Question GPU Issue? HWinfo log included with picture.

[azitello]

I'm trying to understand what and why...

So, to summarize:

I'm fixing a friend's PC, I believed it to be a GPU issue initially, but then back tracked after the GPU worked fine in two other systems with testing. Included hwinfo log of latest crash - help me rule out gpu, but looks like it's gpu, just need an explanation of what or why it's gpu. I can take gpu apart and check vrms if needed.

hwinfo log csv: https://file.io/wT6zFf8ylmdK

In logs, gpu voltage, power, clock, etc drops big when I close out of furmark then crashes.

The issue: System will ramp GPU fans to 100% and black screen lock. Only option is to hard reset.

To create issue repeatedly: Run Furmark for about 2 mins, until GPU is 65c, close furmark, then when gpu starts normalizing after the stress test (about 6-10 secs after closing furmark)... it causes problem

Problem PC:

3080ti, 5900x, 850w a-tier evga

I've changed motherboard, ram, psu, ssd's, only thing not changed is cpu.

This system, after all of these changes, still crashes. About to swap CPU to see. Will update.

Other components used in testing on problem pc:

skill ram = same problem (dual and quad on diff mobos) , superflower 850w = same problem , multiple mobos (b550 and x570.. unify, tomahawk, vision) = same problem , ssd's and even a hdd (diff pcie m.2 slots) = same problem.. again, just not swapped cpu yet (will update)

using a 3060ti and 3070... no crashes in problem pc, just the 3080ti crashes

However... when I use 3080ti in my rigs, it's fine, can't crash. my rigs are 5900x on x570, I've tried different configs. and on an i5 something.

Initially, I leaned towards gpu, then psu... but .. odd behavior of neither crashing in my systems. temps and voltages seem fine, until crash, maybe someone else with better eyes can see the issue from log of most recent crash.

I just got another mobo because system worked fine on my mobos.... but.. new mobo in problem pc, crashes.. again, tested diff configs.. could it be the pc case? idk at this point, but please, someone review the hwinfo logs?

extra: one more thing - if I limit gpu to 60% in afterburner, no crashes, haven't tried 70%, 80%, etc

image of some of hwinfo log i'm guessing where problem occurs:

https://prnt.sc/ruoTFDYse6lt

 

[Lutfij]

Welcome to the forums, newcomer!

3080ti, 5900x, 850w a-tier evga
You should list the specs to all of your system's like so:
CPU:
CPU cooler:
Motherboard:
Ram:
SSD/HDD:
GPU:
PSU:
Chassis:
OS:
Monitor:
include the age of the PSU apart from it's make and model. BIOS version for your motherboard at this moment of time.

EVGA as a brand isn't bad but they tend to have hit or miss units within their own top tier lineups, which is why you're advised to source(borrow, not buy) a reliably built PSU and power the system/HGPU to see if the issue persists.

superflower 850w
You've mentioned the brand and the advertised wattage. What is the model of the unit and it's age? What did it power prior tot he RTX3080Ti?

It could be a lack of power to the GPU, it could also be an overheating problem on the power delivery area of your car. See if you can replace the thermal pads and paste.

Did you use DDU to remove all GPU drivers and manually reinstalled with the latest version sourced from Nvidia's support site?