Help! Games crash at high graphic settings

[ThT_AZN_KiD]

Monitor: LG 27gr83qb (Ultragear 27" QHD 1ms 240Hz with NVIDIA G-Sync)
CPU: Ryzen 7 5800x3d
GPU: MSI GeForce RTX 3080 Suprim X 10g
RAM: Corsair Vengeance RGB Pro SL DDR4 32gb (2x16) 3600mhz CL18 XMP 2.0
MB: MSI MAG B550 Tomahawk MAX Wifi
PS: Corsair RM850x SHIFT
CPU Cooling: Thermalright Phantom Spirit 120 SE ARGB
Case: Lian Li Lancool 216
SSD: Western Digital Black 2TB SN770 NVMe - Gen4 PCIe (x2)
OS: Windows 11


I'm fairly new to PC building and gaming but tried my best to do a mid build that I wouldn't have to make any significant changes for several years. I understand that I am fair game for being dragged here.. but after talking to PC building/gaming friends I can't figure out what is happening.

So far I've played games like RB6 Siege, CS2, Battlefield V, Chivalry 2, and Hogwarts Legacy and during gameplay, if my graphics settings are set to High or Ultra and/or if I increase FPS (usually around 120+) the games crash (screen will freeze or go black, audio playback continues, and then I'm exited from the program or have to open task manager to close it). If you can google it I've tried it (minus suggestions such as vulkan launch options, thanks to some of my PC friends). I've basically rebuilt the PC to ensure all hardware is connected properly, I've reinstalled Windows/games, updated drivers, tried older version drivers, tried testing with in-game tests and using cinebench, playing with graphics settings such as shadows, dx12, etc. etc.

I've now settled for playing games at low to mid graphics settings at or under 100 FPS. I'm not playing for competition so I understand i dont need tons of FPS, but now I just want to figure out why the PC is not performing like it should (to my limited knowledge).

I welcome any advice from the community!

[ˢᵈˣ FatCat]

Originalmente postado por WinterSolstice:

Originalmente postado por FatCat:

this is the firstime i see somebody suggesting updates firmware for M.2, can M.2 really did this, i really curious, because there is one paticular game make my PC black screen crash

I am only speaking about WD Black M.2 because I use another model and found that the firmware was an issue at one point in time personally - everything I suggest I have used successfully and experienced in the past myself, fyi. :csd2smile:

and after that, you free from the crash?
mine is 990pro samsung, and the 2nd is WD black

[Phénomènes Mystiques]

Originalmente postado por FatCat:

Originalmente postado por WinterSolstice:

I am only speaking about WD Black M.2 because I use another model and found that the firmware was an issue at one point in time personally - everything I suggest I have used successfully and experienced in the past myself, fyi.

and after that, you free from the crash?
mine is 990pro samsung, and the 2nd is WD black

That one time, yes, but I only use one M.2 and I have found other stupid things related to game crashing such as the Apple Bonjour application or GPU clocks overly high in it's native application (and there are more problems with the OS itself some on purpose to make it run properly and Windows 11 is a broken mess for some systems - frankly, I am done with updating the GPU driver).

The M.2 had only one firmware update since I bought it in 2021 so that was at initialization and I also saw an interesting difference running the WD Black in "Game Mode" as opposed to not using the mode.

[ˢᵈˣ FatCat]

Originalmente postado por WinterSolstice:

Originalmente postado por FatCat:

and after that, you free from the crash?
mine is 990pro samsung, and the 2nd is WD black

That one time, yes, but I only use one M.2 and I have found other stupid things related to game crashing such as the Apple Bonjour application or GPU clocks overly high in it's native application (and there are more problems with the OS itself some on purpose to make it run properly and Windows 11 is a broken mess for some systems - frankly, I am done with updating the GPU driver). :MaterialGirl_expression2:

yeah i hate updating GPU driver especially for AMD, it's always train-wreck after, once i cant use Free-sync, but the problem if i keep hold updating driver, my windows forced me to use the newest driver, and i need eventually reinstall driver so i can use AMD adrenaline

thx for the new knowledge

[tigerxs1]

do u use gpu fan controller,like aorus engine or gigabyte control center,my gpu used to play up intill i setup fan curve and prefer to run fans all the time,not semi/off and on,i dont know what the msi fan controller would be?

[ThT_AZN_KiD]

Problem not "solved" but definitely getting somewhere thanks your responses! Updated the Graphics Driver and have RAM voltage at 1.35. Now getting 144-200 FPS on Ultra and High Graphics settings (varies depending on the game) but crashing on Ultra settings if the FPS limits are higher than 200/220 (again, varies..). Still avoiding the typical in-game overlays and performance inhibiting settings. That said, I'm not unhappy about the progress..this already looks and feels better.

