Disappearing Cards - Hardware Troubleshoots No Good

As of this morning, two of my GPUs on a six-GPU rig disappeared. Tried maintenance mode and hard reboots, didn’t return.
Swapped out:

Power Supply
Cards
Splice
PCI-E Risers
USB 3.0 cables
Power cables

The funny thing is the burnout seems localized to the GPU 4 and GPU 5 spots. I’ve exchanged the cards, power supplies, everything - and whatever I put in those positions doesn’t run.

It’s not the motherboard ports because I tried those as well with the working GPUs.

The only idea I have left is to flash the bios on the motherboard as maybe the Above 4G decoding got disabled, but that doesn’t really seem like it’ll work as there’s 4 GPUs mining on it already and I know the ports are good.

At a loss here - anyone have any ideas? I’ve swapped out all physical hardware.

Spots in what?

  • Motherboard
  • PCIe assignment
  • “list as seen by the OS?”
  • “list as seen by the Miner?”

List as seen by HIVE OS. See pic.
Missing from there are GPU 4 & GPU 5, which are an RX 5700 XT and RX 6800. I’ve swapped the cards out with the ones you see running in the pic, and the cards that formerly worked, now do not in those spots.

There was no software or miner update - just dropped off out of the blue and I can’t get the rig to even see them.

Here’s the full setup by the way. Plan for tomorrow is reload the bios on the motherboard and see if that fixes anything.

Have you confirmed it is not the PCIe slots themselves by having the 6 GPUS in and removing 1&3 or 2&4 from the system?

Have you checked that PCI resources have not been consumed by the BIOS settings defaulting to PCI auto, enabling HD audio, or or or?

Good luck in the pursuit. Bummer when working goes to not working.

Yeah, I switched the PCIe slots around, so it’s not the ports. For example, taking GPU 2 & 3 and plugging them into the slots that were for 4 & 5, they work just fine.

I’m going to wipe the bios out today and start fresh. Last idea before trying to swap out the motherboard itself.

@Grea - Thanks for your feedback and trying to help!

One other interesting caveat - the cards are getting power (as evidence by LED lights on their sides) but not spinning up on startup, nor spinning their fans when the rig is running.

Well, in case anyone looks at this in the future. I fixed the issue.
What did NOT work:

Loading new bios onto the motherboard
Swapping out power cables
Checking voltage coming from wall into PSU

What did work:

Unplugging all cards plugged into the non-functioning PSU.
One by one, re-adding them into different ports on the PSU and seeing if that worked.
I noticed that sometimes, one or two GPUs would come online with the start - but not all three.
Finally, I swapped my three lowest power cards onto that PSU, moved some cables around and voila - it’s working in full.

Still the damndest thing that they just disappeared from one day to another - and all the hardware and software fixes didn’t work. In the end, I guess it’s something funky with that power supply, but still, am not really sure what the issue is or was.

Cheers all

Temps increase in your area which reduces the power efficiency?

Unfortunately, power supplies fail too…be careful.

This topic was automatically closed 416 days after the last reply. New replies are no longer allowed.