[OS] nvOC easy-to-use Linux Nvidia Mining

[fk2]

Here is what works for NICE_X16R

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ZENEMYminer/recommended/ccminer -a x16r -o stratum+tcp://x16r.eu.nicehash.com:3366 -u 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK

Please default in next versions

[Stubo]

Here is what works for NICE_X16R

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ZENEMYminer/recommended/ccminer -a x16r -o stratum+tcp://x16r.eu.nicehash.com:3366 -u 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK

Please default in next versions

I think what is missing is this line in 1bash, in the NICE_X16R section:

Code:

NICE_X16R_ADDRESS=$NICE_ADDRESS

NICE_ADDRESS is supposed to be set by the user on line 439 of 1bash:

Code:

NICE_ADDRESS="NICEHASH_BTC_ADDRESS"              # Your internal Nicehash BTC addrees to use with nicehash		

... but I could be wrong. Reading the new way that addresses, ports, pools, etc. is now specified is a bit confusing:

Code:

xaddr=$COIN$UADDR

That is from 0miner.

[LuKePicci]

Yes, the address variable for some nicehash/mph/zpool coins is missing. Most of them has a line like that one suggested by @stubo directly in 0miner but I think it is best to have that in 1bash, so follow his suggestion. Doing so should let nvoc build a correct command line. Let us know if this solves the issue.

Please show me a valid command line for nice_ethash also.

[fk2]

tyvm stubo, that worked like a charme for X16R.

@Luke: unfortunately that didn't work out for bminer and ethminer. this is what it should look like:

Code:

ethminer -SP 2 -U -S daggerhashimoto.LOCATION.nicehash.com:3353 -O YOUR_BTC_ADDRESS_HERE.WORKER-NAME:x

according to https://www.nicehash.com/help/ethereum-mining

I used this here:

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ETHMINER/recommended/ethminer -SP 2 -U -S daggerhashimoto.eu.nicehash.com:3353 -O 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK.Icemine01:x

but there is no ethminer in the given path.../recommended, therefore no screen given named miner

[Stubo]

tyvm stubo, that worked like a charme for X16R.

@Luke: unfortunately that didn't work out for bminer and ethminer. this is what it should look like:

Code:

ethminer -SP 2 -U -S daggerhashimoto.LOCATION.nicehash.com:3353 -O YOUR_BTC_ADDRESS_HERE.WORKER-NAME:x

according to https://www.nicehash.com/help/ethereum-mining

I used this here:

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ETHMINER/recommended/ethminer -SP 2 -U -S daggerhashimoto.eu.nicehash.com:3353 -O 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK.Icemine01:x

but there is no ethminer in the given path.../recommended, therefore no screen given named miner

I did some more digging in 1bash and 0miner and, to me, there seems to be a big gap for support of NiceHack in 0miner. The only relevant code I could find is this:

Code:

## NICEHASH
if [[ ${COIN:0:4} == NICE ]]
then
  NICE_NEOSCRYPT_ADDRESS=$NICE_ADDRESS
  NICE_LYRA2REV2_ADDRESS=$NICE_ADDRESS
  NICE_X11GOST_ADDRESS=$NICE_ADDRESS
  NICE_SKUNKHASH_ADDRESS=$NICE_ADDRESS
  NICE_CRYPTONIGHT_ADDRESS=$NICE_ADDRESS
  NICE_EQUIHASH_ADDRESS=$NICE_ADDRESS
  NICE_ETHASH_ADDRESS=$NICE_ADDRESS

but nothing references those variables after they are set there. I think there needs to be some more logic in that section that sets the x% variables specifically for NH.

[LuKePicci]

tyvm stubo, that worked like a charme for X16R.

@Luke: unfortunately that didn't work out for bminer and ethminer. this is what it should look like:

Code:

ethminer -SP 2 -U -S daggerhashimoto.LOCATION.nicehash.com:3353 -O YOUR_BTC_ADDRESS_HERE.WORKER-NAME:x

according to https://www.nicehash.com/help/ethereum-mining

I used this here:

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ETHMINER/recommended/ethminer -SP 2 -U -S daggerhashimoto.eu.nicehash.com:3353 -O 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK.Icemine01:x

but there is no ethminer in the given path.../recommended, therefore no screen given named miner

Ethminer command line layout has changed when they introduced support for secured protocols, sounds like you are missing the latest ethminer versions, re-run the nvOC upgrade command and check the contents of the ETHMINER folder with "ls -l", you should see to which folder the various symlinks are pointing to.

tyvm stubo, that worked like a charme for X16R.

