It's not bad power, that wouldn't cause one card to flake out. It could be a bad riser, cable or the little PCIE chip that plugs into the slot. I've had bad risers that run fine for days or weeks and then all of the sudden cards will drop off. I hate to be obvious, but have you updated the BIOS and tried to set the Gen settings for the slots etc? I have one ASUS board and the rest are ASROCK H110 pro's. The H110's are setup for mining and fired right up, I have 6 10 card rx580 rigs running SMOS. The ASUS board is 8 cards and was kind of a PITA to get running. Had to set everything related to PCIE to Gen1, disable SATA controller, audio controller etc. Finally it has been running for over a week. I read ton's on boards before setting up my 10 card rigs and stayed away from the B250 as it seemed to have overheating issues and just be unstable compared to the H110 or another board. Hell, I paid $129 for my Asus Z270-P and it runs 8 cards...
Thanks for your answer.
In the B250 bios there is a mining mode where everything gets set properly right away (gen1, above 4g decoding). I turned off the audio already but I can try and turn off all the things I don't use.
Tonight I'll do that and also embark on a mission to identify the faulty riser, if any. I don't want to run with 1 risers at a time to identify the faulty one... I guess I'll try removing them 1 by 1 instead even if it might not be as accurate (if I have 2 faulty ones).
Otherwise I already ordered a new 6 pack of risers earlier this week since I have 3 more GPUs on the way.
Make sure you aren't trying to boot with a monitor plugged into one of the mining cards on B250. Also I found that mining mode doesn't really get you exactly where you need to be. You still need to set all your pcie to gen2 enable csm(I think that's what it's called). And make sure to enable onboard video.
Thanks.
I use onboard video already. I had to enable CSM to even be able to boot so that's the first thing I ever did.
I played around with the PCI settings.. not sure if I tried GEN2 already.
I'll test that AND I'll start writing down everything I do otherwise I'll end up doing everything twice while forgetting other things to try
Update to my situation.
I replaced 1 PCIe riser, just the small part you plug in the PCIe slot on the mobo, it had something that did not appear normal. The board was not showing that PCIe has having something plugged in but SMOS was detecting the card and using it fine (normal hashrate). I replaced it to be safe not before it was not working.
But this fixes nothing.
I have 6 GPUs plugged with 6 risers.
I tested with GPU 1-2-3-4 (with their respective risers unchanged) for 12 hours and everything runs perfectly.
Then I tested with GPU 3-4-5-6 (with their respective risers unchanged) for 12 hours and everything runs perfectly as well.
Am I right to assume all these risers are fine because of this?
In fact I pretty much know they are good, I used them to mine smoothly on a Z170 PC Mate with 5 GPUs for weeks. (all 6 were used, I switched some around).
However the moment I plug 1 more GPU (number 1, with the same riser it's been tested with before), I have 4 GPU mining normally and 1 mining at 4Mh/s.
I tried both Gen1 and Gen2.
I disabled all the Sata but the one I use.
I disabled the audio controller.
I have CSM enabled.
I use on-board graphics, not PEG.
I have a fan pointing straight at the little chip that can get hot (in the middle of the PCIe slots).
Any help is appreciated.
Problem might be solved. It's been running stable for 20 hours so while it's still early to claim victory it's still a very good improvement compared to my previous situation.
What I did to fix it was to turn the damn Mining mode OFF. After that I set all the bios settings required for mining like regular motherboards require:
Basically: GEN2, Above 4G, disable everything not used.
If the problem was really with that I am not happy for Asus to have provided a buggy Mining mode.
