Hello all,
I have a rig that was working for 1.5 years and now I don’t know what to do more in order to get it working again
I cannot understand what /where is causing this problem, i reduced overclocks, changed PSU, I tried all 6 gpu individual and some gave the error other don’t, but if I change them in PCIe slot or riser, sometimes also give error other times not.
I’m lost.
Any idea to pinpoint the problem?
Thanks
I have exact same thing with 6x RX570 and 6x P106-90.
I forgot to mention that these errors show with phoenixminer and claymor miner, but if I get the rig in maintenance/stop miner it does not show errors, so I assume is something related with miners or what?
Thanks
Hello,
I had this exact same problem on my rig. It turned out that I had a faulty riser. Try installing one card first, and if that doesn’t give you an error, its definitely a riser problem.
I couldn’t believe that one of my risers was faulty until I swapped each one out individually.
Good luck!
I’ve found other thing, turning off one corsair fan that I had connected to the motherboard helped, not so imediatly errors and stays stable.
Now I’m at a fase that I’m turning overclocks up and found that two gpus are not liking.
Next step is to found where they are in the rig…
the sdma0 timeout indicates a pci failure to me. Probably a bad connection for whatever reason. I had this happen and got it resolved by sorting out which pcie spot on my MB wasn’t working. It could also be a riser.
You cant mine eth with Claymore since the DAG file has grown too large for 4Gb gpu’s. Try Teamredminer instead. It still works on ETH. Maybe that is your problem here.
I changed to phoenixminer, I was using claymore just for testing.
By the way, do you have any trick to identify the gpus on the rig?
I look at pci id on Hive but the layout on the motherboard sometimes are not in the order we think.
I Have several rigs with sapphire nitro + 570 4gb and all started to get unstabled, I had to turn down the overclocks that I was using for the last 2.5 years.
Could be the components to wearing down or something ?
You can identify the gpus by typing “gpu-fan-find” on the console. It will halt the mining, stop the fans, and spin up the fans on the gpus one by one, starting with gpu0.
Regaring the unstable Nitro+ 570, have you cleaned the accumulating dust, and perhaps the paste/thermal pads are dried out?
This is hardware problem.
After 2.5 years 24/7 it’s can be Power Supply (oxidation, low power on +12V line, etc )
PSU has 1 year.
It’s very time consuming to find the bad riser…
How do you guys diagnostic the risers ?
It’s easier to swap all the risers and forget it for some time maybe!
Risers are cheap. Just change them.
