Dxgi_Error_Device_Removed
#1
Posted 16 August 2016 - 08:33 AM
Everything is running just fine. I've had the gun running virtually non-stop for three days, and I've been playing on this particular rig for four months.
wtf?
#2
Posted 16 August 2016 - 11:02 AM
#3
Posted 16 August 2016 - 12:29 PM
It's a pretty bizarre issue, as you won't get that error in a different game not based on Cry Engine, try downclocking your GPU and see if it occurs again.
#4
Posted 16 August 2016 - 12:43 PM
800 watt PSU and GeForce 560ti is the card
It's definitely game related because I have more demanding games where this does not occur.
#5
Posted 16 August 2016 - 12:50 PM
#6
Posted 16 August 2016 - 11:37 PM
But ok. Ask for help and dismiss it up to you.
#7
Posted 16 August 2016 - 11:43 PM
Swap the game from DX11 to DX 9 and see if the issue still occurs, if it doesn't it's because DX9 loads the CPU.
Your other options are to use DDU to uninstall reinstall drivers, and do a full install of DX.
DatGuyHigh, on 16 August 2016 - 12:43 PM, said:
800 watt PSU and GeForce 560ti is the card
You have a power hungry fermi card, and if it was all about the PSU claimed wattage we'd all be running diablotek 1KW PSU's
#8
Posted 17 August 2016 - 05:19 AM
Conversely, you could lower the clockspeed by like 35MhZ on the core and 15MhZ on the memory and see if it helps instead of bumping up the voltage a notch. Either one of these should eliminate any ludicrously slight instability that may exist. It may be worthwhile testing this with more extreme values if it doesn't seem to help, something like 100MhZ less on the core and 50MhZ less on the memory.
I don't think Cryengine causes this for -no reason- but I also think that it has a certain way of handling things and freaks out faster than other engines do when interacting with OS user-layer and kernel-layer stuff for hardware items.
The other thing to look at is a complete clean-out of drivers before installing the latest available for your card. It's possible that's the issue. It's probably less likely, but it's definitely possible.
Hell, it could even be a bad bit in system RAM or your permanent storage medium of choice (HDD or SSD).
Claiming it isn't hardware because your current hardware is faster on paper is a very 2-dimensional argument that won't fly here, btw. We're mostly people that have been in technical positions all the way from tier 1 technical support to engineering jobs in this section of the forums. If you ask for help, we help. If you reject help on the notion that you're feeling 14 years old today, we just walk away and let you deal with it.
#9
Posted 17 August 2016 - 05:31 AM
Oderint dum Metuant...I've only seen the issue ONCE. How the hell is it dismissing your advice? It's not like I can tweak some settings, try it, and if it runs then "you were right". The game's running right now with no changes. Until the error becomes chronic changing stuff won't tell me or you anything.
#10
Posted 17 August 2016 - 07:03 AM
Makenzie71, on 17 August 2016 - 05:31 AM, said:
Oderint dum Metuant...I've only seen the issue ONCE. How the hell is it dismissing your advice? It's not like I can tweak some settings, try it, and if it runs then "you were right". The game's running right now with no changes. Until the error becomes chronic changing stuff won't tell me or you anything.
You have a problem.
You come to a hardware forum to ask for help.
People offer help based on their knowledge.
You dismiss that help with "it's not my hardware"
If only life were that simple, hardware develop faults sometimes these faults come early some times late into their lifetime.
As pointed out, this is an excessively common issue with Cryengine games, MWO is not unique to this error, it has occurred since the early Crisis days.
I love how people not just you keep coming back to the fact they have an over the top PSU, it doesn't matter that you have a 1.2KW PSU.
It can still fail, it can still develop a problem, there is so much more to a PSU than it's advertised wattage otherwise as i say if that's all it came down too, we'd all have 2KW diablotek PSU's..but we don't because some are made well, and others are trash.
Edited by Oderint dum Metuant, 17 August 2016 - 07:05 AM.
#11
Posted 17 August 2016 - 07:21 AM
Considering the game is working just fine now, and that all of my game settings have been reset, it's likely that it was an issue with the program itself and not hardware.
#12
Posted 17 August 2016 - 09:51 AM
The game stores all of your configurable settings in a file, or a few files. If the game crashes and then when you bring it back up everything is reset to default, that could indicate that there was a write issue in your storage or RAM. RAM issues and HDD/SSD issues, when very minimal or even accidental in nature, are the types of things that happen one or twice and then you don't see them anymore or at least you don't see them for a decent amount of time.
You are correct that it could also indicate an issue with the application itself, but I'd think that if it were the application it would be a lot more prevalent than it is. We don't see it very often, and the issue has been observed most often in Cryengine based games when there are odd issues with hardware, be them major or very minor, so the application is definitely the less likely culprit.
We're not white knighting for the devs, we know this game still has a lot of issues. This just probably isn't one of them.
#13
Posted 17 August 2016 - 09:56 AM
#14
Posted 17 August 2016 - 01:05 PM
(*Note I can't remember the specific error I was getting, but that seems familiar.)
#15
Posted 21 August 2016 - 11:12 PM
#16
Posted 24 August 2016 - 06:37 PM
And I have had them crash on Radeon, so it's not directly linked to nVidia only.
With my previous drivers I often had those driver crashes, I also tried setting power limit from GPU settins higher thinking maybe my card is getting old. But never I had game settings reset.
Such settings could reset if they are being written at the time of crash, but I would think settings are written to only during modifying them.
If you run the repair tool it will reset settings.
Also, it might be different from Steam install than standalone. I have standalone install. Maybe Steam install messes up or resets the settings at crash. If you have Steam try disabling online game data saving if it's an option, as the game account stuff is always on server anyway.
#17
Posted 24 August 2016 - 09:05 PM
Casual warm up of the graphics card always reduced the number of crashes. Notice the initial load of the GPU in Mechlab or Lobby screen is higher than in game, about 70-99% vs 45-55%. On cold boot, my rig was typically destined to crash.
Not taking the most concurrent Nvidia driver. I'm running 347.25 on a Titan Black. Reduced crashes, no more hard boots (black screen) of the system. Nvidia stopped tuning the drivers for that particular card around the 355 version, which means no performance benefits from later versions anyways.
Using e.g. PrecisionX from EVGA to a) change the fan profile and downclock the GPU to consistently stay below 76°C under severe load and preventing the card from 'boosting'.
Alt-tabbing out of the lobby screen to reduce GPU usage from 99 to ~ 50%.
Investigating the voltage stability of your power supply. Look at the 5V and 12 V rails in the bios screen. Does it fluctuate? Do you hear strange noises from your power supply, e.g. clicking noises? Mine started doing that after around 3 years of service, so I swapped it out immediately. (1200W, Silverstone)
The jury is still out but no crashes after that so far.
Your mileage may vary. RAM or VRAM can be faulty as well.
1 user(s) are reading this topic
0 members, 1 guests, 0 anonymous users