@Luke: unfortunately that didn't work out for bminer and ethminer. this is what it should look like:

Code:

ethminer -SP 2 -U -S daggerhashimoto.LOCATION.nicehash.com:3353 -O YOUR_BTC_ADDRESS_HERE.WORKER-NAME:x

according to https://www.nicehash.com/help/ethereum-mining

I used this here:

Code:

screen -c /home/m1/NVOC/mining/screenrc-miner -dmSL miner /home/m1/NVOC/mining/miners/ETHMINER/recommended/ethminer -SP 2 -U -S daggerhashimoto.eu.nicehash.com:3353 -O 39qjQtznWGrsQAcbgQCa4VjQyMonzDNSUK.Icemine01:x

but there is no ethminer in the given path.../recommended, therefore no screen given named miner

I did some more digging in 1bash and 0miner and, to me, there seems to be a big gap for support of NiceHack in 0miner. The only relevant code I could find is this:

Code:

## NICEHASH
if [[ ${COIN:0:4} == NICE ]]
then
  NICE_NEOSCRYPT_ADDRESS=$NICE_ADDRESS
  NICE_LYRA2REV2_ADDRESS=$NICE_ADDRESS
  NICE_X11GOST_ADDRESS=$NICE_ADDRESS
  NICE_SKUNKHASH_ADDRESS=$NICE_ADDRESS
  NICE_CRYPTONIGHT_ADDRESS=$NICE_ADDRESS
  NICE_EQUIHASH_ADDRESS=$NICE_ADDRESS
  NICE_ETHASH_ADDRESS=$NICE_ADDRESS

but nothing references those variables after they are set there. I think there needs to be some more logic in that section that sets the x% variables specifically for NH.

They are referenced indirectly with delayed variable expansion. As I was saying before this code block was intended to leave those settings out from 1bash but IMHO having them in 1bash, as you suggested for X16R, as like for any other coin, is the best.

[fk2]

Sure, just install cuda-9-2 following instructions about installing cuda from specific ppa packages.

Here is the detailed installation guide: https://docs.nvidia.com/cuda/archive/9.2/cuda-installation-guide-linux/index.html#ubuntu-installation
Here is the ppa (and a recap of the same install steps) for 16.04: https://developer.nvidia.com/cuda-92-download-archive?target_os=Linux&target_arch=x86_64&target_distro=Ubuntu&target_version=1604&target_type=debnetwork

Just use cuda-9-2 package instead of the generic cuda metapackage in the last step.

Hey Luke,
 I installed everything as described, skipped the examples since I don't think we need this, made an etry in etc/rc.local for the nvidia-persistenced and had a look at nvcc --version to make sure 9.2 was installed and everything looked ok. After a reboot nvcc --version shows 8.0 again although I had a apt purge for cuda* and nvidia* before installation.

you got a tip what to do? do I need to reinstall?

edit: nevermind, I didn't uninstall cuda properly with

Code:

sudo /usr/local/cuda-X.Y/bin/uninstall_cuda_X.Y.pl

first. But now I did everything again and after reboot there was no cuda installed:

Code:

m1@m1-desktop:~$nvOC stop
Stopping nvOC processes...

gnome-screensav killed (pid 1842)

...done.
m1@m1-desktop:~/NVOC/testing$ nvcc --version
The program 'nvcc' is currently not installed. You can install it by typing:
sudo apt install nvidia-cuda-toolkit
m1@m1-desktop:~/NVOC/testing$ systemctl status nvidia-persistenced
● nvidia-persistenced.service - NVIDIA Persistence Daemon
   Loaded: loaded (/lib/systemd/system/nvidia-persistenced.service; static; vendor preset: enabled)
   Active: active (running) since Tue 2018-11-06 08:03:11 MST; 1min 10s ago
  Process: 820 ExecStart=/usr/bin/nvidia-persistenced --user nvidia-persistenced --no-persistence-mode --verbose (code=exited, status=0/SUCCESS)
 Main PID: 879 (nvidia-persiste)
   CGroup: /system.slice/nvidia-persistenced.service
           └─879 /usr/bin/nvidia-persistenced --user nvidia-persistenced --no-persistence-mode --verbose

Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:01:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:02:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:03:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:04:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:05:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:06:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:08:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: device 0000:09:00.0 - registered
Nov 06 08:03:10 m1-desktop nvidia-persistenced[879]: Local RPC services initialized
Nov 06 08:03:11 m1-desktop systemd[1]: Started NVIDIA Persistence Daemon.
m1@m1-desktop:~/NVOC/testing$ sudo systemctl enable nvidia-persistenced
m1@m1-desktop:~/NVOC/testing$ nvcc --version
The program 'nvcc' is currently not installed. You can install it by typing:
sudo apt install nvidia-cuda-toolkit
m1@m1-desktop:~/NVOC/testing$ export PATH=/usr/local/cuda-9.2/bin${PATH:+:${PATH}}
m1@m1-desktop:~/NVOC/testing$ export LD_LIBRARY_PATH=/usr/local/cuda-9.2/lib64\
> ${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
m1@m1-desktop:~/NVOC/testing$ nvcc --version
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Tue_Jun_12_23:07:04_CDT_2018
Cuda compilation tools, release 9.2, V9.2.148
m1@m1-desktop:~/NVOC/testing$

how do I make setting paths and persistanced permanently the right way please?

edit2: did nvOC upgrade + miner recomp, started nice_x16r and get this:

Code:

    SCREEN -c /home/m1/NVOC/testing/screenrc-miner -dmSL miner /home/m1/NVOC/testing/miners/ZENEMYminer/ccminer -a x16r -o stratum+tcp://x16r.eu.nicehash.com::3366 -u . -p x -i 0 SCREEN -c /home/m1/NVOC/testing/screenrc-miner -dmSL miner /home/m1/NVOC/testing/miners/ZENEMYminer/ccminer -a x16r -o stratum+tcp://x16r.eu.nicehash.com::3366 -u . -p x -i 0
/home/m1/NVOC/testing/miners/ZENEMYminer/ccminer: error while loading shared libraries: libcudart.so.8.0: cannot open shared object file: No such file or directory
/home/m1/NVOC/testing/miners/ZENEMYminer/ccminer: error while loading shared libraries: libcudart.so.8.0: cannot open shared object file: No such file or directory

I thought I need cuda9.2 for zenemy/x16r!?

edit3: seems like nOC didn't update properly, I am still on an outdated 1bash. tried again:

Code:

m1@m1-desktop:~$nvOC upgrade
Warnings:
- your customized 1bash settings won't be affected, but you may have
  to create a new one starting from a frash copy of 1bash.template
  if it gets updated during this upgrade, otherwise you will get a
  warning each time nvOC restarts
- the upgrade will fail safely if you made conflicting customizations
  to any other nvOC script: if it happens you need either to run this
  script with restore command or manually manage this by git commands.

Do you want to continue (y/N)?  y
Checking for nvOC updates...  (target: '/home/m1/NVOC/testing')
Updating 063cd21..7779bce
error: Your local changes to the following files would be overwritten by merge:
        0miner
Please, commit your changes or stash them before you can merge.
Aborting
nvOC has been upgraded.
nvOC: line 894: /home/m1/NVOC/testing/1bash.template: No such file or directory
nvOC: line 899: ((: 0038: value too great for base (error token is "0038")
nvOC: line 907: ((: 0038: value too great for base (error token is "0038")
There may be new required miners packages to be installed.
  Would you like to check for further miners updates and install them now?
Do you want to continue (Y/n)?  n
m1@m1-desktop:~/NVOC/testing$

erm, what?

Code:

m1@m1-desktop:~$nvOC restore
Checking for modified contents...
 M 0miner
 D 1bash.template
 M telegram
Stashing nvOC tree modifications

*** Please tell me who you are.

Run

  git config --global user.email "you@example.com"
  git config --global user.name "Your Name"

to set your account's default identity.
Omit --global to set the identity only in this repository.

fatal: unable to auto-detect email address (got 'm1@m1-desktop.(none)')
Cannot save the current index state
Done. Your modified contents has been restored. You can now upgrade.
If you just made a mistake you can still recover your mods, see '\e[1mgit stash --help\e[0m'.
m1@m1-desktop:~/NVOC/testing$

Do I need a github account for this?

[Stubo]

Oh, geez, fk2. I wouldn't know where to start with that. While I understand nvOC better than most, if I found myself in that situation, I would start over with a fresh image. I say that because the most time consuming part for me is in getting the OC settings right for each GPU and algo/miner. I presume you have those saved off in 1bash copies (I use a spreadsheet and 1bash copies) so it may be faster to just start over.

