Topic: [OS] nvOC easy-to-use Linux Nvidia Mining - page 388. (Read 418253 times)

Hi @fullzero,

just a suggestion - could you add the ability to mine UBQ? I'm pretty sure claymore can do it if you set allcoins to -1.

I'm saying this because it has recently jumped to the top of the whattomine.com list for dagger-hashimoto coins a few times.

It's only EWBF, ccminer and Claymores do not have this drop in hashing power.

Hi @fullzero,

I have been watching this space and trying nvOC because I would so much rather use Linux if possible. I've probably already said this but my experience with EthOS on AMD rigs has been great - awesome stability, remote monitoring, better ETH hash-rate than Claymore with no fee (by using sgminer-gm).... So I REALLY want nvOC to work at least as well on my Windows 8.1 rig, as Windows does.

Unfortunately I just can't seem to get the overclocks to stick! I'm mining ETH so in Windows using Afterburner I lower the core clock or leave it at factory and I overclock the memory to +750 on all 4 of my Nvidia cards (2 x 1070, 2 x 1060 6GB). This gives me a reported hash-rate in Claymore of around 107mh/s and uses around 600W for the whole rig.

When I use nvOC a couple of strange things happen (or maybe they aren't strange - you tell me?). When the terminal auto-starts I have to put in the miner1 password 3 times  before Claymore starts. For whatever reason, I always see the message "GPU#1 disabled". If I hit the "1" key it says "GPU#1 enabled and works. Also it usually seems to take 2 or 3 attempts before it successfully connects me to nanopool, i.e. it can't connect the first 1 or 2 times and then it does connect. In Windows it always connects 1st try. And finally - it's reported hash-rate mines at around 99mh/s which is the hash-rate I get when the GPUs are at factory settings. Another strange thing about this is that it appears to be using almost the same amount of power - 580W, but seemingly without the overclocking.

When the terminal runs I see it apply the numbers - 0 to core and 750 to memory, but the GPUs seem to be ignoring this for some reason. They are different branded GPUs - the 1070s are Gigabyte and the 1060s are Galax and they all seem to respond to Afterburner in Windows.

Here is a paste of the relevant parts of my onebash:


I will next try it with individual clocks set to "YES" but I did try that with version 0014 of nvOC and it didn't work. it doesn't need a "+" symbol does it it?  Like "MEMORY_OVERCLOCK_2=+750"?

Any help from OP or anyone else would be appreciated. I know it's only 8mh/s difference but I really want to be hashing as high as possible because I live in a country with VERY expensive power so highest hash per watt is important to me.

I can not get the individual power setting to work. How do I manually override the onebatch setting?

That made it work! Thanks so much!

Glad you got it 

I recommend everyone learn more Linux and Networking; I try to on a regular basis.

Excellent -- I have now the bashing action.

4 x Aorus 1080ti
165/0/195 - mining ZEC

I got over 720-740sols per card with a high PL 265
But since its mining ZEC, zero for memory and +165 for core
and lower down the PL to 195

Temp: GPU0: 68C GPU1: 69C GPU2: 64C GPU3: 63C
GPU0: 672 Sol/s GPU1: 645 Sol/s GPU2: 651 Sol/s GPU3: 666 Sol/s
Total speed: 2634 Sol/s
+-----+-------------+--------------+
| GPU | Power usage |  Effecincy   |
+-----+-------------+--------------+
|  0  |    202W     |  3.33 Sol/W  |
|  1  |    197W     |  3.27 Sol/W  |
|  2  |    197W     |  3.30 Sol/W  |
|  3  |    194W     |  3.43 Sol/W  |
+-----+-------------+--------------+

To setup remote re-configuration of rigs / remote viewing using SSH and screen:


PREP RIGS:

On each rig you wish to ssh into ensure:



in 1bash

have your phone or a pencil and paper with you as you will also need the ip address of each rig:

to find the ip:

scroll to the top of the mining process where is shows:


find this ip and record it with a picture or write it down / save it to a text file ect.

