Ok, thanks for the confirmation. I first tested it out with just one 390x connected via 16x riser, and that worked great. Then, I installed TeamViewer, shut down, connected the other GPUs via 1x-16x risers, and restarted (without a monitor plugged in. I thought I could just then use TeamViewer to connect and start up the mining software). After work, I'll try restarting again but with a monitor connected during bootup to see if the other GPUs get started.
I've got 2x 1000w platinum PSUs to power the 6 gpus. Should be enough. While testing though, I've only got the 3 connected with 1 of the PSU. I think 1000w is enough to power those 3 cards though...
No Prob.
The 390x's shoudl draw stock about 250watts each, the 280x's about 230watts or so. (as long as they are good quality cards).
'Most' 390x's you can undervolt and get back about 10-20% wattage, some more, some less.
I run all my 390's (non 'x') at 100mv undervolt, 1100/1500 settings (stock is 1015/1500) and they use 200-210watts each while getting around 29MH+ now.
Solo mining has been earning me 50% more than pool mining for the past week so I'm sticking to it. Some days I get 1, some days I get 5 blocks.
Well shit. So I got home, plugged in the monitor (with all 3 gpu's connected, 2 via 1x-16x risers), and booted. Still, only the one connected to the 16x pci slot powers on and is recognized. I've got the molex power connector on the mb plugged in, and all 3 gpus have dedicated power which I made sure what all connected. My PSU is an evga "supernova", so no issues with psu quality I think. I know it's not the gpu's because I swapped out one of the ones connected via the 1x-16x riser to the 16x-16x riser, and that one powers up just fine. Only the ones connected via 1x-16x. Anyone have tips for how to troubleshoot this?
Edit:
MB is an ASROCK H81 pro btc,
2x 280x are both Sapphire,
4x 390x are all MSI.
i'm using windows 7 64bit
1x to 16x risers
old 80gb hdd's , 2gb ddr2 RAMs
good old socket 775 cpu's and motherboards @ 4 cards each. all fine, i'm using only one monitor that i plug if remote access troubleshooting is not possible.
sometimes it is better to have at least two rigs, a few extra risers and a spare PSU that you are very sure that is 100% working. swapping components to compare and check to know what component have trouble.
from experience i've seen motherboards that cannot detect more than two or three GPU's (know that i'm using cheap old boards), there are two 1x to 16x risers here that makes my gpu's undetectable (this have been used for quite a while).
it is also good to have a GPU temperature monitoring software around. to know what card is heating more etc.
last day one of my miner hanged twice, unplugged my hottest card, found out the my high speed hydraulic fan is not smooth anymore (rotating) (from modded old sapphire 280x), removed the cover underneath the sticker, put some compressor oil (just lying around here) returned the cover, seal it with a piece of masking tape (fan fixed-smooth again)...... and reapply thermal paste to the card because it was already dry.. and the rig was fixed and mining again