Crash to black issue

[Ravin1337]

Hi,

 

I recently received my TUF RTX 3080 OC (after waiting 6 months) and all was well for a couple of days, until I started getting random crashes.

 

These crashes took 2 forms:

- Both of my screens would freeze - no sound, no picture movement, peripherals unresponsive

- Crash to black and restart

 

After using HWMonitor and ASUS GPU Tweak II over a few days, I think I've narrowed the issue down to the GPU.

 

In my most recent crash (10 minutes ago), it crashed to black and restarted. GPU Tweak II opened with Windows and showed 2057MHz on the GPU Speed Clock, which is unusually high and also outside of the predetermined settings for the particular game (it hates OCs so I clock it back to factory default just for this game). Now checking the monitoring inside GPU Tweak, the GPU Clock shows as 1845MHz while GPU Speed shows as 1710MHz and the Performance BIOS is set on the card.

 

So far, my system has only crashed when playing this particular game (FFXIV), no time pattern.

 

Any support with this issue is greatly appreciated.

 

System Specs:

R7 5800X
ASUS TUF RTX 3080 OC

MSI B550-A Pro

Corsair H115i AIO

1TB MP600 NVMe SSD

2TB Seagate Barracuda HDD

750W Corsair PSU

Edited June 15, 2021 by Ravin1337
Change the title to be more appropriate to the issue

[DariusBiceps]

I am not an 100% level expert on the 30 Series cards, but my guess would be that your PSU is a bit overloaded, as ASUS reccomends a 850W unit for the TUF-Series 3080 OC.
To be precise, my thought is that FFXIV puts so much load on your 3080 that your PSU cant handle it. More so Linus told in the early 30-series days, these are the first GPUs were even he would go one step above the recommendation.
Still I would love to hear some other opinions on that.

Edited June 13, 2021 by DariusBiceps
Typing errors corrected

[Ravin1337]

After putting a new 850W PSU in, I still crashed into a restart but this time it was after a game of Rainbow Six Siege (Nvidia drivers crashed twice during the game, never happened before on this system) when I got back into the main lobby and was tabbed out into my browser.

 

Things I've tested:
CPU - Furmark -  Stressed to 90c for 5 minutes under AIO cooling (2x120, 2.5k RPM each) and no crash - PBO turned off as a test and still crashes
GPU - Furmark - Pushed to 1950MHz for 5 minutes, no crash
RAM - MemTest86 - All 16GB tested for 30 minutes in cycles, no errors - Memory is at 3600MHz
SSD - Running at 50c max, 20c below Tjmax - no loading issues

HDD - Running at 43c max, far below Tjmax - not much on this drive anyway

OS - I've run DISM and SFC - DISM found nothing, SFC fixed a few files but the issue persists

Things I need to test:
Motherboard - how do you test a motherboard with software?

PSU wiring - I used the old wiring, could be crappy wires supplied with the first PSU

GPU driver - crashing twice just before a crash to restart makes me think it's this - would the system fully restart if the driver stopped working?

 

A few things I've noticed:

The rear case fan in SYS6 spiked to 10k RPM in the Max column of HWMonitor, which isn't possible on this 140mm fan - potential short from the fan header?

Event Viewer's system shutdown error, Source: EventLog, ID: 6008 shows a shutdown time of 16:56:23 when I crashed at 17:31 - why is this? is this an issue?

 

Are there any other things I can look at that might resolve this issue? Nothing appears in Event Viewer after these crashes aside from the usual ID 41 power loss and ID 6008 listed above so I don't have much to go on.