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

[fullzero]

nvOC 0019 didn't work on my AM4-platform. Has anyone else tested? Using ASUS ROG STRIX B350-F motherboard. I kept getting corrupt xorg.conf errors on boot. This system doesn't have iGPU, so I don't really know what to do..
Has anyone else tried nvOC on AM4?

If you get more than one xorg.conf reboot then; nvOC will probably not work without modification on what mobo.  By this I mean: if the xorg.conf message and reboot is a loop ( happens 2 or more times in a row) then you need to modify something.

[1714192378]

1714192378

[1714192378]

1714192378

[1714192378]

1714192378

[fullzero]

I found out that this relay


work with this code:

RESET RIG

Code:

#!/bin/bash
# RESET RIG
GPIO_RESET=23
gpio -g mode $GPIO_RESET out
sleep 2
# Resetea el equipo
gpio -g write $GPIO_RESET 0 ----> HERE IS THE CHANGE
sleep 1 # Lo mantengo pulsado 1 segundo
gpio -g write $GPIO_RESET 1 ----> HERE IS THE CHANGE
exit

POWER ON/OFF RIG

Code:

#!/bin/bash
GPIO_POWER=24
gpio -g mode $GPIO_POWER out
sleep 2
# Enciende o apaga el equipo segun su estado previo
gpio -g write $GPIO_POWER 0 ----> HERE IS THE CHANGE
sleep 1 # Lo mantengo activado 1 segundo
gpio -g write $GPIO_POWER 1 ----> HERE IS THE CHANGE
exit

And this one wich looks pretty the same


works with this code:
RESET RIG

Code:

#!/bin/bash
GPIO_RESET=23
gpio -g mode $GPIO_RESET out
sleep 2
# Resetea el equipo
gpio -g write $GPIO_RESET 1 ----> HERE IS THE CHANGE
sleep 1 # Lo mantengo pulsado 1 segundo
gpio -g write $GPIO_RESET 0 ----> HERE IS THE CHANGE
exit

POWER ON/OFF RIG

Code:

#!/bin/bash
GPIO_POWER=24
gpio -g mode $GPIO_POWER out
sleep 2
# Enciende o apaga el equipo segun su estado previo
gpio -g write $GPIO_POWER 1 ----> HERE IS THE CHANGE
sleep 1 # Lo mantengo activado 1 segundo
gpio -g write $GPIO_POWER 0 ----> HERE IS THE CHANGE
exit

Hope it helps.

what are you using to detect when it freezes?

1.ping
2.check hashrate from pool, if hashrate is low, it is freezenn.
3.run nvidia-smi via ssh, if it exits with error code, the rig is   unhealthy, reset it.

Hi! I sat up Raspberry Pi and relay, but I can only manually power on/off my rigs. The program from here: https://bitcointalk.org/index.php?topic=1933467 works, but sometimes when miner hangs, rig accepts ping and RPi doesn't reset it. Can you share your program here? And can RPi and nvOC be sat up to work together like SRR?

SRR_SERIAL="000055"
__SRR_SLOT="1"

Sorry I wasn't clear enough, your code is exactly what I mean to reset "manually", I'm using it. But we need code to work automatically and together with nvOC and it's scripts, like SRR

Ok, I got it, I think.
The thing is the raspi must accept commands to reset/power on/off from Nvoc. If that is right then:
1. A variable is needed to hold the raspi's ip (in 1bash?)
2. Guess a variable is needed to  choose raspi or srr (in 1bash?)
3. Send command from nvoc to fire up the reset script in raspi (same conditions as in srr)

Example to send commands from a local linux box to a remote one (executing in the remote one):
ssh -o LogLevel=quiet -p $PORT root@$REMOTE_IP 'if [[ -f /tmp/file.csv ]]; then /bin/rm --interactive=never /tmp/file.csv \
                           && echo "OK: /tmp/file.csv has been deleted in remote system"; else echo "OK: /tmp/files.csv missing in remote system";fi' | tee -a $LOG_FILE

This is a very good option but a certificate id needed, so I don't think is the best solution from the user is point of view becouse it needs to install the certificate manually in the raspi.

