Topic: 12 GPU Mining Rig Help. (Read 351 times)

I bought them locally, not sure about their "brand" sorry. 

I have a Solution... Spilt the rigs into 8 cards... Initially I was doing 11 card rigs.... too many "problems" and to troubleshoot..... Yes it can run... but you need it to run like 20+ days without touching your rigs... I have like some rigs.. ran for 40+ days without touching it...

and above 8 cards, i find you have to use the onboard GPU with windows..  8 cards, can run with the PCIE GPU.. its mainly the Onboard GPU driver sometimes conflict with the PCIE GPU...

I am not very technical... just from experience.... 11 gpu sometimes restart for no reasons.... most likely blue screen... and if onboard GPU needed... you will need a HDMI dummy... otherwise if you set your program to auto run when windows booted.... it will run on PCIE GPU and it will stay in limbo...

I have 12 rigs... and I have limit them to 8 to save me a lot of time....

I just use those Ebay cheap 1600w PSU for the rigs to save money.... ran fine for like 10 months already lol touch wood... worse some to worse change PSU only.. A$130 each... Each rig.. has two 1600w CHEAP PSU each power only 4 cards... hence they can live long....

Some rigs have Corsair 1000W HX1000i.... but it's too expensive and will reduce my days ROI.....

Risers just use Version 006c with 4 capacitors 270uf... those with 3 capacitors and they will have smaller capacitor like 200uf type..  will give problems after 3 months of running.....  

Motherboard Asus Z270Ar or Asus Z370 A..... memory 8gb kingston 2400mhz

When you tried to add that GPU one by one, don't use the same port for all GPU. Add every GPU on the different port (each GPUs must be adding on the different port) and also use the different riser while installing each card.

do you have a prefer riser? Please link me it <3

I have a 12 x 1070 rig too with 2 x 1250w psu they are running fine.  Most likely the problem is with the risers (that includes cable to check too)  and you can try reinstalling driver / cuda to save you some time, just in case its a driver problem.

I use the PCIe x16 for the main GPU and all the rest are on the x1 ports

When I was installing them 1 by 1 I was using the PCIe x16 port

have you used the PCIe x16 port on MB when plugging in your risers or are all of them plugged in small x1 ports?

Hey, I did what you told me to do, but one of the gpu's says that it has driver issues. it only says it when all of them are plugged in. when I did it one by one they all seemed fine.

What OS are you using ? i was going to assume Win 10
What drivers are your running ?
Any additional changes you may have made ?
What trouble shooting have you currently completed ?

That's the thing, my 1070's aren't running perfectly its one of my 1070 Ti's Rigs that is running 100%. I should've said this, in the beginning, my fault.

OP didn't use AMD cards, and It was clear enough on OP's first post.
"12GTX 1070, 24 GTX 1070Ti's"

Will be better to read all of OP's explanation and also the previous comment before posting something, to keep your comments stay on-topic. 

for error 43 you must use patch to slove this problem

for gpu not found there two way ... some time you must change your driver to blockchian driver or adrenaline

or this problem come with windows 10 update

You say you have 3 Rigs and 1 is working

Are they all identical?

If yes, I was thinking on your Sh*tty risers

If not, as others posted might be your power issues, to prove this, lessen the load by installing fewer GPU's on each rig

I was mining with my 1050 tis and they were crashing and I couldn't figure out what the problem was. Until I found out that the Risers I was using were of shitty quality. I checked the PSUs before realizing that the GPUs were crashing due to not enough power. Not sure if that's your problem but you should check your risers.


Edit: just read the replies and maybe the problem is not the risers.

I second this.

@FaourZ From the net,


Some powersupplies have "shared" +12v Rail to the regular +12v for other components making the max power label inaccurate for computing the power consumption of the videocards alone.
Most major producers however, separate the rails and using only the max power of the +12v for PCI-e power as the Rating Label.

What are the brand(s) and model(s) of your PSUs?

Unless you can't underclock like the post above me stated, plug out one of the GTX1070ti to test the miner for stability.

Even at 180w each 6 of them is 1080w which is just fine for a 1300w PSU.  1080w draw on a 1300w PSU isnt shutting anything down.  Put TDP at 65% and overclock +150/+500 and the magic will happen.

I think OP was using 2 x 1300 Watt PSUs for each rig, in another word 2600 Watt each rig, that is right OP? If it right, a bit chance that power is the issue.

(They think you were using 1300 Watt for each rig) 

OP
Have you tried plugging the cards one by one? And find out how much plugged cards that giving you untroubled rig. If possible use the same wiring scheme with the rig that was running smoothly (GPUs, Risers and other wiring schemes.)

In other thread, I found someone has troubled with the same mobo causing by un-mounted additional PCIe power (the two Molex connector) May you have to check it. (but I'm not sure about it) Since we are using a powered riser nowadays.

This is also good advice must you try:

actually TDP on my 1070Ti's and P104-100's is 180W, bro. loading your PSUs to the limit isn't good thing either, better to have some power reserve remaining.

if I understand OP right - 1070's run fine, but there are issues with 2x 12x1070ti's rigs. I am pretty sure this is related to additional power that Ti's require.

yeah, I was confused when he said it wouldn't handle it, lol. Thank you for your responses and I will test this out. Thanks, my dude! appreciate it a lot.

Say what? How on earth do you come to that conclusion? Lets see your math. Ill show you mine right now.

Rig with 12 1070Ti's at 150 TDP each. (No overclock or TDP adjustments) https://www.buriedone.com/hashrates

6 1070 TI's per PSU= 6x150w=900w

PSU1= 900w
PSU2=900 + CPU/Motherboard/SATA 200w= 1100w

and each of his PSU's are rated for 1300w

I would still give them a 65% TDP Overclock  +150 clock/+500 memory and the power usage will go even lower. https://www.buriedone.com/tuning

Problems are probably a combination of BIOS setting and/or update AND/OR Driver issues during install.  Go into device manager and click on view tab then show hidden devices and have a look at the mess under Display Adapters and Sound and Video Controllers.

DDU to remove it all and disconnect all GPU's.  Install one at a time and reboot after drivers install for each GPU.  Boot back up and verify GPU installed successfully.  Then and only then shut down and install next GPU. Repeat cycle until all are installed.