Radeon RX Vega 56 causes Crashing/Rebooting

[KennyThe1]

Hi, I wanted to try flipping PCs so I bought some used PC components. One of them was SAPPHIRE PULSE Radeon RX Vega 56 8GB and when I start a game the PC crashes/reboots (the GPU seems to turn off but the rest of the PC stays on and I don't see anything on the screen), when the crash/reboot happens it makes kind of a pop sound. I also installed the newest version of bios and the GPU drivers.
Here are the specs:
CPU: Ryzen 5 3600
GPU: AMD Radeon Rx Vega 56 8GB
mobo: gigabyte a520m h
PSU: gigabyte p650g 80+ gold
ssd: adata legend 800 500gb
ram: Crucial Ballistix DDR4 3200MHz 2x8gb CL16

At first, I hoped the problem wasn't the GPU so I tried another PSU but that didn't seem to help so I changed the GPU for a working gtx1650 and it showed me that the GPU unfortunately was the problem, I even tried the GPU in another PC and the problem was still there. So what I am basically asking is if anyone has encountered a similar problem, if yes, is the GPU fixable or should I just give up on it and move on?

 

[RJARRRPCGP]

Any WHEA Logger events in the event log?

 

[Dr. Dro]

This was likely a mining card that was run to the bone. Since it's an HBM card, it's generally not serviceable. If the core went bad, then it's done for.

 

[LabRat 891]

Probably ran ragged.
Might have a chance w/ a V64 cross-flash to get a higher stock vCore
(assuming it's degradation-related).

Can also try editing PPTs to underclock the HBM and core, to see if there's any point it's stable @.

WX9100 Power mods - help w/ powerplay tables

I've got a MI25 I've successfully reflashed into a WX9100 16GB, and have it well-cooled. I'd like to work on restoring the original MI25's power levels or higher. There's at least 1 other TPUser (DR4G00N) that has the same kind of setup, and was successful in modifying the power limits...

www.techpowerup.com

 

[Jism]

Yep; Vega's 56 often had a 64 bios flashed into it. Because of that higher clocks and / or voltages could be archieved, but HBM would degrade that even stock clocks at some point would not hold up. Downclock HBM to half of it speed. If that does stop the BSOD's or chrashes then you found the problem.

Fix? Nothing. Flash back to Vega 56 Bios and downclock HBM.

 

[Bagerklestyne]

Yeah I flashed my Vega 56 to a 64 bios and instead of ram stock 800 overclock limit to 945 I got it stable at 1100 for several years even with an undervolt.

They were definite good mining cards in the time and more than likely yours has been subjected to it.

check the voltages as well but yeah, underclock the ram and gpu clocks and see how you go with lighter game titles (esports crap that runs on a toaster)

 

[delshay]

This was likely a mining card that was run to the bone. Since it's an HBM card, it's generally not serviceable. If the core went bad, then it's done for.

Your dead right. ..I bought a second hand Vega Nano & it keeps cutting out. Reducing the clock to lower than stock seems to fix it, but there's a small glimmer of hope because whatever is cause it to cut-out is heat related. It's not the core/memory or the main MOSFET, but I have a sneaking feeling it's the 5v regulation. I've not had time to measure/confirm as I'm to busy trying to complete hardware mod on R9 Nano.

@OP

Do the natural thing & check thermal paste/pads, changes here did nothing to fix the issues, but there's no harm in checking if the card has been opened in the past. If the card has never been open I would send it back if possible.