Other option is use sshpass: https://stackoverflow.com/questions/12202587/automatically-enter-ssh-password-with-script

I don't use any of the mechanisms that Nvoc has for restart the system etc. Only use Maxximus007_AUTO_TEMPERATURE_CONTROL

we can add the RPi's ssh public key to every rigs, so the RPi can exec command against rigs via ssh. nvOC doesn't need to know RPi. RPi check hashrate from pool, if it finds a rig with low hashrate on pool, reset the rig.

I have runable codes in github, but the code quality is low, only supports few pools, and there is no setup descriptions, I'm going to improve it and maybe release it next month:)

Great. It's a lot more direct aproach. If the rig is not in danger of "burning" I would try first a reboot. Not system likes a hard reset!!.

I will get an rpi and relay and integrate this eventually; I like open source solutions that use over the counter parts. 

[fullzero]

Any one knows a simple embedded text editor for web sites?
I want to add 1bash editor to nvOC web page, so we can edit it from its web page
Going to add a restart miner too if can find a simple editor.

This is how it looks like now :

looks good. 

I will add to -1.3

[fullzero]

Any one has any news from fullzero ?
His last post was Oct. 1st a week ago ...
And last active Oct 3rd.
Getting worried for him.

someone ping him!

I've been in hardware mode this past week;  I'll try to push -1.3 soon to add all the new contributions / zm 5.0.

[fullzero]

Anybody can help me out..

I have 5 12GPU rigs ,

1 of them  i keep frying mobo's .. i have tried everything..

Tested all parts sperately all seems to work
however  the minute i attach my PSU's to the motherwboard whilst I have GPU's and risers powered , even without plugging the PCI's in the MOBO.

It instantly fries my mobo...

I have fried 4 motherboards already trying to trial and error...


it used to work before though...

Why would you attach PSU's while the system is powered?
Maybe I'm not understanding something. Can you give more info?

You missunderstand or i explained wrong  :d

Either either way..

I tested all parts separate by powering the parts with psu. All fans are running etc .

However if I prepare my system.
Combine all parts like they should.
The moment I power on. It will fry the mobo. The cpu fan will spin briefly and it is dead from this moment.

I have fried 4 mobos already cause i don't know how I can find problem without trying to connect it.

Is it possible a gpu is dead or something? It used to work before and suddenly it died after I had some gpu losses..

If it is this gpu or a extender how do I find it because If i have all gpu powered with psu no mobo and no display. Just power to see fans running all have working fans.

How is it possible that broken gpu fry my motherboards? Or is it something I'm missing

how are you switching on your other/s power supply ?
Using this ? https://www.amazon.com/Supply-Adapter-Connector-Triple-Multiple/dp/B073PRM7YZ
Manually ?
Something else ?

Yes I use those.
However I have 4 systems still running like that.
Today I tested with a mobo with 6gpu card only.
It still fried, I also did not use the adapter for this test and it still fried my mobo.

I also tested


Also
I even fried a mobo while only the following where connected.

Connected on psu: hdd, pci extenders, 6pin gpu powers, mobo main plug +cpu.

Connected on mobo cpu + main mobo power.

The extenders were not plugged in.
This gives me reason to think it is not gpu?

Is it possible something gives short circuit? From psu?

it can be a lot of things but I would replace that psu. What brand and model is it?

on this particular system it are 2 different psu , on the other's it's 3 PSU all the same

The malfunction system has a 1200 platinum seasonic psu and a corsair platinum 850 w psu

The othersystems all have 3 x 850w corsair platinum psu hxi or something

how do i sample test this psu to see for error?
can i use voltage meter?

If you fried so many mobos I wouldn't try connecting those power supplies anymore. Request RMA.

You can give it one more try before replacing the mobo and power supplies for that rig

What you should do is power the motherboard without any risers and GPU's first. Use only one power supply, plug the 24-pin ATX cable, 8-pin CPU cable and two 4-pin cables to the motherboard and power it on.