Just my 2 cents.

[fk2]

Hi Stubo, I totally agree with you but unfortunately I colocated my miners and reimaging the hdd/ssd means extra costs by travelling or hireing the colocation staff.

I tried to git clone 3.0 stable to 'NVOC/mining' to my 16.04 2.1 testing rig but that did not work out like expected. followed github readmes. Is this even possible to clone 3.0 to a 2.1 testing rig with 16.04 base? I must confess I didn't tried to recompile all miners yesterday it was a long day...

when I do clone 3.0 to my rig it doesn't even give a master-log:

Code:

m1@m1-desktop:~$nvOC master-log
tail: cannot open '/home/m1/NVOC/mining/nvoc_logs/screenlog.0' for reading: No such file or directory
tail: cannot open '/home/m1/NVOC/mining/nvoc_logs/tempcontrol-screenlog.0' for reading: No such file or directory
tail: cannot open '/home/m1/NVOC/mining/nvoc_logs/watchdog-screenlog.0' for reading: No such file or directory
tail: cannot open '/home/m1/NVOC/mining/nvoc_logs/8_wtmautoswitchlog' for reading: No such file or directory
tail: no files remaining
m1@m1-desktop:~/NVOC/mining$

besides: I tried on a second miner yesterday and I did successfully update to cuda 9 2 on a 2.1 testing nvOC with 16.04 and preinstalled cuda8. I only need to manage to get this cuda92 paths set correct permanently.

I will try and figure out and when I am done I will post a specific short question because I just realise I bomb way too much info and problems here. Very important first:

how do I make setting cuda92 paths and persistanced permanently after installation the right way please?

edit: back to the problematic first rig. I copied 0miner and 1bash from the cloned 3.0 mining folder into my old 2.1 testing folder and it worked. Here is report. I guess it needs some updates to realise there is a cuda92 version of zenemy but failed as described in the post above

Code:

Operating System info:
Kernel        :  4.4.0-97-generic
OS            :  Ubuntu 16.04.3 LTS
System        :  (gcc 5.4.0 (Ubuntu 5.4.0-6ubuntu1~16.04.4)
nvidia driver :  410.73

nvOC info:
nvOC path     :  /home/m1/NVOC/testing
nvOC          :  0019-3.0
Report        :  0012
1bash         :  0045
1bash.template:
3main         :  0026
5watchdog     :  0022
6tempcontrol  :  0008
wtm switch    :  0023
wtm profit    :  0004
pool switch   :  0002
minerinfo     :  0010

Hardware info:
Motherboard   :  ASUS PRIME Z270-A  Rev 1.xx
BIOS ver.     :  1009 07/23/2017
CPU Model     :  Celeron(R) CPU G3900 @ 2.80GHz
CPU Cores     :  2 (Cores + Threads)
Mem Total     :         3984604 kB
Mem Free      :          127312 kB
Swap Total    :               0 kB
Swap Free     :               0 kB
Ethernet      :  Intel I219-V

CUDA SDK version
nvcc path     : /usr/local/cuda-9.2/bin/nvcc
nvcc: NVIDIA (R) Cuda compiler driver
Copyright (c) 2005-2018 NVIDIA Corporation
Built on Tue_Jun_12_23:07:04_CDT_2018
Cuda compilation tools, release 9.2, V9.2.148

nvOC is not installed as a service

HDD and Partion info:
NAME    SIZE FSTYPE TYPE ROTA HOTPLUG TRAN   VENDOR   MODEL             REV MOUNTPOINT
sda    74.5G        disk    1       0 sata   ATA      TOSHIBA MK8052GS 0A
├─sda1    9M vfat   part    1       0
└─sda2 74.5G ext4   part    1       0                                       /

VGA info:
01:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
02:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
03:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
04:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
05:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
06:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
08:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)
09:00.0 VGA compatible controller: NVIDIA Corporation Device 1c03 (rev a1)