When you have finished this with each rig; go to the computer you are going to use to ssh from:


SETUP CLIENT:

If it is a windows computer I recommend using putty: http://www.putty.org/ for this as it will let you use as many instances into as many rigs as you want.

for each rig open a new putty instance then:

enter:


replacing ipaddress with the miners ip from eariler; if it was 192.168.1.124, then you would enter:


and click open





confirm your intent to access the system if asked, then enter password:

miner1

when prompted

Then move on to the Reconfigure section below:


If you are on a linux computer:

open a terminal and enter:


when it is done:

for each rig open a new terminal instance or tab then:

enter:


replacing ipaddress with the miners ip from eariler; if it was 192.168.1.124, then you would enter:


confirm your intent to access the system if asked, then enter password:

miner1

when prompted

Then move on to the Reconfigure section below:



RECONFIGURE:


after SSHing into a rig:

open 1bash with nano:


make changes / edits; then press ctrl + shift + X to quit selecting Y to save changes

Then:

we need to kill the current mining process by entering:


now we have stopped the GPU mining process



SCREEN:

after killing the current mining process:

enter this to prevent a possible error:


Then attach to the current mining process screen (v0018 will auto restart the mining process):



if you are updating for _Parallax_ Mode enter:


then


and you have remote viewing of the mining process in this screen.

you can close the ssh session (window / terminal) whenever desired

then later reSSH in and enter:


to return to the mining process whenever desired.


If you are at this point and have a farm command center PC:  Congratulations!  


Right now the simplest way to setup Internet access to your rigs is this:

Setup a single computer with SSH sessions into each rig (I recommend saving each session config or making a file with the settings to launch each.)  

On this computer install Teamviewer or a similar application.  

Connect via your mobile / tablet / work computer ect to the control computer with Teamviewer or similar application, then on that rig Monitor / reconfigure each rig via its SSH session.

This should be reasonably simple (given the complexity of what you are doing) to setup.


See the _Parallax_ Mode guide for autoupdating oneBash Link


Additional useful cmds for Screen from UberDaemon:

To detach from a screen (i.e. you want to close the SSH connection while mining continues afterwards) it's


To reattach to a screen (resume connection) it's

 # once you've typed 3-4 characters hit the tab key and it will fill out the rest and save you some typing

If you can't remember what you named the session use:


If you're new to screen you may not know you can open a 2nd shell and switch back and forth (i.e. your miner is running on one shell and you open a 2nd to run nvidia-smi):

once in screen you create a 2nd shell (or 3rd, 4th, etc) with

To switch back and forth use:

That did the trick.  Thanks!

I will make a single post with a list of instructions; but:

after killing the current mining process:


or whatever you want to call this session in place of rig1

then enter:


and you have remote viewing of the mining process.   

I'd like to get this working so I can monitor via SSH, but when I try re-initiating gnome-terminal I get:

Executing the 'gnome-terminal' command with sudo has the same result.

Thanks, I'll try it and let you know.

I would try re-imaging the USB key.  

I am reasonably sure forcing the integrated graphics by connecting to the mobo is what caused the problem; I'm not sure what happens with the xorg.conf when that happens.  

If I am correct, then after reimaging the cards should be detected.

Yes, I only see the 4 NVIDIA GPUs. Copied, logged out, logged in, don't see the OC messages (also still get an error from nvidia-settings).

What coin and settings?

Started both: keep getting sidetracked and still have 1 week left of non Crypto obligations.

try this again:

lspci | grep VGA

if it shows only 4x GPUs then do this again:

press f12 to open the guake terminal, if cpuminer is running press ctrl + c to close it or open a new tab and enter:


then:


you should see a file that looks something like: '/etc/X11/xorg.conf.06082017'  but most likely with a different date number at the end maybe 06092017? (it will be from the first day you launched nvOC so it should start with a 06 )

anyway note what the date is on that file and change the following code to match it:


and enter the password: miner1 when prompted.

then logout

and log back in

and tell me if you see the OC messages.

Have any of you guys see your rigs hash and then all of them go to 0 Hash ?  Even reboots, it will do that.  Zotac mini 1070s

How's the monitoring software coming? And the amd build?

Comes back with the same errors.