If it works, power down your PS, wait 30 seconds the plug one GPU directly to the pci-x16 slot (no riser), don't forget to plug PCI power to the GPU, then power on the rig.

If it works, turn off PS, wait 30 seconds, unplug the GPU and connect it trough riser to the motherboard on the x16 slot (plug  power to the riser) then power on the rig.

If it works, power off the PS, add another GPU trough a riser, power on... and so on, keep adding one GPU at a time.

Here is manufacturer's web site explaining how to connect:
http://www.biostar.com.tw/app/en/event/crypto_mining/page4_2.htm

I had a BIOSTAR TB250-BTC PRO (12x) die on me.  Worked well for ~3 weeks then a cap near the northbridge went.  I wonder if anyone else has had a similar problem? 

I would try leenoox's troubleshooting suggestions; and RMA if needed.

[fullzero]

My 19x rig:

13x 1080ti
6x p106
4gb ram
celeron g3930
2x 2400w server psus from Parallel Miner
1x 1600w evga p2 (Could do without if I really wanted to.)
12x 120mm case fans that just so happened to fit snugly in the case
Custom frame from eBay: http://www.ebay.com/itm/Mining-Rig-frame-19-GPU-Ethereum-Bitcoin-ASUS-B250-Mining-Expert/322780828479

Boring video with no commentary: https://www.youtube.com/watch?v=9_mv1pd3BIA

Nice job; you are the first member to show me 19x working.

Have you tried using 19x of the same GPU?

[fullzero]

OK guys
Here is My WhatToMine Auto Switch forked from damNmad smartminer
It will check the coins you want to mine then based on your cards and your own url mine the top coin if difficulty low and profit is high

These are the files:
WTM_AUTO_SWITCH
WTM_AUTO_SWITCH.py
Put both in /home/m1/

Edit /home/m1/1bash and add :

Code:

# WTM AUTO SWITCH SETTINGS# remember to disable Parallax MODE (_Parallax_MODE="NO")
 
WTM_AUTO_SWITCH="YES"
WTM_AUTO_SWITCH_SYNC_INTERVAL="3" # Time to sync with WTM for best coin
#WTM_AUTO_SWITCH_URL="FOR NOW EDIT IT IN WTM_AUTO_SWITCH.py
#WTM_AUTO_SWITCH_COINS=" FOR NOW EDIT IT IN WTM_AUTO_SWITCH.py > includeTags = [ 'ZEC', 'ZEN', 'ZCL', 'SIB' , 'LBC'  ]

#WTM_AUTO_SWITCH_diff="TO BE ADDED IN NEXT VERSIONS"  # PERCENTAGE TO CHANGE IF TOP COIN IS HIGHER THAN CURRENT COIN

Edit /home/m1/3main and add these lines somewhere  after Maxximus007_AUTO_TEMPERATURE_CONTROL ( easier to find "SALFTER_NICEHASH_PROFIT_SWITCHING" and add before it:

Code:

if [ $WTM_AUTO_SWITCH == "YES" ]
then
HCD='/home/m1/WTM_AUTO_SWITCH'
running=$(ps -ef | awk '$NF~"WTM_AUTO_SWITCH" {print $2}')
if [ "$running" == "" ]
then
guake -n $HCD -r WTM_AUTO_SWITCH -e "bash /home/m1/WTM_AUTO_SWITCH"
running=""
fi
fi

Install requests python module with :

Code:

sudo apt install  python-requests

Go to whattomine select your cards, hash rate, power.
You can also select to mine base on current, 24 hour, 3 day or a week profit and difficulty.
Dont forget to choose same for both profit and difficulty or it will give wrong results.
Click calculate, then add .json to coins at the begining of the address after you click calculate!!!
From:

Code:

https://whattomine.com/coins?utf8=✓&adapt_q_280x=0....

