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

[fullzero]

Dressed up the telegram message a little 

Code:

#!/bin/bash
# Telegram Info Script
# By BaliMiner et al...
# for nvOC by fullzero
# ref: http://bernaerts.dyndns.org/linux/75-debian/351-debian-send-telegram-notification
#

source ~/1bash

CHATID=$TELEGRAM_CHATID
APIKEY=$TELEGRAM_APIKEY
WORKERNAME=$WORKERNAME
SYSTEM_BOOT_TIME=$(uptime -s)
SYSTEM_UP_TIME=$(uptime -p)
CURRENTHASH=`/usr/bin/curl -s http://localhost:3333 | sed '/Total/!d; /Speed/!d;' | awk '{print $6}' | awk 'NR == 3'`
MINER_UP_TIME=$(ps -p `pgrep miner` -o etime | grep -v ELAPSED)
GPU_UTILIZATIONS=$(nvidia-smi --query-gpu=utilization.gpu --format=csv,noheader,nounits)
TEMP=$(/usr/bin/nvidia-smi --query-gpu=temperature.gpu --format=csv,noheader)
PD=$(/usr/bin/nvidia-smi --query-gpu=power.draw --format=csv,noheader)
FAN=$(/usr/bin/nvidia-smi --query-gpu=fan.speed --format=csv,noheader)
LF=$'\n'
PROFIT=$(cat /home/m1/current-profit)
MSG="$LF Worker: $WORKERNAME
$LF System Boot Time: $SYSTEM_BOOT_TIME
$LF System Up Time: $SYSTEM_UP_TIME
$LF Miner Uptime: $MINER_UP_TIME
$LF GPU Utilizations:
$GPU_UTILIZATIONS
$LF GPU Temps:
$TEMP
$LF GPU Fan Speeds:
$FAN
$LF  Power Draws:
$PD
$LF  $CURRENTHASH"
/usr/bin/curl -s -X POST --output /dev/null https://api.telegram.org/bot${APIKEY}/sendMessage -d "text=${MSG}" -d chat_id=${CHATID}

Nice changes: I will try your version out later today.  I haven't heard from BaliMiner, so I will make a telegram guide soon (I had 300 telegrams this morning on the test bot).

papampi do you want to make the guide?  I know at least: you, _Parallax_, and lost_post have setup telegrams.  _Parallax_ and lost_post both have heavly integrated it (different messages for different events ect)  Each has done so in a different way; I'll optimize a version of their methods / code for v002.

[1713560197]

1713560197

[1713560197]

1713560197

[1713560197]

1713560197

[1713560197]

1713560197

[1713560197]

1713560197

[fullzero]

Make sure the code for UBIQ in 3main file matches with the format in this link :

https://ubiqpool.io/#/help

Thanks! This was the solution. For the Stratum proxy I need to use Claymore, so first it was my fault to test it with Etminer (with just http this is also possible).

Also I had to change some lines in the 3main file because of this format from Ubiqpool.

Code:

EthDcrMiner64.exe -epool stratum+tcp://us.ubiqpool.io:8008 -ewal 0xb85150eb365e7df0941f0cf08235f987ba91506a -epsw x -eworker rig1

The problem was in this part of 3main because the Stratum proxy needs another format like mentioned above. So what I did is remove and move the $UBQ_WORKER part and now it's working

Code:

UBQADDR="$UBQ_ADDRESS/$UBQ_WORKER"

if [ $DOT_POOL_FORMAT_or_FORWARD_SLASH_POOL_FORMAT == "DOT" ]
then
UBQADDR="$UBQ_ADDRESS.$UBQ_WORKER"
fi

screen -dmS miner $HCD -epool $UBQ_POOL -ewal $UBQADDR -epsw x -mode 1 -allpools 1 -dbg -1 $UBQ_EXTENSION_ARGUMENTS

This is the first time I've seen this syntax used.

[fullzero]

I just placed an order for some g4600's. I will see how that works - the rig that had been crashing every couple hours crashed after 24 hours after I turned off teamviewer - hopefully this puts me over the top.

