Mining configuration and automation for Ubuntu 16.04 with Nvidia GPUs ( GTX 1060 and 1070 )

Step by step guide

1. Assemble the rig with all graphics cards and make sure they're all "spinning"
2. Install Ubuntu desktop from a bootable USB stick by following the on-screen step-by-step guide. If the Ubuntu installer doesn't start, hit F6 during boot and select nomodeset from the options menu in the bottom right area.
3. Make sure you check the "Install third party software for graphics..." - this will preinstall the nvidia drivers on your system. Once the OS is installed, fire up a terminal and install git. sudo apt-get install git && sudo apt-get install openssh-server && sudo apt-get install screen && sudo apt-get install htop && sudo apt-get install curl
4. Go into your router management console and forward a port to port 22 on the mining machine IP address. For example, I forward port 2017 on my router to port 22 on local address 192.168.0.69 which is the reserved IP address of the miner.
5. Edit the /etc/ssh/sshd_config and where it says PasswordAuthentication no change it to PasswordAuthentication yes. This is not the most secure option, but if you know this you will also know how to set up key based authentication over ssh.
6. Press ctrl + alt + f1 to switch to a text-only terminal.
7. Run lspci -v | grep VGA to make sure the system sees all 12 GPUs... if it doesn't something is wrong with the build.
8. Update the repository database and install any updates: sudo apt update && sudo apt upgrade... then restart: shutdown -r now
9. If you haven't already installed the nvidia drivers (checked the checkbox from step 3), then you will need to install them manually: Add the Nvidia drivers repository: sudo add-apt-repository ppa:graphics-drivers/ppa and continue with the next instructions, otherwise jump to step 11.
10. Go in text mode ctrl + alt + f1 and install the current Nvidia drivers: sudo apt-get install nvidia-375 (or nvidia 384). If the install fails or anything else goes wrong you can always run sudo apt-get purge nvidia-* to return to the previous step
11. Go into the "Software and Updates" application -> "Additional Drivers" tab and enable the nvidia proprietary driver. Then restart!
12. Select the Nvidia GPUs sudo prime-select nvidia and enable them all for over-clocking nvidia-xconfig --enable-all-gpus
13. Create a folder somewhere in you home directory and clone this repository... cd YOUR_FOLDER && git clone https://github.com/the-codepunker/ubuntu-nvidia-mining.git .
14. Make a copy of the default Xorg configuration file sudo mv /etc/X11/xorg.conf /etc/X11/xorg_bak.conf
15. Enable full over-clocking capabilities for your GPUs sudo nvidia-xconfig -a --cool-bits=31 --allow-empty-initial-configuration
16. Execute the commands below to populate the xorg.conf file with the required parameters for enabling over-clocking:
    • sudo sed -i '/Option "AllowEmptyInitialConfiguration" "True"/a Option "ConnectedMonitor" "DFP-0"' /etc/X11/xorg.conf
    • sudo sed -i '/Option "ConnectedMonitor" "DFP-0"/a Option "CustomEDID" "DFP-0:/etc/X11/dfp0.edid"' /etc/X11/xorg.conf
17. Copy the fake display dfp0.edid from my git repo to the X11 folder: sudo cp dfp0.edid /etc/X11/dfp0.edid
18. Make sure lightdm doesn't overwrite your xorg.conf file on the next reboot: sudo chattr +i /etc/X11/xorg.conf and restart. If everything went well... your xorg file will remain unchanged, but the system will no longer boot in graphics mode so you will need to switch to text-only mode.
19. To fire up the miner you will need to first execute my script for over-clocking located in the root folder of the repo: sudo ./occ.sh. If it works it will tell you that it changed the memory frequency of each GPU and the maximum power it can draw.
20. Finally edit the file called execute.sh and add your address instead of mine. Then run sudo ./execute.sh and BOOM.

Automation through cron - Set the crontab like this as sudo:

@reboot screen -S miner -dm bash -c 'echo waiting; sleep 60; cd /home/___USER__/__FOLDER/; ./occ.sh; echo waiting; sleep 10; ./execute.sh;'

0 9 * * * /sbin/reboot --reboot --force

Monitoring the miner

1. Rename daemon.ini.template to daemon.ini
2. Edit daemon.ini and assign the correct values to the settings in it
3. Add this entry in the root crontab with sudo crontab -e. Don't forget to replace the folder paths. */10 * * * * cd /home/[__USER__]/Desktop/eth && php ethdaemon.php >> /home/[__USER__]/Desktop/eth/miner.log 2>&1
4. You can then check the log like this: cat miner.log | grep -n --after-context=20 'Miner failed'

If you want tips on how to set up the hardware check out my post.

Donations are welcome: 0x9335fE2BCdca68407ed5Ae5FB196d2c69CAf96Da