I have 4 RTX 3060 ti GPUs. All was fine till yesterday. Since yesterday one of the GPU started going offline randomly with cuda errors/ mem oc errors. We changed Riser / pcie cable / miner. But still same problem. Since morning now card is not being detected in Hiveos. If I connect the card on Windows I get display. But nothing in HiveOS. Do I have to do anything additional to enable card ?
Appriciate for advise.
I tried with another PCie slot with another riser / usb/pcie cable and same. Let me try point no 2. I did not try that. I will keep you posted.
GPU is working in Windows, so GPU is not the issue. Today we replaced it with another GPU and same problem. So as you advised we connected 6 pin connect to both connectors on motherboard. Still no luck. Then we changed few settings in motherboard. Set all PCie slots to Auto (earlier it was Gen 1), disabled integrated GPU as we have I3 9100F. Enabled Above 4G, CSM is already disabled, and moved GPU to another slot on the motherboard (earlier we checked in slot 2/3 but failed to detect the card) 5/6. Then it worked. Now rig is online and we will test for atleast 48 hrs to see if its stable then check if its the issue with slot 2/3. Also we changed the USB cable and pcie riser card. We will check al one by one once we are sure rig is stable with new settings / slots /usb cables / risers.
Its been 3 days and rig is stable. We asked biostar support but as always they just sent list of FAQ links. Today we will add 2 more cards. While doing this we will check slots 2/3 again if the slots is the issue. So verdict is motherboard is good, risers / cables are good, gpus are good, hiveos / oc settings are good.
