Back to the main site

[solved, kind of] Overwatch General Error 0xE0010170

Hi everyone,

Recently I updated my Arch Linux and Overwatch started crashing on startup; I suspect the issue is the NVIDIA driver but I’m not sure. Downgrading the driver to version 450.57-3 fixes the issue. However, since Arch Linux doesn’t support partial upgrades, I can’t hang onto an old NVIDIA driver over an extended period of time. Here are the relevant pastebins and a screenshot:

Any idea on what the issue could be or how to fix it?

Also sorry in advance, the forum limits me to posting two links in a post. I’ll try to see if I can put everything into one link or post the rest below.

Remaining bullet points from above:

Found another person who shares the same issue.

https://forum.winehq.org/viewtopic.php?f=8&t=34282

Hi all,

I was able to successfully run Overwatch just using wine-staging and some other tweaks (but not with Lutris). The inspiration came from docs in the dxvk github repo.

The performance is amazing (stable 60 fps on ultra settings) with the exception of occasional stuttering because of shader caching.

Unfortunately, I don’t know what issue there is running Overwatch under Lutris. Doing a system downgrade to an older linux kernel and nvidia driver (and basically all other packages) did not resolve the general error. My guess is that it’s related to dxvk caching and nvidia driver updates, but I don’t have the knowledge to fix this on my own.

I too have this issue and have had it for a very long time now (at least a couple of months), every fix you use also fixes it for me, but booting via Lutris is heck times for convenient.

I’m also using arch just like you, it would be great if this gets fixed.

If I disable DVXK in runner options the game boots, but stutters quite a bit and incredibly low FPS whenever shaders need to get calculated

Same error here, if I disable the DVXK the game starts. What could it be?

**Note, this General Error message seems to change every year or so, The error message 0xE0010150, 0xE0010160, 0xE0010170, and now 0xE0010180 have started getting reports all seemingly related to the same core issue (graphics). The last error message I got was 0xE0010180.

Possible fix or reason:

My system had a 1080ti and a 2080ti without SLI in the system. The 1080ti was the one in the primary motherboard slot and the one I was getting display output from. This means the 2080ti was essentially doing nothing but suck power.

Once I removed the 1080ti, and put only the 2080ti into the primary motherboard slot, everything immediately started working.

I’m using KDE Plasma and occasionally but not all the time random OS elements or applications would start to flicker in-and-out of existence. This happened often and a Redditor even posted what seems to be describing the same issue in which no one knew what was the cause or even experienced it, but AMD users seemed to experience a similar but different problem that other AMD users also stated they encountered. https://www.reddit.com/r/archlinux/comments/ir35zq/has_anyone_else_with_nvidia_graphics_noticed_odd/

This means its possibly an issue specific to having 2 GPU’s non-SLI in the system, or the 1080ti is acting up. Based on responses on that Reddit post it makes me believe this could be errors with VRAM allocation, meaning the VRAM may be faulty.

I ran Machine-learning on the 1080ti very heavily and it’s a couple years old now, the machine-learning training would result in almost a maxed out VRAM usage at all times, but not particularly a maxed out utilization. Perhaps all the read-writes to VRAM killed something and is now causing issues?

Regardless once I removed it and only kept my 2080ti, everything started working again.

EDIT 18 Oct 2020

I can now confirm it’s cause to be simply having 2 GPU’s in your system. I’m getting display output from my RTX 2080ti, With the 1080ti just sitting there. If I remove the 1080ti without changing anything it works again.

I need to use both my 1080ti and 2080ti, so it would be great for a fix to this.

As for the VRAM issue’s I was speaking of, it turned out to actually be faulty regular CPU RAM. I have 4x16GB Corsair Vengeance Pro RGB, one stick was faulty and caused memory corruption. Since removing that stick of RAM I haven’t had any OS corruptions or glitches.

Solution found! :smiley: You can force DXVK to use one of your two (or more) GPU’s when playing Overwatch.

For example, I have 2 GPU’s, a 1080ti, and a 2080ti. If I remove either, it works, but not if I have both.

To fix this by forcing DXVK to use a specific GPU, it’s as simple as using the DXVK_FILTER_DEVICE_NAME environment variable.

Instructions on it’s usage can be found here: https://github.com/doitsujin/dxvk#device-filter

i.e. in my instance, putting DXVK_FILTER_DEVICE_NAME=GeForce RTX 2080 Ti was the magic needed.

You can put this environment variable into Overwatch via Lutris by configuring Overwatch entry -> System options tab -> Environment Variables.

Enjoy!