Local git branch and submodules status:
## 19-2.1...origin/19-2.1 [behind 1160]
 M 0miner
 D 1bash.template
 M telegram
+c9aea3e750138de4b9eae7731f89b84ba2a418ac miners (remotes/origin/19-2.1)
## HEAD (no branch)
 cd6fab68823e247bb84dd1fa0448d5f75ec4917d ANXccminer/src (heads/master)
 baf6c9e4e36c9cc1b67698ee2977d445f598c573 ASccminer/src (v1.0)
-24c65cf166bbb3332d60e2baef859ceb604e5d49 ETHMINER/src
+2e457923f3125fbaedf5d8ba1f7d0fafc85b0ba8 KTccminer/src (1.4.6-tpruvot-1389-g2e45792)
+7061f3c78e52a03f7ff5d0743900312de5bb24fc KTccminer_cryptonight/src (3.05)
 7d41d49b92db27b9ab80270adaa92f6b06d1ef78 KXccminer/src (0.1)
 78dad7dd659eae72a07d2448de62b1946c1f2b41 MSFTccminer/src (2.2.5-rvn)
 8affcb9cd09edd917d33c1ed450f23400f571bdb NAccminer/src (v2.2-mod-r2)
 9e86bdd24ed7911b698f1d0ef61a4028fcbd13c5 SPccminer/src (1.5.81)
 c800f1a803e1b2074ed2a7c15023c096d0772048 SUPRminer/src (1.5)
+370684f7435d1256cbabef4410a57ed5bc705fdc TPccminer/src (2.3-tpruvot)
 48b170a5828256600ca71e66d4c114af4e114236 VERTMINER/src (v1.0-beta.2)
+c5f0505de039545585811585f2c189828dfc3ec2 XMR_Stak/src (2.4.7)
 bfd1c002f98f2d63f2174618838afc28cf4ffffe cpuOPT/src (v3.8.8.1)

1bash settings:
1bash version .............: 0045
LOCAL or REMOTE ...........: REMOTE
TEAMVIEWER started ........: NO
SSH daemon started ........: YES
SLOW_USB_KEY_MODE .........: NO
SRR .......................: NO

Watchdog ..................: YES
SYSRQ Reboot ..............: NO
Watchdog Cycle ............: 15
GPU Utilization Threshold .: 50
Alternate Pool ............: NO

Temp Control ..............: YES
MANUAL_FAN ................: NO
TARGET_TEMP ...............: 75
__FAN_ADJUST ..............: 5
POWER_ADJUST ..............: 5
ALLOWED_TEMP_DIFF .........: 2
RESTORE_POWER_LIMIT .......: 85
MINIMAL_FAN_SPEED .........: 50
MAXIMAL_FAN_SPEED .........: 95

CLEAR_LOGS_ON_BOOT ........: NO
AUTO_REBOOT ...............: NO
_Parallax_MODE (upPaste)...: NO

TELEGRAM_MESSAGES .: NO
TELEGRAM_ALERTS ...: NO

HEADLESS MODE:.............: NO
GPUPowerMizerMode_Adjust...: NO
ETH Pill ..................: NO

POWERLIMIT MODE ...........: GPU_SPECIFIC
GPU0 Power Limit ..........: 76
GPU1 Power Limit ..........: 76
GPU2 Power Limit ..........: 76
GPU3 Power Limit ..........: 76
GPU4 Power Limit ..........: 76
GPU5 Power Limit ..........: 76
GPU6 Power Limit ..........: 76
GPU7 Power Limit ..........: 76
OVERCLOCK MODE ............: GPU_SPECIFIC
GPU0 Core Overclock ........: 110
GPU0 Memory Overclock ......: 1100
GPU1 Core Overclock ........: 110
GPU1 Memory Overclock ......: 1100
GPU2 Core Overclock ........: 110
GPU2 Memory Overclock ......: 1100
GPU3 Core Overclock ........: 110
GPU3 Memory Overclock ......: 1100
GPU4 Core Overclock ........: 110
GPU4 Memory Overclock ......: 1100
GPU5 Core Overclock ........: 110
GPU5 Memory Overclock ......: 1100
GPU6 Core Overclock ........: 110
GPU6 Memory Overclock ......: 1100
GPU7 Core Overclock ........: 110
GPU7 Memory Overclock ......: 1100

GLOBAL WORKERNAME .........: YES
AUTO WORKERNAME ...........: CUSTOM
WORKERNAME ................: Luisenmine01
plusCPU ...................: NO

AUTO_SWITCH ...............: NO
COIN ......................: NICE_X16R
ALGO ......................: X16R
MINER .....................: ZENEMYminer
NICE_X16R WORKER ................: Luisenmine01
NICE_X16R ADDRESS ...............: 3CW6n5M7...
NICE_X16R POOL ..................: x16r.eu.nicehash.com
NICE_X16R PORT ..................: 3366
X16R INTENSITY ..........: 0