[ˢᵈˣ FatCat]

Originalmente postado por ThT_AZN_KiD:

Problem not "solved" but definitely getting somewhere thanks your responses! Updated the Graphics Driver and have RAM voltage at 1.35. Now getting 144-200 FPS on Ultra and High Graphics settings (varies depending on the game) but crashing on Ultra settings if the FPS limits are higher than 200/220 (again, varies..). Still avoiding the typical in-game overlays and performance inhibiting settings. That said, I'm not unhappy about the progress..this already looks and feels better.

glad to hear that bro

[Phénomènes Mystiques]

Originalmente postado por ThT_AZN_KiD:

Problem not "solved" but definitely getting somewhere thanks your responses! Updated the Graphics Driver and have RAM voltage at 1.35. Now getting 144-200 FPS on Ultra and High Graphics settings (varies depending on the game) but crashing on Ultra settings if the FPS limits are higher than 200/220 (again, varies..). Still avoiding the typical in-game overlays and performance inhibiting settings. That said, I'm not unhappy about the progress..this already looks and feels better.

Hooray

I am happy about the progress.

[jeffpmaxs6]

I have a similar mobo and updated the BIOS and drivers. I don’t know if it will help, but wouldn’t hurt. There is a “Kombo Strike’ in the BIOS and if set at default my cpu volts are 1.3. There’s settings from 1-3 and 3 is suppose to boost performance, but not much difference, accept the volts go down to 1.2 which is better for cooling. When you let the fps run wild is when the heat can creep up. I think 90c is when the cpu will throttle, so under 80c would be ideal. I think 240hz on the monitor is over kill and the same for that many fps. I would disable g-sync to see if that helps. If not, try 120hz and capping the fps about the same. Using After burner or thru the card if you. Then if no crash, then raise the fps and keep an eye on heat and then raise the hz to find what is causing the crashes. You only need more then 120hz/fps if you notice any difference while gaming.

[AmaiAmai]

Originalmente postado por ThT_AZN_KiD:

LowLevelFatalError [File:Unknown] [Line: 198] Unreal Engine is exiting due to D3D device being lost. (Error: 0x887A0006 - 'HUNG')

The game sees that Windows unloaded the driver from the application, or it would be better to say that the application was restricted from accessing the graphics interface .

Reasons:

1. Another application took EXCLUSIVE control of the GPU -- this can happen with overlays or older software. It can happen with recording software as well. When overlays or recording software use hardware acceleration they are essentially calling for exclusive use of the GPU if its GPU-based recording. In modern Windows 11, this is not supposed to happen but can still happen if you use older recording software or ones written by developers not following the standard. That includes game devs...

Keep in mind that the average PC has Xbox game bar running (overlay 1), Steam overlay (overlay 2), then their GPU driver overlay (Overlay 3), etc. This actually lowers performance and causes instability.

2. The system unloaded the driver because the application referenced "bad memory." 99.9% of the time a bad RAM overclock. The framebuffer appears to be corrupted or inaccessible.

3. GPU driver reported that it failed to find data (in VRAM) too many times -- linked to bad GPU overclock, including factory ones, or the VRAM really is bad.

^ Most of the time it will be those 3 issues that cause a driver to unload like that. Game Engines are supposed to stop when those issues happen because to them it looks like the hardware might get damaged.

The entire purpose of a operating system is to stop you, the user, from damaging hardware. If you do in fact have a bad OC, it's simply working as intended.

[ˢᵈˣ FatCat]

Originalmente postado por AmaiAmai:

Originalmente postado por ThT_AZN_KiD:

LowLevelFatalError [File:Unknown] [Line: 198] Unreal Engine is exiting due to D3D device being lost. (Error: 0x887A0006 - 'HUNG')

The game sees that Windows unloaded the driver from the application, or it would be better to say that the application was restricted from accessing the graphics interface .

Reasons:

1. Another application took EXCLUSIVE control of the GPU -- this can happen with overlays or older software. It can happen with recording software as well. When overlays or recording software use hardware acceleration they are essentially calling for exclusive use of the GPU if its GPU-based recording. In modern Windows 11, this is not supposed to happen but can still happen if you use older recording software or ones written by developers not following the standard. That includes game devs...

Keep in mind that the average PC has Xbox game bar running (overlay 1), Steam overlay (overlay 2), then their GPU driver overlay (Overlay 3), etc. This actually lowers performance and causes instability.

2. The system unloaded the driver because the application referenced "bad memory." 99.9% of the time a bad RAM overclock. The framebuffer appears to be corrupted or inaccessible.