As for power, both are built the same way. Each is running on 2 750w EVGA G3 power supplies. Each SATA/molex power connector is running no more than 2 GPU risers. The molex connectors (and single SATA) hooked to the board each share with the riser for a single GPU (this means 13 GPU's, 2 motherboard molex, and 1 motherboard SATA to connect. The PSU's each have 3 SATA and 1 Perif (molex) connector, hence my using two connections per cable). I did have a 550w PSU in the mix (as a third) to see if that would change things, but it did not.

The only other discernable different is that not all GPU's are the same brand/make (but all are 1060 6GB), and the risers are not all identical (though I have switched them out in troubleshooting).

Good, G4600 is 2 core, 4 threads CPU. That will definetely help, much better than Celerons you have.

As for the mixed GPU's, try to put as many of the same brand/model in the same rig. Then you will have to manualy set overclocking for each GPU. Don't use the global OC for all. Try with the lowest stable value for all (I believe it was about 600 memory for you), then increase +50 memory on one brand/model and see if it's stable, then try +50 on different model and see if it's stable, then repeat until you get max for all models. It will take a while to fine tune it, that's the downside of mixed cards. Good luck and keep us posted.

Hi there nvoc-ers !

I may be stupid or definitely not paying enough attention to something as i can't for sake of me to be able to make AsRock H110 to recognize 13 x 1060 6GB GPU's. Tried with nvoc17, then 18, then 19..does not see the gpu on the last pciex slot.
It mines beautifully with 12 GPU's but there is no way i can make it see all the 13 GPU's.
The rig is built using a G4560, i tried using an i3... same result.
2 x Corsair 1000 W PSU's powering 6 gpu + mobo + risers and the 2nd powering 7 gpu's

Any hints ?

I struggled with this for a very long time. For me, it came down to a bad riser. I could have sworn I had tested and replaced them all, but this past weekend I started with one, and kept adding 1, one boot at a time, until I got to 13 working. I swear I did this a few times, but i find the pic-e risers LOVED to come loose as well, and may have had bad luck swapping a bad riser for a bad riser.

Aside from that, if you haven't, I would make sure the the 2 molex connectors on the motherboard are each connected to the separate power supplies.

Also, I noticed you have all risers on one PSU. I believe I have read that you should not have more than 2 per connector - I am unsure how many connectors are on those corsairs, but it might be worth trying splitting them up if you haven't.

Let me know if you get anywhere with this.

Another potential problem; that is often overlooked is slightly bent risers.  If when you attach risers (depending on what kind of frame you are using) you place any (even slight, warping pressure (bending the front or rear of the riser sightly up or down while securing the other end) this will cause intermittent hard crashes.  I have a rig that crashed last night: and I suspect this is the case with it.

[cosmincucu]

First of all thanks FullZero for the awesome mining OS provided. It's made my mining days much easier, at least from the OS point of view.

Not the issue : i am running 7x 1070 different brands, on an Asus Prime 270A, G4400 Celeron CPU, and some ram.

NvOC v18 worked fine, V19 starts, but in terminal, while mining ETH (i don't think the coin matters) it doesn't go past :

LAUNCHING:  Mining Process

process in screen miner; attach with: screen -r miner


I would like the Claymore stats , as i log in from time to time to see temps. Anyone have any idea why it does not go past that?

Also, i keep my cards at 115W, is that enough ? Overclock-wise i am supposed to try again and increase each card's memory as i shuffled them around a bit.

[fullzero]

Someone has been able to unzip the file tree from the nvOC 0.19 (1.img) image on windows. I did it for the 0.18 with 7-zip but now says that it can not. The zip checksum is correct. Thx

Why not use winrar?

I haven't tested 7-zip.  The built in windows zip function should work; as well as winrar.  I have been using linux only (I image with dd or the Disks tool (which is a GUI front end for dd + a few additional features)

[fullzero]

How would I go about deleting an algo from the SALFTER MPH switcher?  One of my cards just doesn't like keccak.

Open the mph_switcher in gedit and prepend a # to lines referencing keccak.

