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

[fullzero]

First off thanks for all the help thus far. But I am running in to a problem and would really appreciate it if someone could give me a helping hand.

So initial I had 5x 1070 2 connected via risers and 3 connect directly via the motherboard. I then bought some more risers so I could connect all of them via risers and someone suggested to do a fresh install whenever doing any hardware changes so I did that. However, now regardless of what configuration and the way I setup the gpus either the original with some via the motherboard and some via risers or all of them on risers it is only picking up 4 out of the 5 gpus. Any troubleshooting suggestions I tried look through to see if anyone else had this issue, but couldn't find anything that worked. Also I have tried multiple fresh installs on a few USB drives. Once again thanks in advance!

What motherboard are you using?

I am upgrading to a 6x PCIe mobo next week so I hope to not run into this issue?

I am currently using the Asus Prime Z270-A

1. Will the new cards work if you run 2 old ones with the 2 new ones?
2. Did you follow what is in the OP?

If you are using an ASUS PRIME Z270-A; ensure 'Above 4G Decoding' is enabled in the bios.  Also ensure PTP aware OS: is set to 'Not PTP Aware' in the bios.  Finally, ensure you 'Clear Secure Boot Keys' in the bios.

Also, I've read that PCIe setting needs to be set to Gen1. It sounds like it could be a simple BIOS issue.

Thanks! That was it guess it must have reset the changes in the bios when I did the fresh install or something.

When connecting some GPUs directly; you may need to manually set the pcie bandwidth to gen1. 

If you are using risers you shouldn't need to do this.

The reason is that when you connect a GPU directly to a mobo it will usually default to using the max bandwidth for that port; either 16x / 8x / 4x. This hogs the bandwidth and can prevent the proper recognition of all cards.

[1714287168]

1714287168

[1714287168]

1714287168

[fullzero]

So my 2nd rig is now running with 6 GPUs. 

Rig 1 https://i.imgur.com/Q6kLeV3.jpg
 • Asus Prime Z270-A Motherboard
 • Corsair HX1200i PSU
 • Celeron G3930 CPU, 4GB RAM
 • (7) Gigabyte GTX 1070 Founders Edition

   CC: 200
   MC: 1600 (4608 MHz mclk in nvidia-smi dmon)
   pl: 120W

   945W at the wall

   ~220 MH/sec ETH (31.4 mhs per GPU, 26% over stock)


Rig 2 https://i.imgur.com/SUp7cj4.jpg
 • Asus Prime Z270-AR Motherboard
 • Corsair HX1200i + HX1000i (using Add2PSU)
 • Celeron G3930
 • (4) EVGA GTX 1070 FTW, (2) EVGA GTX 1070 SC

   CC: 1
   MC: 1100 (4353 MHz mclk in nvidia-smi dmon)
   pl: 125W

   885W at the wall

   ~180 MH/sec ETH (30 mhs per GPU, 20% over stock)
   

Rig 1 is nice and stable.  It's been mining consistently for over 62 hours.  Rig 2 is semi-stable: It's been running for just under 24 hours.  It doesn't lock up, but I do get error messages from Claymore randomly that cause Claymore to restart on average of once an hour.  I will try lowering MC to eliminate those.

Both rigs hashing: https://i.imgur.com/vx4hs88.png

nvidia-smi dmon output for both rigs: https://i.imgur.com/xSAoM5T.png

Nanopool: https://i.imgur.com/uFf52Hn.png


I need to get a case built to improve cooling and to ensure these GPUs aren't at risk of falling over like dominoes.  I had planned on working on that last weekend, but then I made the decision to build a 2nd rig so I held off.  What I am thinking of doing is repurposing a 4' wide x 4' deep x 6.5' tall indoor gardening tent (similar to this: https://i.imgur.com/whULDww.jpg)and placing the rigs inside so I can keep the heat output under control and hopefully increase cooling.  I'll have an exhaust fan sucking out air from the top of the tent, and passive intakes at the bottom pulling in cool air.  I'll route a 6" duct from the exhaust fan to a window to blow the heated air outside (until winter time!) so that my central air doesn't have to deal with the extra heat I'm currently introducing.  Rather than build a case/frame out of wood or aluminum I am thinking of using a wire shelving unit.  I like the idea of building a nice aluminum frame, but I'm not 100% sure if I am going to stop at 2 rigs or keep going, so until that clears up I am not going to construct a custom frame.

As I was typing this rig 2 experienced one of the Claymore errors:

GPU 0, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 0, GpuMiner kx failed 1
GPU 0, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 0, GpuMiner kx failed 1
GPU 1, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 1, GpuMiner kx failed 1
GPU 1, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 1, GpuMiner kx failed 1
GPU 2, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 2, GpuMiner kx failed 1
GPU 4, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 4, GpuMiner kx failed 1
GPU 4, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 4, GpuMiner kx failed 1
GPU 2, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 2, GpuMiner kx failed 1
GPU 5, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 5, GpuMiner kx failed 1
GPU 5, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 5, GpuMiner kx failed 1
GPU 3, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 3, GpuMiner kx failed 1
GPU 3, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 3, GpuMiner kx failed 1

WATCHDOG: GPU error, you need to restart miner

If your going to use a wire shelving unit; I would look at citronick's setup.

