Computer Crashes During Long Gaming Session

[strong_s5470]

OK Guys,

 

I've had this long term issue where my PC will intermittently crash during a long gaming session (4+ hrs). I can also replicate it by putting it under a GPU heavy workload like F@H with GPU only enabled to isolate the issue and let it run overnight to find it had crashed the next morning. I have sent the GPU to the store i bought it from and they could not replicate the issue. I have also tried a clean install of the drivers etc but I cant seem to shake it or determine exactly which component is causing the issue. In windows reliability monitor there is a Hardware Error LiveKernelEvent Code 144 error and most of the time the Radeon Software: Host Application stops working. The crash can either be the screen just freezes and you need to power cycle or the PC will completely crash and restart. I cant remember seeing a blue screen error. This has been across multiple AMD driver and Windows versions and Bios. I have also run mem test and other hardware diagnostics on my drives etc which all seem to be ok. I have not seen it crash while F@H on CPU only.

 

Any help would be good, if someone knows a diagnostic tool that might actually give me usable info that would be great

 

Specs

CPU: i7-10700K (no overclock but limits maxed)

Motherboard: MSI Z490-A Pro

GPU: PowerColor 5600 XT Red Devil

RAM: 16GB Corsair Vengeance RGB Pro 3200 (2 sticks, XMP enabled)

PSU: EVGA 650 G+

[Radium_Angel]

Sounds like it's overheating.

Or the GPU gets dodgy under long term stress loads.

[Hulkstern]

I second @Radium_Angel, Though to go into a little more detail... I might recommend you download MSI Afterburner or GPUZ. Both have sensor logging where you can see what your temperatures are at over time. You should try and run something intensive like you mentioned in your post, then keep an eye on GPU thermals (for a 5600xt there is core and junction temps). From what you said it does seem like a GPU crash. If the temps seem normal (aka, around 70-80c core, 90-100c Junction) make sure you do not have any sort of overclocks applied. If none of these suggestions help, check your warranty and see if you can RMA your card, as it may be an issue with the actual card.

[strong_s5470]

Yep, tried that. Temps all ok, within those ranges suggested in HWinfo. No overclock applied and have already RMA'd the card where they could not fault it while running Heaven overnight, they only ran it one night though. With this being an intermittent fault it's hard to capture. I suspect it is the GPU being dodgy over long term use but haven't been able to definitively prove it. It might happen after 4-5 hours, it might happen after 8-10 hours. Does anyone know of a diagnostic tool that might capture some logs that are actually useful instead of generating ambiguous codes that don't narrow the issue down?

[Hulkstern]

21 hours ago, strong_s5470 said:

Yep, tried that. Temps all ok, within those ranges suggested in HWinfo. No overclock applied and have already RMA'd the card where they could not fault it while running Heaven overnight, they only ran it one night though. With this being an intermittent fault it's hard to capture. I suspect it is the GPU being dodgy over long term use but haven't been able to definitively prove it. It might happen after 4-5 hours, it might happen after 8-10 hours. Does anyone know of a diagnostic tool that might capture some logs that are actually useful instead of generating ambiguous codes that don't narrow the issue down?

Have you tried a different PSU? Or better yet, do you have an old card like a gtx 480 (essentially just something really power hungry) you could throw in your system and try the stress test again? It could be that there is an issue with your PSU, and the problem may only show up after continuous load on whatever rail your GPU is on. That's just a theory though. I'm not familiar with the sensors available on the 5600xt, you may also be able to check to see what voltage the card is receiving. If that voltage is sagging over time, that could cause some instability.

[strong_s5470]

Unfortunately I don't have a spare PSU or GPU. I have tried to borrow one but it turns out it didn't work and that was the only person i know who had a spare card.