Warning: Your graphics device has encountered a problem

[Ash0938503]

@BubbaBlitz7348

Read an article on one of the tech sites this morning that indicated that the issue was either the win 11 update or possibly the latest nvidia drivers. But this was only for win 11 and Ive gotten it in win 10 as well, so I guess that rules out the windows update that caused it.

The article talks mainly about 522.30 driver with win 11 22h2, focusing on performance issues not bugs, & it is not related to the su10 CTD GPU Bug.

as you mentioned, its happening in all windows versions.

[Ash0938503]

@TallestParsley2

@RolandoRossini

It is 100% due to SU 10 nothing else I have tested everything.

To prove that, I think, one would have to be able to take their MSFS back to the day before SU10 was released (probably via a system image or similar) and run it. So much else on our PCs, as well as on the MSFS servers, has also changed since then, that it still isn’t truly a scientific “double-blind” test, so to speak.

Not necessary if were possible, as many, almost 100% reported they didn’t experience the bug before su10, without changing anything.

[Ash0938503]

@Nevermind123451

If this is true everyone should try to limit their gpu with msi afterburner and start the sim to test if its gone

Unfortunately that was tried with no success.

[RolandoRossini]

Exactly, I have already tested it, had it extra screwed down and the same thing!

[DensestSnail693]

Hopefully you weren’t undervolting in Afterburner as IMO that seems to be the issue.

[Ash0938503]

its just very random, not strange if we cannot spot the exact cause easily. all we know its related to the gpu processes, & then the door of the possibilities is opened: is it the driver, or the sim dealing with the driver, or both, is it related to a programming error, or there is some sort of conflict needs to be resolved even if it is not a bug?!

thats what a developer do, LOL.

[hendis55]

@Ash0938503

almost 100% reported they didn’t experience the bug before su10

Unless I’m terribly mistaken, the “Your graphics device has encountered a problem” message itself was introduced in SU10, so that goes without saying.

And that message can probably be triggered in many different GPU-related cases with a myriad of underlying causes, some internal to MSFS and some not. Before SU10, similar cases would most likely have resulted in direct CTD or (more rarely) BSOD.

It does appear that GPU-related crashes happen more often now, though. But it’s far from clear that this is because of a specific coding error introduced in SU10.

[DensestSnail693]

As far as I can tell it’s entirely related to ‘gpu memory fragmentation’. Before Asobo/MS introduced the message then yes it would have caused a CTD normally preceded by a warning in event manager. BSOD’s I think not, that would usually be a Windows crash or cpu/ram/mainboard fault, most likely a driver issue or unstable overclock.

[hendis55]

@DensestSnail693

BSOD’s I think not, that would be a Windows crash or cpu/ram fault.

I don’t know. I used to get frequent BSOD’s related to the same nt…dll. A clean driver reinstall solved that. Whether such an issue would have provoked the new “Your graphics…” message today is doubtful, I agree.

[DensestSnail693]

Newer graphics cards and drivers are designed to shut down before a critical occurence, both Nvidia and AMD. However drivers themselves can be corrupted by another crash of some sorts.

[Ash0938503]

@hendis55

@Ash0938503

almost 100% reported they didn’t experience the bug before su10

Unless I’m terribly mistaken, the “Your graphics device has encountered a problem” message itself was introduced in SU10, so that goes without saying.

And that message can probably be triggered in many different GPU-related cases with a myriad of underlying causes, some internal to MSFS and some not. Before SU10, similar cases would most likely have resulted in direct CTD or (more rarely) BSOD.

It does appear that GPU-related crashes happen more often now, though. But it’s far from clear that this is because of a specific coding error introduced in SU10.

well said.

[Ash0938503]

@DensestSnail693

As far as I can tell it’s entirely related to ‘gpu memory fragmentation’.

as far as i know, MS Implemented new form of memory fragmentation in the sim, not exactly sure since su10 or su11 beta, but intersting if that has to do with the crash.

@DensestSnail693

Before Asobo/MS introduced the message then yes it would have caused a CTD normally preceded by a warning in event manager. BSOD’s I think not, that would usually be a Windows crash or cpu/ram/mainboard fault, most likely a driver issue or unstable overclock.

as that is a probability, i think that by being the error appearing in many driver versions, and some of whom reported the crash has no overclocks @ all, its point us that there is some sort of conflict @ least, i mean when we look to other softwares (high end demanding that needs a lot of vram and gpu proccessing), the error does not appear in them but only with MSFS, thats why most think it can be from the sim side.

[Ash0938503]

usually, fixing driver issue can be done fast when the error is among multiple softwares, as the cause in these situations is often obvious and not hidden, the problem is when the error happens to a single software.

@ least a special care from devs to the sim stability is a need.

[DensestSnail693]

@Ash0938503

we look to other softwares (high end demanding that needs a lot of vram and gpu proccessing)

Name just one that will max out a gpu and keep it there using multiple sources of data? IMO the framentation happens because the 522.25 driver is struggling to work as advertised and that is why some gpu drivers are better than others.

[Ash0938503]

multiple sources of data thats in business industry, softwares that is specialized in some fields. but in usual use terms, there is the stress testing that the benchmarking softwares offers, that can last for hours of processing to test performance and stability, & even conflicts when changing certain settings.

[DensestSnail693]

So then you’ll explain how you get problems and I don’t unless I induce them?

[Ash0938503]

nobody can really point to a cause, but there is a situation not for a few, that points to a stability issue from the sim side, when we use a fresh installed OS, Vanilla Sim, from 16 up to 24 Gb GDDR6X VRam, no overcloking, no settings being changed, & you have the error many times in random sequence, is’nt this showing that there is an actual problem?

[DensestSnail693]

Well considering reports indicate the .25 driver probably reporting wrong voltages to the gpus I think I’ll go with Nvidia’s theory rather than yours.

[Ash0938503]

the error has appeared in 4 driver versions up till now, LOL, so we have percentages to point probabilities.

[DensestSnail693]

Some more than others and yes all drivers require stabiity from their targeted hardware which in turn require stability from the PC’s they’re in. Simply put there’s a lack of redundancy built into those drivers.