the latest drivers.

I will update the post with the results after a safe time period of playing.

I even tried rolling back my driver to the previous version, 442.74. I had apparently uninstalled the NVidia “GeForce Experience”, did a clean install and installed that as well.

STILL crashed.

Disable image sharpening.

hasn’t been enabled in the first place.

I used Display Uninstaller on safe mode etc, but still I get crashes. Also, here and there I got freezes. The freezes are more like when a computer is under heavy workload (GPU @ 65-70 C in training room), rather than missing 1 frame or something.

Is there anything I need to upload to make the picture of the problem more clear? (dxdiag etc)

The Blizz staff has some suggestions for this issue here:

By the way, if you’re seeing an average of 75C, that’s a bit warm, you might also be experiencing temperature spikes. You can use HWMonitor to check.

Actually it is my bad… It is around 75 C in the middle of a fight and the same more or less is when I check my card’s behaviour with heaven benchmark. In the practice room it is around 68.

Before I open this thread I checked the temperatures with HWM, because I have an ITX setup with 9900k, so I thought it was either the motherboard or the CPU. Since the computer completely freezes and needs sign out, I couldn’t see at that exact point the temperatures, but after a fast login CPU seems ok (and it should be because it is at 4.4GHz). The motherboard VRMs seem ok (by touching they feel to be a bit higher than room temperature) as well.

Nevertheless, I’m waiting for a custom waterloop for CPU, GPU and Mobo, so I can have definitive results upon the temperatures.

The cause of the “Rendering device lost” can be the GPU even if it works with other software. I experienced consistent and very frequent “Rendering device lost” errors (every few minutes) with one of the strongest RTX 2080 Super GPUs from one of the most popular brands. The card had issues only with OW. Replacing it to any of the other RTX 2080 cards I had access to eliminated the problem instantly. It wasn’t a driver error or similar, only the GPU was different. All cards were under about 60% load (Ultra settings) and everything was very far from overheating. My current RTX 2080 Super card (that works perfectly with OW) has some factory OC but not as aggressive as the problematic card.

If your PC crashes then the reason can be overheat even if your sensors don’t read any suspiciously high values. I experienced PC crashes about 2 years ago (it was unpredictable, sometimes every 10 minutes, sometimes once per day) and buying one of the biggest CPU coolers solved my problem. (The original cooler was the same brand - Noctua - but a smaller model.) I use only a hexa-core i5 8600k @3.6GHz without OC, not too beefy but still more than enough for the games I play. It can heat up within a split second when something starts using all 6 cores. It’s probably worse with an 8-core i9. Before the CPU cooler upgrade I tried to solve the crash issues with a lot of other things (reassembling the PC, different/stronger PSUs, etc…).

BTW, why are you using ITX? I’m a huge fan of small minimal design machines but in case of strong PCs (gaming, data science, etc…) going for ATX (perhaps EATX) is much more practical. You have access to a wider range of motherboards, cooling is significantly easier and you are less likely to have strict limits on graphics card size (you can also have 2 GPUs if needed). A modern ATX case (that doesn’t leave room for ancient DVD-drives, floppy drives, HDDs, and similar) can be very small. What I’m using is a good example for a modern small ATX case without much wasted space inside (Fractal Design Define C) and I upgraded to this from the micro-ATX version of the same model that was only 4 centimetres shorter in height! Saving 4cm wasn’t worth being limited to micro-ATX motherboards. There are much more ATX boards available with better VRMs.