i heart that Windows supports only 8 nvidia cards, how do you run a 9 cards rig?

I reconsider this idea and found that it is perfect. I've bought a Raspberry Pi and a 8 way relay board to run it.

I'd like to run a web server in the Raspberry Pi, the rigs report their status (including gpu, cpu ,memory, disk usage, miner hashrates, temp, fan speed etc) to the Raspberry Pi at regular intervals. If a rig doesn't report status in minutes, Raspberry Pi knows the rig was dead, restart it.
We can add some web pages to show rig status and add some button to restart the bigs by hand.

I've never learned the go language, if it was nodejs, I can contribute codes .

not a big deal, I'll add rm ~/.ssh/* to my checklist.
what's the password of root? need to change  it as well.

i don't think adding default ssh credentials is  a good idea. my dev machine is mac, i can add it easily by hand, just run ssh-copy-id. But it's more difficult to remove it. My rigs  are  30 miles away home, so they must be opened up to Internet to be managed.

thanks, I'll give it a rty this weekend.

cannot kill 3main, use pkill -f 3main.

I'll try  alexis miner on my 1080ti

nice, it's difficult for me to buy a  manageable poe switch.

in 3main
I know it is used to kill the miner process, but what does -e mean? Cannot find this argument in

use websocket to consume the api, there is an example under api folder in the source code.

You can connect to my rig via ssh.
I think the most important thing is hashrate, if you have ideas to improve hashrate, try them at first.

i can give you the ip address and password of my rig to test if you need, 3 1080ti and 1 1080.

almost the same hashrate, if you can decrease dev fee to 1%, I'd like to use yours

specs:
B85 mobo
G1840
4G ram
3*gtx 1080 ti cc+120 mc-1000
1*gtx 1080 cc+150 mc -1000
ubuntu 16.04 x64
cuda 8.0
driver: 384.59

EWBF 0.3.4b:

GPU0: 454 Sol/s GPU1: 672 Sol/s GPU2: 687 Sol/s GPU3: 679 Sol/s
Total speed: 2492 Sol/s
+-----+-------------+--------------+
| GPU | Power usage |  Efficiency  |
+-----+-------------+--------------+
|  0  |    153W     |  2.97 Sol/W  |
|  1  |    203W     |  3.31 Sol/W  |
|  2  |    198W     |  3.47 Sol/W  |
|  3  |    201W     |  3.38 Sol/W  |
+-----+-------------+--------------+

yours:

>  GPU0 Sol/s: 454.4  Avg: 454.4  I/s: 240.1  Shares/s: 0.000
>  GPU1 Sol/s: 684.6  Avg: 684.6  I/s: 365.5  Shares/s: 0.000
>  GPU2 Sol/s: 683.4  Avg: 683.4  I/s: 367.8  Shares/s: 0.200  ++
>  GPU3 Sol/s: 686.0  Avg: 686.0  I/s: 365.4  Shares/s: 0.000
   ==== Sol/s: 2508.5 Avg: 2508.5 I/s: 1338.9 Shares/s: 0.000
   GPU0 Sol/s: 433.7  Avg: 444.1  I/s: 237.4  Shares/s: 0.050  +
   GPU1 Sol/s: 668.1  Avg: 676.4  I/s: 362.4  Shares/s: 0.000
   GPU2 Sol/s: 689.6  Avg: 686.5  I/s: 364.5  Shares/s: 0.100
   GPU3 Sol/s: 677.5  Avg: 681.8  I/s: 362.2  Shares/s: 0.150  +++
   ==== Sol/s: 2468.9 Avg: 2488.7 I/s: 1326.5 Shares/s: 0.150
   GPU0 Sol/s: 445.0  Avg: 444.4  I/s: 237.3  Shares/s: 0.033
   GPU1 Sol/s: 670.9  Avg: 674.6  I/s: 361.3  Shares/s: 0.067  ++
   GPU2 Sol/s: 680.8  Avg: 684.6  I/s: 363.7  Shares/s: 0.100  +
   GPU3 Sol/s: 662.1  Avg: 675.2  I/s: 361.6  Shares/s: 0.133  +
   ==== Sol/s: 2458.8 Avg: 2478.7 I/s: 1323.8 Shares/s: 0.133

my kaby lake g4560 is running with 12 GPUs, no issue.

I run the latest version several minutes on GTX 1080ti, cannot say any hashrate improvement from EWBF_'s.

just a small tip, DO NOT use the buggy m.2 to pcie adapter.

GPU's MODEL NAME: maxsun super Jetstream GTX 1080ti
RAM:11 GB DDR5X
CORE/MEMORY CLOCK: 1797/5005
HASHRATE:670 Sol/s
MINER SOFTWARE: EWBF's CUDA Zcash miner 0.3.3B
OS: Ubuntu 16.04 x64
TDP(OPTIONAL): 200 watts

the script is awesome, but windows bat file is hard to  write and read. I'd suggest use python, nodejs or other script language to rewrite it, that make life easier  and could also support linux.

nvOC linux, gtx106 3gb samsung memory, 24mhs @75watts, cc 0,mc+1400.