GPU crash only occurring in one game at high settings

[ETHANO8225]

Ok so I figure I'll start off by saying I play on a laptop, specs below.

Spoiler

Dell G15 5530

CPU: 13th Gen Intel i7-13650HX (24 MB cache, 14 cores, 20 threads, up to 4.90 GHz Turbo)

GPU: nvidia RTX 4060 8gb GDDR6 (laptop version)

RAM: 2 x 16GB 4800MHz DDR5 sodimm, upgraded from 2 x 8GB 4800MHz ddr5 samsung memory (used manual provided to find correct upgrade, upgraded in early may)

Storage: 2 x 1TB m.2 drives, this drive was added in recently

OS: Windows 10 Pro (downgraded from win11 pro despite the loss of cpu e-core support in win11, also i've had a fresh install done in late may)

Display: 165Hz 1920x1080 

Other things to maybe note:

I turned off Hardware accelerated GPU scheduling recently to see if that could be conflicting with nvidia reflex for some reason, it didnt change anything. Nothing has been modified in bios except the disabling of the internal mic, disabling of virtualization support (for virtual machines) and I also tried to exclusively use my nvidia gpu through the bios as well. I have tried multiple different driver versions, the last and latest major release, etc etc. using DDU to remove the driver fully for a clean install each time. I also have done a Memtest on the new memory to see if there were any faults but none were found. the blue screen doesnt really provide any good information as well, I can provide crash dump if needed (although i can tell you it has something to do with directx and the watchdog process).

The issue I am currently having only relates to destiny 2 at certain graphics levels. Note, this has only been happening recently, yet I don't see anyone having my specific issue except ps4 users. Essentially, when I turn the graphics up too high*, the GPU (driver?) crashes completely, resulting in a black screen if I have my pc set to use the nvidia gpu in BIOS, and a weird freeze / game crash / switch to iGPU if the "optimus" mode is selected (basically a mode where it can select which gpu to use to save power) + shortly after, a blue screen displaying VIDEO_DXGKRNL_FATAL_ERROR appears. This happened around the time I installed new ram, albeit it worked perfectly for all my other games like cs2 at max no fps cap, bo3 at max, rocket league (undemanding but another game it worked on), etc, which is why I did the memtest I mentioned above.

*What do I mean by too high graphics?:

Spoiler

So I did a lot of testing to see what caused the crash the quickest, and it seemed to be a combo of render resolution and / or FPS. 

These are the tests I did (all at the tower in destiny 2, in the subclass menu, with a new subclass selected, at 1080p)

Highest graphics - 200fps max 200 render res - instant crash

Highest graphics - 90 fps and 90 render res - no crash after 2 mins, occasionally crashes in game when a lot of effects are present

Med graphics - 120 fps and 100 render res - crash after 65s or so 

Lowest graphics - 165fps and 200 render res - crash after 7s

same as above but 150 render res - crash after 30-40s

I've researched this issue quite a bit within destiny 2. It says there may be a vram leak 1 (there are more examples I can get if needed), although I crash at low VRAM usage, so that wouldnt make sense.

The known issues help page also says "Laptops w/ Intel 12th generation Iris integrated graphics may crash in high-demand environments. Players can mitigate these crashes by ensuring Windows 10 is set to use only discrete graphics when running Destiny 2 via Windows->Settings->Graphics Settings, and selecting "High Performance" for Destiny 2." This has already been done, and on top of this, I tried to run in nvidia-gpu-only mode, to no prevail. I've done stress tests using OCCT, didn't run it for 24Hrs, but I ran it for 1H with no detected issues. Considering its only with this one game, I can tell it's an issue with the game, but the fact that I'm the only one experiencing this issue (except ps4 users??) makes me very confused.

After all the testing I've done I am just confused. Destiny isn't installed on the new SSD, and games on the new one don't crash. Everything else at max graphics does what its supposed to: thermal throttle, and reduce fps, as I'd expect. I don't know where to go from here, as I used to be able to run at 165Hz, high settings, 100 render res, with no crashes. Any ideas?

[Dreamer]

Have you checked your Event Viewer for more details on your crashes/errors?

[Salted Spinach]

Looks like either the GPU or RAM is unstable

 

To test the GPU, lower its clocks

 

To test RAM, disable XMP

[ETHANO8225]

On 6/16/2024 at 6:53 AM, Salted Spinach said:

Looks like either the GPU or RAM is unstable

 

To test the GPU, lower its clocks

 

To test RAM, disable XMP

I tried to disable XMP but I've checked the bios and found no settings related to it. It says the max non-xmp ddr5 speed is 4800, so i'd bet it is just running at its max speed w/ no xmp required.

For the GPU i want to try and lower its clocks but honestly don't know how to safely do so on a laptop. The BIOS is pretty barebones even in advanced mode, there is an overclock section but its only related to the CPU (which has always been disabled). If you have any ideas/past experience, I'd greatly appreciate any tips

On 6/16/2024 at 2:19 AM, Dreamer said:

Have you checked your Event Viewer for more details on your crashes/errors?

Yes, I have.. although there is nothing consistently showing up when I crash. Occasionally it'll show something about AUDIODG crashing, sometimes it shows a LiveKernelEvent with 0 log info attached (quite literally - 1.3million lines of whitespace across 3 log files).

Is there any way to like, log any new logs in eventvwr to a separate area to view them for later? or a "show logs from x time to y time"?

sidenote - I've tried to install the newest "supported" drivers from dell - I believe nvidia driver vers 552.12 (about a month old), to no avail.