I have never seen anyone use an indoor gardening tent to contain / control airflow.  I think this is a good idea for home mining.  How much do these cost?

If that claymore error happens frequently; I would lower the OC a little and see if it stops.

[fullzero]

You guys seeing NVIDIA cards going to 0 hash often?  I'm seeing this very often on Zotac Mini and 1080TI Gigabyte rigs

With Claymore 9.5 yes (on AMD rigs), back to 9.4 for me. Have had it with too much OC on Gigabyte mini, running for 20-40 minutes then back to 0.

I will switch one of my 1070 mini rigs to ETH and see if this happens for me as well.  What OC settings are you using when this happens?

[fullzero]

How it is with mainboard Biostar TB250-BTC? Is it working with M.2 slot PCI-E adapter? Somebody with experiences?

When I tested; the TB250-BTC would not support 7x GPUs using an m2 adapter. 

[fullzero]

Hey there, I've set this up and once the rig booted and the miner started, it displays an error with assigning value for GPUFanTargetSpeed? Anyone has a solution for this?

Code:

ERROR: Error assigning value 70 to attribute 'GPUTargetFanSpeed'
       (m1-desktop:0[fan:0]) as specified in assignment
       '[fan:0]/GPUTargetFanSpeed=70' (Unknown Error).

I responded to your pm.

[fullzero]

Hey,

Having all sorts of problems getting 6 cards running on Z270-AR, just get "internal driver error" in the one bash output then it spits a load of too many arguments statements.

I've re-imaged fresh and confirmed the MD5.

I've done the standard from BIOS including the latest flash, 4g decoding, non ptp aware and clear boot keys, no joy is there a different image for this one based on the Z270-A or are we all standard now?

Any ideas?

Thanks again great work love the project.

Are you connecting the monitor to the motherboard?

If you are change it to the primary GPU (the first 16x pcie slot closest to the CPU)

and follow the process on this thread:

https://bitcointalk.org/index.php?topic=1854250.msg19449945#msg19449945

[fullzero]

Any ideas on why boot is taking 20+ minutes and then once the miner starts everytime it runs one of the commands the entire system freezes for a solid 3-4 minutes. Any thing obvious that I am missing? Running of an SSD after following the provide instruction and I have tried no OC and fresh install. And once it finally finishes running all the commands it can never connect to the pool and I get a Share rejected error.


Edit: Just tested on a USB same issue

So I figured out the issue for some reason it lags hard when I connect 7x GPUs at once. @fullzero any clue on what you think might be causing this?

Not sure if this will help, but its worth a try.

Try installing updates:

go to the ubuntu launcher at the top left and type: u

you should see the ubuntu update app

open it and install all the updates.

[keepitsane]

Any ideas on why boot is taking 20+ minutes and then once the miner starts everytime it runs one of the commands the entire system freezes for a solid 3-4 minutes. Any thing obvious that I am missing? Running of an SSD after following the provide instruction and I have tried no OC and fresh install. And once it finally finishes running all the commands it can never connect to the pool and I get a Share rejected error.


Edit: Just tested on a USB same issue

So I figured out the issue for some reason it lags hard when I connect 7x GPUs at once. @fullzero any clue on what you think might be causing this?

Not sure if this will help, but its worth a try.

Try installing updates:

go to the ubuntu launcher at the top left and type: u

you should see the ubuntu update app

open it and install all the updates.

Okay I will try that. In the bios I set it to Gen1 just to be sure even though I am using risers. It picks up all 7 and it runs for about 20-30 mins and then I get the same error as UberDaemon


GPU 0, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 0, GpuMiner kx failed 1
GPU 0, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 0, GpuMiner kx failed 1
GPU 1, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 1, GpuMiner kx failed 1
GPU 1, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 1, GpuMiner kx failed 1
GPU 2, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 2, GpuMiner kx failed 1
GPU 4, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 4, GpuMiner kx failed 1
GPU 4, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 4, GpuMiner kx failed 1
GPU 2, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 2, GpuMiner kx failed 1
GPU 5, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 5, GpuMiner kx failed 1
GPU 5, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 5, GpuMiner kx failed 1
GPU 3, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 3, GpuMiner kx failed 1
GPU 3, GpuMiner cu_k1 failed 4, unspecified launch failure
GPU 3, GpuMiner kx failed 1

WATCHDOG: GPU error, you need to restart miner


This is with and without OC.

[citronick]

Hi fullzero,

Back to ethpool, dwarfpool settings:

in onebash file - can I change ETH section from:

ETHERminedotORG="NO"
ETH_WORKER="nvoc"
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="eth-us-east1.nanopool.org:9999"

to

ETHERminedotORG="NO"
ETH_WORKER="citronick -esm 1 -dbg -1 -wd 1 -r 1 "
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="stratum+tcp://asia1.ethpool.org:3333"

Perhaps, you may consider to add ETH_ESM and the other key Claymore keyswicthes or alternatively include a ETH_EXTENTION to append any other Claymore stuff. Hopefully the failover pool features too soon.

