I setup my rig on Windows 10 to modify and test GPUs. Windows saw all 8 cards and I let it run overnight using Phoenix miner. Zero issues…8 cards.
Installed HiveOS and I get errors and it crashes after a minute of mining. I start with 1 card and test it out then add another etc…4 cards work great, add 5th card and it crashes again. Doesn’t matter what port I use or what card. I’ve swapped everything all the way down to motherboard. Why would this work perfect under windows and not HiveOS? Attached are some error screenshots. Mind you Hive works great on 4 cards but add the 5th and it crashes.!
Hi kewing, I have the exactly same issue but in my case is when i add the 6th card on the system… did you found any solution for this?
Not yet. I put windows 10 back on for now so I can at least mine. Can’t imagine this is the first time this has happened?
whats your rig’s hardware? which mainboard are you using?
Zotac mb, 4-ASRock 5700 XT, 1-MSI 5700XT MECH OC, 3-MSI 5700XT Gaming X, 16GB RAM, running off an SSD. Thermaltake 1200W psu, 1200W server psu, HiveOS 0.6.172 image
try beta image
Wouldn’t the full release version be better to have?? Just asking.
Beta didn’t work.
What did work was start it up with all the cards connected, wipe out all the clock settings and letting it download the new config. After that I went through and did my OC settings by just changing the max wattage. Everything is working great now. Wow, what a ride that was.
6 Likes
OMG, can’t thank you enough!
This fixed my troubles with getting all 6 GPUs online. I looked for answeres everywhere for about a week and would have never thought of such a simple solution. I did a similar thing of erasing OC and just limiting voltage but not in the order you proposed so I never looked back at it until I saw your post. Did it exactly as you stated and all of a sudden it worked! Thanks again!
1 Like
What did you do?
Booted up (without applied flightsheet, but maintenance mode would work too), then opened OC for all card, deleted everything.
Then applied undervolting (850-900mV, so the cards don’t hash at full power and short the circuits) and fan speed (optional, I do 60%).
Then check all setting through amd-info in the rig itself (or via remote access).
Make sure the setting are correct, because for some reason they don’t always update as you expect them.
Then you can try start mining and if it works start adding other OC settings and play with the voltage etc.
I figured when I try to undervolt to 850mV it doesn’t usually work and either pushes it to 750 or 950, depending on other settings (aggressive undervolting sometimes helps), so check the real info with the amd-info command regularly, at least from start. Sometimes you have to set all the voltages or even core/mem state.
Also my XFX RX 580 (Samsung memory) don’t really like 850mV so far, so I’m playing with the OC setting at 900mV, which hasn’t crashed yet. I’ll play with lowering the voltage later if it allows.
This is where I’m at so far while system still being stable (applying slight changes one at a time with active miner)
and amd-info shows matching settings (same for all cards)
Good luck
3 Likes
This needs to be shared far and wide… so much frustration and then bam just like that. Fixed. This is what I did exactly.
Delete all overclock settings for GPU’s on Hiveos.farm for the rig you are working on.
Start rig
When Hiveos runs I ran the firstrun -f command.
Hiveos runs through the configuration.
All cards recognized.
Miner loads and finishes configuration without freezing and now my last gpu works, nothing crashes, go back to hive.farm to adjust your settings on your cards now.
All cards work now yay!
I would thank you all for helping me with a hopeless situation. I had 4 GPUs working rock stable, when adding the 5th GPU the system would be locked in a restart loop, unable to start hiveos properly. So I was mining with just 4 cards for the last month and a half. The 5th card would mine normally if I kept just 4 GPUs on the system.
What I did:
I cleaned the OC profile, replaced the pcie riser adapter (the part that goes into the pcie slot) and also aligned it, so it would be go straight on the slot, not bending laterally.
I suspect my problem was a combination of defective riser with the OC profile interfering with the card recognition and an adapter that was slightly bent laterally by the usb cable. When I previously replaced the riser, the system would start correctly, but would fail to recognize the 5th card properly (recognized as GA104, or something like that), probably due to the OC profile.
Hope my experience helps someone else with this kind of problem.
Many thanks!! Exact same problem and now all good. Great job.
Wow this thread has saved me from an insane headache. Had the same issue where 4 cards were running stable on NBminer and the rig would crash any time I tried to add any of my other 5 cards. Have been dealing with this for the last two weeks. However, whenever I enable the ETHpill I get the same errors as before and my rig crashes. At least all of my cards are running now, just need to figure out the issue with the Ethpill.
when you say max wattage you mean power limit or core voltage and memory voltage, please let me know. I also have same problem my 5th and 6th gpu always crash within an hour. I tried everything
Correct. Power limit. Clearing OC helped big time then slowly adjust from there
Thanks, I have rx6600 cards how do I know how much power limit I should set. I mine etash and ton
as you can see GPU4 and 5 are sitting idle doing nothing since I disabled them in the lolminer because they keep crashing every half hour. Also should I stop the miner when I add the Power limit to each GPU and restart back once its added, what about the clocks, should I leave them empty?
This topic was automatically closed 416 days after the last reply. New replies are no longer allowed.