Looking at Scott's gitlab:

https://gitlab.com/salfter/mph_switch/blob/master/mph_switch.py

he has updated the mph switch to allow: manual override by specifying algo & coin on command line; use "list" to get options

https://gitlab.com/salfter/mph_switch/commit/8fa93d659b8a206ecb9091e84b1e9734efcbd997

I will add these changes in the next set of updates.

this is the relevant part:

Code:

else: # manual override 
  if (sys.argv[2]!="list"):
    miner=miners[sys.argv[2]]
    coin=coins[sys.argv[3]]
  else: # list available miners & coins
    print("algos: ", end="")
    for i in miners:
      print(i+" ", end="")
    print("")
    print ("coins: ", end="")
    for i in coins:
      print(i+" ", end="")
    print("")
    sys.exit()

[fullzero]

First of all thanks FullZero for the awesome mining OS provided. It's made my mining days much easier, at least from the OS point of view.

Not the issue : i am running 7x 1070 different brands, on an Asus Prime 270A, G4400 Celeron CPU, and some ram.

NvOC v18 worked fine, V19 starts, but in terminal, while mining ETH (i don't think the coin matters) it doesn't go past :

LAUNCHING:  Mining Process

process in screen miner; attach with: screen -r miner


I would like the Claymore stats , as i log in from time to time to see temps. Anyone have any idea why it does not go past that?

Also, i keep my cards at 115W, is that enough ? Overclock-wise i am supposed to try again and increase each card's memory as i shuffled them around a bit.

If you want to have the mining process auto attach; change this setting in 1bash to:

Code:

LOCALorREMOTE="LOCAL"

or you can open the guake terminal ( press f12) and type:

Code:

screen -r miner

and press enter.

If you watch the v0019 demo video (linked at the top of the OP)  I demo interacting with SSH and screen and explain why I set the default to remote.

[stefcho25]

Does this OS support z270p motherboard.
I've tried to run it with 6 gpu and it restarts.

[cosmincucu]

First of all thanks FullZero for the awesome mining OS provided. It's made my mining days much easier, at least from the OS point of view.

Not the issue : i am running 7x 1070 different brands, on an Asus Prime 270A, G4400 Celeron CPU, and some ram.

NvOC v18 worked fine, V19 starts, but in terminal, while mining ETH (i don't think the coin matters) it doesn't go past :

LAUNCHING:  Mining Process

process in screen miner; attach with: screen -r miner


I would like the Claymore stats , as i log in from time to time to see temps. Anyone have any idea why it does not go past that?

Also, i keep my cards at 115W, is that enough ? Overclock-wise i am supposed to try again and increase each card's memory as i shuffled them around a bit.

If you want to have the mining process auto attach; change this setting in 1bash to:

Code:

LOCALorREMOTE="LOCAL"

or you can open the guake terminal ( press f12) and type:

Code:

screen -r miner

and press enter.

If you watch the v0019 demo video (linked at the top of the OP)  I demo interacting with SSH and screen and explain why I set the default to remote.

I'm sure you had your reasons for setting it up this way. But me being male, and a techie, why would i read the manual, or see the orientation videos ,when i know stuff?

Thanks for your help, the answer was obvious and right in front of me, i'm just slightly blind.

[papampi]

Dressed up the telegram message a little  

Code:

#!/bin/bash
# Telegram Info Script
# By BaliMiner et al...
# for nvOC by fullzero
# ref: http://bernaerts.dyndns.org/linux/75-debian/351-debian-send-telegram-notification
#

source ~/1bash

CHATID=$TELEGRAM_CHATID
APIKEY=$TELEGRAM_APIKEY
WORKERNAME=$WORKERNAME
SYSTEM_BOOT_TIME=$(uptime -s)
SYSTEM_UP_TIME=$(uptime -p)
CURRENTHASH=`/usr/bin/curl -s http://localhost:3333 | sed '/Total/!d; /Speed/!d;' | awk '{print $6}' | awk 'NR == 3'`
MINER_UP_TIME=$(ps -p `pgrep miner` -o etime | grep -v ELAPSED)
GPU_UTILIZATIONS=$(nvidia-smi --query-gpu=utilization.gpu --format=csv,noheader,nounits)
TEMP=$(/usr/bin/nvidia-smi --query-gpu=temperature.gpu --format=csv,noheader)
PD=$(/usr/bin/nvidia-smi --query-gpu=power.draw --format=csv,noheader)
FAN=$(/usr/bin/nvidia-smi --query-gpu=fan.speed --format=csv,noheader)
LF=$'\n'
PROFIT=$(cat /home/m1/current-profit)
MSG="$LF Worker: $WORKERNAME
$LF System Boot Time: $SYSTEM_BOOT_TIME
$LF System Up Time: $SYSTEM_UP_TIME
$LF Miner Uptime: $MINER_UP_TIME
$LF GPU Utilizations:
$GPU_UTILIZATIONS
$LF GPU Temps:
$TEMP
$LF GPU Fan Speeds:
$FAN
$LF  Power Draws:
$PD
$LF  $CURRENTHASH"
/usr/bin/curl -s -X POST --output /dev/null https://api.telegram.org/bot${APIKEY}/sendMessage -d "text=${MSG}" -d chat_id=${CHATID}