[Stubo]

I don't think this will be extremely helpful because it doesn't answer your question directly, but maybe it will provide some clues. I am running the 3.0-stable release which is Ubuntu 18.04.1 and CUDA V9.2.148 by default. The interesting part is that it also has CUDA 8 installed. So, I started looking into how it supports CUDA 8 by toggling back to it. I started by grepping for cuda in the mining directory. There were two hits of interest - 0miner and screenrc-miner-cuda-8. So in screenrc-miner-cuda-8, we have these env var changes to toggle back to 8:

Code:

## set cuda-8.0 env for older miners
setenv PATH /usr/local/cuda-8.0:$PATH
setenv LD_LIBRARY_PATH /usr/local/cuda-8.0/lib64:$LD_LIBRARY_PATH

And this screenrc is called by 0miner with this code:

Code:

# List of cuda-8.0 miners
CUDA_8_MINERS="ANXccminer ASccminer KXccminer MSFTccminer NAccminer SILENTminer SPccminer SUPRminer VERTMINER"
LAUNCH="screen -c ${NVOC}/screenrc-miner -dmSL miner"
for miner in $CUDA_8_MINERS
do
  if [[ ${!xminer} =~ $miner ]]; then
    LAUNCH="screen -c ${NVOC}/screenrc-miner-cuda-8 -dmSL miner"
  fi
done

So, if the coin you chose has a miner that requires CUDA 8, then it toggles $LAUNCH to use a different screenrc which in turn handles the necessary env var tweaks.

Hope this helps.

[LuKePicci]

Ok, step by step.

Both nvidia drivers and CUDA SDKs (with runtimes) can be installed in two different ways, distro packages or "runfiles".
In the first case you need to uninstall with apt, in the second one you need to uninstall with uninstall perl script.
In nvOC, as in original fullzero's design the cuda (8.0) sdk was installed by runfile while nvidia driver was installed by distro package. nVidia explicitly recommends to not mix runfile with package installs of drivers/cuda. For this reason, every other cuda sdk installed in newer nvoc images is installed by distro package as like as drivers. The old cuda 8 sdk installed by runfile is still sitting there as long as it does not cause problems. In old images your .bashrc may still contain explicit addition of cuda-8.0 paths to $PATH therefore even if you install newer sdks doing nvcc --version will still show the old version number. This is not necessarily an issue as long as 0miner targets each miner to the correct runtime. It is instead an issue with miners compile where compiler script do not make explicit reference to a particular sdk version.

So, just check .bashrc and replace paths pointing to "cuda-8.0" with paths pointing to the generic "cuda" wich is symlinked to cuda sdk you installed last. Of course, if you messed up with sdks you don't need to uninstall and reinstalll an sdk to make this be symlinked correctly, just fix that symlink manually.

Generally speaking, for old images (newer images has some changes included that avoids this kind of issues): you don't need to uninstall the cuda-8 sdk, you also don't need to fix path in bashrc because 0miner takes care of selecting the runtime for the miner you re using, but you may need to make sure you are explicitly targeting (by make command line options) the correct sdk while compiling if you do not intend to compile with cuda-8

That's all about cuda, now let's talk about upgrades issues.

That message asking you to perform a restore appears if git detects you changed something in nvoc files that upgrades would overwrite. nvOC restore gives you a quick shortcut to clean-up all main nvoc tracked files (ie excluding 1bash, miners submodule and other temp files, or additional files you created in the nvoc path. To perform the clean-up, the nvoc restore uses the git stash command, allowing you to recover any customization you unintentionally lost by performing the upgrade. The git repo manager requires you to identify yourself before making commits or stashes, that''s why it is asking you to say who you are. Nothing related togithub, which is just  git repo host. You can simply input "me@nvoc" as email and "Proud Miner" as name. Newer images has such kind of settings included by default.

[rsafiyari]

Hi all. I really need help with a rig i just picked up. I'm new to mining and some this stuff is way over my head and will take time to learn. My issue now is i cannot turn on my rig (I know). it's a 13 gpu 1080 cards on a 13 slot moderboard with 4gb ram. it's on Nvoc (linux).

Problem is my monitor shows a purple screen and it doesn't go into nvoc. I turn on the power supplies then the moderboard. Then all GPU lights come on and fans start to work. My monitor comes on and goes into a purple screen. Then fans stop and nothing happens. Lights on GPU stay on and monitor stays on purple screen. I tried it twice waiting an hour before calling it a day.