To:

Code:

https://whattomine.com/coins.json?utf8=✓&adapt_q_280x=0&adapt_q_380=0&adapt_q_fury=0&adapt_q

Copy the whole address and paste it to WTM_AUTO_SWITCH.py replace the default address:

Code:

data = requests.get("https://whattomine.com/coins.json");

Set the coins you want to be switched in WTM_AUTO_SWITCH.py in the included tags :

Code:

includeTags = [ 'ZEC', 'ZEN', 'ZCL', 'SIB' , 'LBC'  ]

Now you can start wtm auto switch with

Code:

bash WTM_AUTO_SWITCH &

P.S
Need help to improve the WTM_AUTO_SWITCH.py script, if any one willing to help please let me know.

Very nice! I'll give this a go next week!

I am sure lots of members will use this.  Thanks for sharing with the community.   

[fullzero]

so what if I could kill the current mining process (mining ZEC using 0019 1.2) then install my own miner and run it ?

I found a great tutorial to install and run VTC and would prefer not touching anything until the next beta is release. Is this something possible ?

so I tried this:

pkill -e miner
/home/m1/SPccminer/./ccminer -a Lyra2v2 -i 20 -o stratum+tcp://ipaddress:9171 -u vtcwalletaddress

Sometimes it mines for a shortime, sometimes I get "segmentation fault" error. It may be the autorestart, how to I kill that process ?

Any ideas/tips ?

when you kill miner, wdog will start again
we already have spccminer, why not configure 3main to use it for vtc instead ?

like I wrote earlier, it doesnt work and I dont know why. Is there a log that could help me ? The algo log is always empty when I open it.

I will note to update the SP ccminer fork; is there another VTC client for me to add?

[fullzero]

Hi guys,
I've problems with detecting all installed GPUs. I bought 13x risers V008S with 100cm USB 3 cables. Installed 11x 1080Ti, but system detect only 10.

There are 3 PSU (2x 1700W, 1x 1200W). Each GPU is connected on same PSU as riser. I don't know why I don't see all cards. There is some way, how detect which cars wasn't recognise?
Mobo: ASRock PRO (for 13 gpus)

Unfortunatelly this is not one problem which I have... :-(

Miner is running and sometimes is killed immediately after execution for 5 or more cycles, so it's mean that I don't mine for 20 minutes - if mining MONA.

Safter sometimes has same problem... :-(

Run the miner for a bit then feel each GPU; the one that is cold is the undetected one.

My guess is that MONA often uses less than 90% to reduce this to a MONA friendly level open:

IAmNotAJeep_and_Maxximus007_WATCHDOG

and edit:

Code:

THRESHOLD=90

to:

Code:

THRESHOLD=65

[fullzero]

Started using nvOC yesterday on one of my rigs yesterday and I must say it's great! Able to get 10 GTX 1060s up and running on a single motherboard with a stable O/C (~280 sols per card). I have a question regarding my current effective vs. average effective hashrate on flypool. It seems that my average effective is capped at 7 Kh/s when my current effective hashrate has been at a steady > 8 Kh/s, sometimes > 9 Kh/s. This was not the case prior to switching one of my rigs over to nvOC. What accounts for the discrepancy? I have attached a screenshot that demonstrates what I'm talking about. It has been running for almost 24 hours now.

Depends on the pool, it takes 24 - 48 hours to calculate the average hashrate.

Yes I know, but my hashrate should be increasing steadily, not capped at 7Kh/s. It's been capped at 7Kh/s for the past 8 or so hours as shown by the picture above. Could it have something to do with the EWBF % devfee being set to 0 in nvOC?

I recommend updating to 1-2 and trying zm.

[fullzero]

Hey guys, got two more questions:

1. On windows i was able to turn off the LEDs on my cards using either the terrible MSI Gaming App or the MSI LED tool. Is there some way to turn off the GPUs' lights using nvOC?

2. The telegram notifications do not work at all, i have created a bot, got the API, figured out the Chat-ID and can see the conversations with the bot on a webbrowser. Yet nvOC never sends anything. Guake has the AUTO_GRAM tab and announces "New telegram in 1800 seconds" but has not been sending anything so far.

Thank you for your help!