3. GPU driver reported that it failed to find data (in VRAM) too many times -- linked to bad GPU overclock, including factory ones, or the VRAM really is bad.

^ Most of the time it will be those 3 issues that cause a driver to unload like that. Game Engines are supposed to stop when those issues happen because to them it looks like the hardware might get damaged.

The entire purpose of a operating system is to stop you, the user, from damaging hardware. If you do in fact have a bad OC, it's simply working as intended.

if i have crash only for one specific game, and the other games are fine, it's OC unstable or different cause?

[BurakZG]

Did you ran Memtest86?
You should always run memory test after changing RAM configuraton. If you see any error in memtest you should change memory configurarion and run again.
Any error in memtest means random crashes in OS.

[AmaiAmai]

Originalmente postado por FatCat:

if i have crash only for one specific game, and the other games are fine, it's OC unstable or different cause?

If it's only one game, it could be the developer at fault. However, if only you have the issue and no one else, then it's likely linked back to something specific about your PC.

If your screen flashes all the time, sometimes mysterious lag moving the mouse, etc. when not gaming or programs crash, then the OC is bad.

You will know the symptoms of a bad OC because the PC will not perform properly elsewhere.

But the easiest way to test is always to simply remove the OC and test. The reason why it is easy is because you can just save the OC profile in BIOS or on disk and swap it back after testing.

You should also write it down just in case something weird happens like the PC won't boot without the OC (I'm not kidding this happens sometimes, though it is a rare issue and usually a sign of a different issue like a configuration one or MB failure/bad BIOS release).

It certainly saves more time than running memtest or compute applications for hours. And the worst part about that is that there could still be an issue with the OC and those applications don't expose it.

Sometimes it really will be only one application that helps you find out the OC is bad, before others down the line do as well. That's because software developers do different practices and methods, and some methods can resist the problems OC issues cause better than others.

[ˢᵈˣ FatCat]

Originalmente postado por AmaiAmai:

Originalmente postado por FatCat:

if i have crash only for one specific game, and the other games are fine, it's OC unstable or different cause?

If it's only one game, it could be the developer at fault. However, if only you have the issue and no one else, then it's likely linked back to something specific about your PC.

If your screen flashes all the time, sometimes mysterious lag moving the mouse, etc. when not gaming or programs crash, then the OC is bad.

You will know the symptoms of a bad OC because the PC will not perform properly elsewhere.

But the easiest way to test is always to simply remove the OC and test. The reason why it is easy is because you can just save the OC profile in BIOS or on disk and swap it back after testing.

You should also write it down just in case something weird happens like the PC won't boot without the OC (I'm not kidding this happens sometimes, though it is a rare issue and usually a sign of a different issue like a configuration one or MB failure/bad BIOS release).

It certainly saves more time than running memtest or compute applications for hours. And the worst part about that is that there could still be an issue with the OC and those applications don't expose it.

Sometimes it really will be only one application that helps you find out the OC is bad, before others down the line do as well. That's because software developers do different practices and methods, and some methods can resist the problems OC issues cause better than others.

my GPU is Factory OC, i dont even know what that is,BUT on my AMD adrenaline all default, havent tried undervolt but maybe if it happens again i will,
it mostly will crash while iam browsing while gaming
for RAM yes i kinda OC it but it's just normal routine like forced DRAM using 1.35volt,so i can get 3200 i did not use DOCP, or should i use DOCP? hell i even created a post for my problem but because the problem was only for that specific game, i blame the game and jus uninstall it lol , and i tried just using 2666mhz stock speed, yet again still crash

for Motherboard, it's a brand new,and max BIOS, and even with, 5minute i playing the game it will crash again so i believe it might be OC problem
for GPU i bring it to the asus they check no problem at all, there is no way they are lying right

still learning how to CPU OC, so i can deselect all those auto OC/ PBO stuff

[ThT_AZN_KiD]

Thank you guys for all the suggestions and assistance. In addition to the changes I made in my most recent post, I've purchased new RAM and am underclocking my GPU and the problem is solved! Appreciate you guys.

[jeffpmaxs6]

Originalmente postado por ThT_AZN_KiD:

Thank you guys for all the suggestions and assistance. In addition to the changes I made in my most recent post, I've purchased new RAM and am underclocking my GPU and the problem is solved! Appreciate you guys.

That’s cool, but did you have your card over clock? If not, then that’s not a good sign. You should get the clock you paid for and maybe send the card back.

But, it’s also a good sign it might need more volts to stop the crashes at the higher clock. But, search a safe volt to go up to. I had to do that to a Nv card, it eventually died though. You might be alright though or just prolonging the death either way. Unless maybe a BIOS update addresses it.