I was told it takes a while to start because of the # of GPUs connected to the moderboard and only 4gb ram. But don't think it should take hours??? Any help would be appreciated!!!

[Tigel]

Hi all. I really need help with a rig i just picked up. I'm new to mining and some this stuff is way over my head and will take time to learn. My issue now is i cannot turn on my rig (I know). it's a 13 gpu 1080 cards on a 13 slot moderboard with 4gb ram. it's on Nvoc (linux).

Problem is my monitor shows a purple screen and it doesn't go into nvoc. I turn on the power supplies then the moderboard. Then all GPU lights come on and fans start to work. My monitor comes on and goes into a purple screen. Then fans stop and nothing happens. Lights on GPU stay on and monitor stays on purple screen. I tried it twice waiting an hour before calling it a day.

I was told it takes a while to start because of the # of GPUs connected to the moderboard and only 4gb ram. But don't think it should take hours??? Any help would be appreciated!!!

Start with one GPU plugged to PCIe x16 slot. Once you have it up & running start adding additional GPUs.

[papampi_2]

Hi all. I really need help with a rig i just picked up. I'm new to mining and some this stuff is way over my head and will take time to learn. My issue now is i cannot turn on my rig (I know). it's a 13 gpu 1080 cards on a 13 slot moderboard with 4gb ram. it's on Nvoc (linux).

Problem is my monitor shows a purple screen and it doesn't go into nvoc. I turn on the power supplies then the moderboard. Then all GPU lights come on and fans start to work. My monitor comes on and goes into a purple screen. Then fans stop and nothing happens. Lights on GPU stay on and monitor stays on purple screen. I tried it twice waiting an hour before calling it a day.

I was told it takes a while to start because of the # of GPUs connected to the moderboard and only 4gb ram. But don't think it should take hours??? Any help would be appreciated!!!

Make sure monitor is connected to the x16 pcie GPU and IGPU is disabled in bios.
4Gb ram is enough and should makes no problems, it should boot in less than a minute.
Whats the nvOC version?
It looks like a xorg problem, if you have ssh access try to restore it.

[fk2]

@Luke @Stubo

tyvm guys for your help and infos, I managed to get everything working properly today!

[parentibule]

Hi all,

I have still and always the same problem: mining stop to work, gpuinfo doesn't respond and restart doesn't work (no restart).
If I try a manual nvOC restart:

(sometimes Failed to get unit file satet for nvoc.service: Connection timed out)
Stopping nvOC processes...
...done.
Restarting nvOC tasks....

And that all, nothing happens.

ethminer process is running, take 100% CPU, in zombie state and strictly impossible to stop (never saw that on linux).

Any ideas?

[papampi_2]

Hi all,

I have still and always the same problem: mining stop to work, gpuinfo doesn't respond and restart doesn't work (no restart).
If I try a manual nvOC restart:

(sometimes Failed to get unit file satet for nvoc.service: Connection timed out)
Stopping nvOC processes...
...done.
Restarting nvOC tasks....

And that all, nothing happens.

ethminer process is running, take 100% CPU, in zombie state and strictly impossible to stop (never saw that on linux).

Any ideas?

I have seen sometimes ethminer hangs when trying to kill it,
No idea why and the only way to kill it is by using kill -9
May be post in ethminer GitHub page, see if they have some info why it happens

For now try this when it happens

Code:

ps ax | grep -v cpuminer | grep SCREEN | grep miner | awk '"miner" {print $1}' | xargs kill -9

[parentibule]

Thanks, but already made a ton of tries.

Why ethminer cannot be build? I try to lauch it with --cuda but it doesn't work (always 0H/s)

[papampi_2]

Thanks, but already made a ton of tries.

Why ethminer cannot be build? I try to lauch it with --cuda but it doesn't work (always 0H/s)

Better ask that in ethminer support page too.

[noct01!]

Hello!

Still loving the work with nvOC, the whole farm runs 19-2.1 and works great.

Upgraded my test rig to -3.0 tonight and seem to have stumbled across the 6tempcontrol issue where the GPU fans sit at 30% and don't change even though the cards are above the target temp.

The Github has the issue logged as a bug but nothing has been said for 3 weeks. Is there a fix? I've upgraded to the latest nvidia-smi and -settings but still no success.

Any thoughts or rollback?

Thanks!