Hey guys, any ideas for me? Thanks a lot :-)

1:  I don't think MSI makes an app for linux; so you will have to use windows to alter the LEDs

2:  ensure you have set:

Code:

	TELEGRAM_CHATID="your_CHATID_here"

	TELEGRAM_APIKEY="your_APIKEY_here"

in 1bash

they should look something like:

Code:

	TELEGRAM_CHATID="422650663"

	TELEGRAM_APIKEY="417738071:AAE7nwRoaJec4f6zMuK9W5dMmpQ_yAt6YAw"

[fullzero]

Guys

I'm running atm only 4 rigs - each 12 gpu 1070 Gtx

Running zcash with EFWB

3/4 running v1,2
1 running v0,19

all of them have these dips. in hashrate look image

what can I do?.

I tried ZM miner , however sometimes I got this network error. I follow the miner's topic on bttalk and latest version could contain a fix for this. waiting for fullzero to release 1,3.

What can I do other than this? any suggestions?

You can manually update zm to 5.0 by downloading it and replacing the client in the zm directory with the 5.0 version.

[fullzero]

Hi,

I didn't manage to connect on any pool that require password for rig, like suprnova.
Tried sigt (before pos), dmd, dgb, mona : no success.
The screen attach to the miner hang, so i can't see any error log. I've seen similar issues in this thread, but no solution.
Rig is : asrock h110 + 2x970 gtx, nvOs on USB key.

Tried the update v0019-1.2 : same problem.

Everything is fine on pools with no password (hush, zen @ minepro, ...)

Any advice on this point ?

suprnova requires you to make each worker before using it.  If you try to use a worker that isn't in your pool setup it will not connect to the pool.  Make a worker at suprnova (use x for its password) then try again.

[fullzero]

To all those bash experts...
Need an script to give me what miner is running so I can make the Miner Output tailored for each miner

I have this code that I am using in my beta telegram script I'll post some day to this forum:

Minig address:
MINING_ADDRESS=$(ps aux | grep -v grep | grep " miner " | sed 's/ /\n/g' | sed '/\home\/m1/d' | awk ' { if ( length > x ) { x = length; y = $0 } }END{ print y }' | sed 's/\..*//')

Exec:
MINER_NAME=$(ps aux | grep -v grep | grep " miner " | sed 's/ /\n/g' | grep "/home/m1/")

Not sure is good for what you need. Not sure yet if I will get these vars from system or Nvoc vars.
 

Thanks,
but I need some thing to only give miner name,
so if I'm mining zec with zm dstm ,
it gives me zm not home/m1/zec/zm/zm

May be easier that I strip variables from 1bash.

ps aux | grep -v grep | grep " miner " | sed 's/ /\n/g' | grep "/home/m1/" | rev | cut -d'/' -f 1 | rev

Nice,
Now another problem all ccminers are ccminer and directory name is identifier.
I should go with 1bash variables, much easier for sure.
Thanks any way.

Why don't you use source 1bash in your script then read any variable from 1bash to what is set instead of using ps and grep... eg. if $COIN="ZEC" then...

I had an idea to create REPORT script in similar manner so that people can run bash REPORT then post their problem with the report attached to the message in the forum. It will make troubleshooting and answering questions much easier than asking them did you set this, did you set that, what's your OC value... I was thinkering about how to execute it and it seems pulling variable settings from 1bash and supplementing with nvidia-smi/nvidia-settings is the best approach. With all the values on hand it's easy to either create report output or create web stats. We can work together on this... I think next week my schedule is more relaxed and i can work a bit on this

Great idea, it would help us to find and fix the problems for all newbies very quickly and easily. Looking forward for it, I may add bits and bobs once I see your initial version. Good Luck

I like your idea leenoox; it will make troubleshooting a lot easier. 

[fullzero]

One of my rigs running nvOC is using a MSI Z270-A PRO mobo with 7x 1080Ti's.  Rig is working great, but I got another 1080Ti I want to add to this rig and I'm all maxed out with 6 GPUs in the PCIe slot and the 7th in the M.2 slot via an adaptor.

So I'm going to swap out the mobo for an Asus PRIME Z270-A.  Already got the BIOS on the Asus configured (this will be my 2nd rig with a PRIME Z270-A).