Nice changes: I will try your version out later today.  I haven't heard from BaliMiner, so I will make a telegram guide soon (I had 300 telegrams this morning on the test bot).

papampi do you want to make the guide?  I know at least: you, _Parallax_, and lost_post have setup telegrams.  _Parallax_ and lost_post both have heavly integrated it (different messages for different events ect)  Each has done so in a different way; I'll optimize a version of their methods / code for v002.

Sure I will write how to setup telegram first thing tomorrow.
Would be nice to have all the info we can get in our telegram, waiting for telegram v002

[salfter]

How would I go about deleting an algo from the SALFTER MPH switcher?  One of my cards just doesn't like keccak.

Deleting the keccak block in mph_conf.json would work.

[salfter]

How would I go about deleting an algo from the SALFTER MPH switcher?  One of my cards just doesn't like keccak.

Open the mph_switcher in gedit and prepend a # to lines referencing keccak.

There shouldn't be references to any particular algorithm in mph_switch.py IIRC; it's algo-agnostic.  mph_conf.json is where you'd want to look.

Looking at Scott's gitlab:

https://gitlab.com/salfter/mph_switch/blob/master/mph_switch.py

he has updated the mph switch to allow: manual override by specifying algo & coin on command line; use "list" to get options

https://gitlab.com/salfter/mph_switch/commit/8fa93d659b8a206ecb9091e84b1e9734efcbd997

Another GPU arrived yesterday.  I put those changes in to facilitate benchmarking: run an algo for a bit, note the hashrate and power consumption, move onto the next.  This time, I also had the sense to note the difference between 3 and 4 GPUs, so that as I add the remaining GPUs to this rig, I can just tweak the config file appropriately and get back to mining.  Here's the result of that:

Code:

for each MSI GeForce GTX 1070 Gaming X 8G:
115 W pwr limit, all except Ethash: -100 gpu_oc 0 mem_oc, Ethash: -200 gpu_oc 1000 mem_oc
Cryptonight: 95 W, 780 H/s
Equihash: 129 W, 420 H/s
Ethash: 131 W, 34 MH/s
Groestl: 134 W, 30 MH/s
Keccak: 124 W, 446 MH/s
Lyra2RE2: 129 W, 29 MH/s
Myriad-Groestl: 135 W, 55 MH/s
NeoScrypt: 130 W, 750 kH/s
Qubit: 133 W, 18 MH/s
Scrypt: 127 W, 640 kH/s
Skein: 137 W, 355 MH/s
X11: 127 W, 11 MH/s

(Power's measured at the wall with a Kill-a-Watt.  My GPUs are powered by a 750W HP server power supply through one of those PCIe breakout boards.)

[VoskCoin]

Possible to mine nicehash equihash only?

When will KMD be integrated?

[hurvajs77]

Hi guys,

I've a problem with booting nvOS - never booted. I tried several USB sticks (16G and 32G), always successfully created (with Etcher, I've Mac). I tried on two computers.

Testing PC has ASUS ROG motherboard with 2x 1060.
RIG has ASRock H110 PRO BTC+ with 7x 1080Ti.

I tried connect DVI cable on primary GPU. On rig, I had a no signal, so I tried all GPUs - without signal. If I used motherboard DVI, so I got signal, but after choosing Ubuntu in boot menu, so it's freezy. After 30 minutes nothing happens.

Similar situation on second computer, I got monitor signal on both GPUs and motherboard, but always Ubuntu freezy and cursor blinking over 1 hour...

Can you help me, please? Thanks for any advice.

[leenoox]

Hi guys,

I've a problem with booting nvOS - never booted. I tried several USB sticks (16G and 32G), always successfully created (with Etcher, I've Mac). I tried on two computers.