So my question is, can I just stick the 32GB USB with nvOC on it from the MSI in the PRIME after I move the GPU's over, and expect it to just come up and start mining?

If not, no big deal, I'll just image a fresh USB, but since this mobo will be a permanent switch, but if I can safe some time just throwing on the USB used with the MSI, I'd rather do that, but if I'm going to run into issues, I'd rather not find out after tearing the rig down and rebuilding it with the new mobo.

It should be ok to move over; if your pcie slot utilization is different the xorg might need to do an extra reboot to adjust.

[fullzero]

I couldnt find anyone talking about this or im a complete newb but once you Edit the bach file how do you edit it again once its ran i cant seem to adjust anything permanently once i run of the usb stick am i missing something? or is there a place im suppose to "edit the values" like another config file stored somewhere in the subfolders im lost?

Before the first boot you can edit the 1bash file in the first partition on the usb key.  When you boot for the first time: that copy of 1bash replaces the default in the primary partition (and is removed from the first partition).  After this you can go to the /home/m1 directory ( click the file cabinet icon on the left) and open 1bash with gedit (double click or right click and select open with gedit) .  Make changes then save.

[fullzero]

Hi I know this is a  little too dumb,

I've read that the compressed file should be at least below 20GB, but in my case after using the provided link, It is way too bigger.



Is there something wrong with the link? after I downloaded it, the file size is Only @ 5.99GB (which was posted). did I missed something?

all replies will be appreciated.

the zip should be 6.4 gb:  unzip the img file from the zip and it should be 15.4gb  follow the guide linked on the OP to check your sha256sum and verify the downloaded zip.

[fullzero]

The rig restart the miner each 11-12 hours, only the miner without to restart the OS. I am curently mining ETH+DCR on suprnova. On nicehash (ETH+DCR) it didnt happent with the same configuration. Is it normal?

Thx!

Not normal,
Set telegram alert so it notify you if gpu lost, low utilization , ....

I just did it. Thx for advice. Let's see if it will restart again and what would be there about temps, gpu's.....

Can you please explain what is the difference between the 3 types of telegrams? papampi, baliminer and kk003.

They send different info (a lot of overlap however) in different formats.

[fullzero]

anyone help me out with the telegram?

I got my API key from @BotFather and my ID from the IDBOT, I inserted them in the correct spots, made sure telegram was turned on and the watchdog was turned on as well, both are "YES" but I'm not receiving any notifications. what did I miss?

see:

https://bitcointalk.org/index.php?topic=1854250.msg22696008#msg22696008

and:

https://bitcointalk.org/index.php?topic=1854250.msg21430404#msg21430404

[fullzero]

Still cannot solve the issue.
I tried with nvOc 0018 and 0019.
To sum up:
I have MSI Z270 gaming plus (quite similar to MSI Z270-a PRO).
Previously 4 ZOTAC Mini 1060 6gb
Now added two EVGA 1060 6Gb
Celeron G3930

• Was mining well with 4 Zotac on nvOc0018 and OC
• I added two EVGA 1060 6Gb to the build, only 4 gpus detected always
• I enabled above 4G memory, 6 gpus detected, starts mining and the OS hangs after 5 min always
• I updated MB Bios and now it throws this error: https://imgur.com/a/QblT1 (No overclock)
• I tried with nvOc 0019 and it restarts after 10 minutes

Okay, I've found the problem:
Two days ago I tried running the whole rig with a wired connection to the router (it was previously connected via tl-wn722n wireless adapter).
It worked, it runned for a day without issues.
So I moved the rig to a friends house, to test his wifi connection, and it also worked.

So, to sum up:
I enabled above 4g memory in my motherboard and connected two more gpus, that caused problems with my wifi connection, and finished in a constant OS hang.

Any ideas for whats going on here?

There is a problem with one of ( I haven't narrowed it down yet) the Ubuntu updates and some specific NICs.  I confirmed this by switching the NIC on mobos I had 19-1.2 Internet problems with: to use a USB NIC.  After switching the problems disappeared, so I know it is a compatibility problem from one of the updates (as they were working correctly before the updates).