Testing PC has ASUS ROG motherboard with 2x 1060.
RIG has ASRock H110 PRO BTC+ with 7x 1080Ti.

I tried connect DVI cable on primary GPU. On rig, I had a no signal, so I tried all GPUs - without signal. If I used motherboard DVI, so I got signal, but after choosing Ubuntu in boot menu, so it's freezy. After 30 minutes nothing happens.

Similar situation on second computer, I got monitor signal on both GPUs and motherboard, but always Ubuntu freezy and cursor blinking over 1 hour...

Can you help me, please? Thanks for any advice.

Disable on-board video (motherboard video) in the Bios. Connect monitor to GPU that is in the first x16 PCIe slot (the first long slot closest to the CPU).

[papampi]

Two problems still exists in v0019 both woth upPaste
1- If your setting is REMOTE when upPaste update 1bash and restart 3main, it does not close 3main and terminal and start a new one, a new Gnome-Terminal pops up while last one is still open.
2- If you enable ewbf monitoring api, when upPaste update and restart ewbf, it cant bind api port and gives api bind error.

Thanks for finding bugs.  With 1 it should be easy to add a pkill -e gnome-terminal to fix.  2 I will have to look into; I don't usually use the api with ewbf.  Let me know if you solve this before me.

May be killing gnome-terminal release the port
Testing it now

[kk003]

Someone has been able to unzip the file tree from the nvOC 0.19 (1.img) image on windows. I did it for the 0.18 with 7-zip but now says that it can not. The zip checksum is correct. Thx

Why not use winrar?

I haven't tested 7-zip.  The built in windows zip function should work; as well as winrar.  I have been using linux only (I image with dd or the Disks tool (which is a GUI front end for dd + a few additional features)

I uncompress nvOC_v0019.zip with winzip so it gives me nvOC_v0019.img. I uncompress nvOC_v0019.img with 7-zip and get this 3 files:
0.fat
1.img
2

I want to open 1.img and save the files tree to the hd so I can see the code etc.

So, no winrar, winzip 7-zip can do this. I don't find a software in windows to do this.
Its ok for me to uncompress on linux (VM in my case). Not sure how to do this with dd from tty. The VM do not have X server.
Thx.

[papampi]

Two problems still exists in v0019 both woth upPaste
1- If your setting is REMOTE when upPaste update 1bash and restart 3main, it does not close 3main and terminal and start a new one, a new Gnome-Terminal pops up while last one is still open.
2- If you enable ewbf monitoring api, when upPaste update and restart ewbf, it cant bind api port and gives api bind error.

Thanks for finding bugs.  With 1 it should be easy to add a pkill -e gnome-terminal to fix.  2 I will have to look into; I don't usually use the api with ewbf.  Let me know if you solve this before me.

May be killing gnome-terminal release the port
Testing it now

Confirmed, killing gnome-terminal released the port, API is fixed now after upPaste update
Obviously multiple 3main and gnome-terminal fixed too.

[gs777]

Hi there nvoc-ers !

I may be stupid or definitely not paying enough attention to something as i can't for sake of me to be able to make AsRock H110 to recognize 13 x 1060 6GB GPU's. Tried with nvoc17, then 18, then 19..does not see the gpu on the last pciex slot.
It mines beautifully with 12 GPU's but there is no way i can make it see all the 13 GPU's.
The rig is built using a G4560, i tried using an i3... same result.
2 x Corsair 1000 W PSU's powering 6 gpu + mobo + risers and the 2nd powering 7 gpu's

Any hints ?

I struggled with this for a very long time. For me, it came down to a bad riser. I could have sworn I had tested and replaced them all, but this past weekend I started with one, and kept adding 1, one boot at a time, until I got to 13 working. I swear I did this a few times, but i find the pic-e risers LOVED to come loose as well, and may have had bad luck swapping a bad riser for a bad riser.

Aside from that, if you haven't, I would make sure the the 2 molex connectors on the motherboard are each connected to the separate power supplies.

Also, I noticed you have all risers on one PSU. I believe I have read that you should not have more than 2 per connector - I am unsure how many connectors are on those corsairs, but it might be worth trying splitting them up if you haven't.

Let me know if you get anywhere with this.

I put 3 risers per 1 connector - works fine

[bwarbwar]

Thanks for the help Fullzero and Salfter!