All of my 1070s are still all Windows and smOS Nvidia Alpha v1, and would like to migrate all of them to nvOC ETH dwarfpool (-esm 0) or ethpool with esm settings.

While all my 1080tis are already on ZEC nvOC with very good results using EWBF.

Thanks

[xleejohnx]

Hi fullzero,

Back to ethpool, dwarfpool settings:

in onebash file - can I change ETH section from:

ETHERminedotORG="NO"
ETH_WORKER="nvoc"
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="eth-us-east1.nanopool.org:9999"

to

ETHERminedotORG="NO"
ETH_WORKER="citronick -esm 1 -dbg -1 -wd 1 -r 1 "
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="stratum+tcp://asia1.ethpool.org:3333"

Perhaps, you may consider to add ETH_ESM and the other key Claymore keyswicthes or alternatively include a ETH_EXTENTION to append any other Claymore stuff. Hopefully the failover pool features too soon.

All of my 1070s are still all Windows and smOS Nvidia Alpha v1, and would like to migrate all of them to nvOC ETH dwarfpool (-esm 0) or ethpool with esm settings.

While all my 1080tis are already on ZEC nvOC with very good results using EWBF.

Thanks

To add those commands just scroll down to the bottom of the onebash and add them to the claymores command line settings 

[citronick]

Hi fullzero,

Back to ethpool, dwarfpool settings:

in onebash file - can I change ETH section from:

ETHERminedotORG="NO"
ETH_WORKER="nvoc"
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="eth-us-east1.nanopool.org:9999"

to

ETHERminedotORG="NO"
ETH_WORKER="citronick -esm 1 -dbg -1 -wd 1 -r 1 "
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="stratum+tcp://asia1.ethpool.org:3333"

Perhaps, you may consider to add ETH_ESM and the other key Claymore keyswicthes or alternatively include a ETH_EXTENTION to append any other Claymore stuff. Hopefully the failover pool features too soon.

All of my 1070s are still all Windows and smOS Nvidia Alpha v1, and would like to migrate all of them to nvOC ETH dwarfpool (-esm 0) or ethpool with esm settings.

While all my 1080tis are already on ZEC nvOC with very good results using EWBF.

Thanks

To add those commands just scroll down to the bottom of the onebash and add them to the claymores command line settings 

hmmm... I finally see the light!

Fullzero pointed me to that part too... got a syntax error but after looking them again in a bigger screen (LOL) ... I think I finally got it - will revert later.

[yun9999]

You guys seeing NVIDIA cards going to 0 hash often?  I'm seeing this very often on Zotac Mini and 1080TI Gigabyte rigs

With Claymore 9.5 yes (on AMD rigs), back to 9.4 for me. Have had it with too much OC on Gigabyte mini, running for 20-40 minutes then back to 0.

I will switch one of my 1070 mini rigs to ETH and see if this happens for me as well.  What OC settings are you using when this happens?

For Zotac Mini it is 200 Core and 450 Mem.  PL is default at 125.  For Gigabyte 1080Ti  both Default with No PowerLimit or 100 Core and 900 Mem and PL 225.  My brother in law reported this issue first and I never saw it until later on the Zotac mini but was seeing it on the 1080Ti.  It has become more often the more rigs I have so he's creating a program to detect 0 hash and auto reboot the rig

[Coin++]

Thank you for your very good work :-).

For nVidia mining with Ubuntu, I use the nvidia-smi intrface with the driver.
With nvidia-smi, I can know the frequencies that I can set up for GPu and memory, but that is never +500 like you do...

How can you set up yours GPUs at so high value with Ubuntu (with Windows, it is very simple via AfterBurner, for example) ?

And other question : with nvOC, can you undervolt nVidia GPUs ?

[xleejohnx]

In the onebash file does it monitor the mining software  and restart them if they fail?

[cryptosharks]

Great job!

https://i.imgur.com/WhBrVj9.png

Do you have any idea what this issue might be? Everything was fine, I was able to set OC values. The OC values persist, but I'm unable to update the OC values. A part of the reason has to be this.

Please let me know.

[bibyfok]

My Nvidia rig are running amazingly with nvOC! Thank you so much

Do you have any ETA for OC and power limit for AMD card?

On an other topic, I'm trying to identify each port number on my motherboard. Is there a way to manually force 1 fan to spin up fast?
Like I want to run GPU3 fan 100% (and all others at 20%) to locate it, is there a way to force it under Ubuntu when the rig is running?

Thanks!

[fullzero]

Hi fullzero,

Back to ethpool, dwarfpool settings:

in onebash file - can I change ETH section from:

ETHERminedotORG="NO"
ETH_WORKER="nvoc"
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="eth-us-east1.nanopool.org:9999"

to

ETHERminedotORG="NO"
ETH_WORKER="citronick -esm 1 -dbg -1 -wd 1 -r 1 "
ETH_ADDRESS="0x.....blah..blah..blah"
ETH_POOL="stratum+tcp://asia1.ethpool.org:3333"

Perhaps, you may consider to add ETH_ESM and the other key Claymore keyswicthes or alternatively include a ETH_EXTENTION to append any other Claymore stuff. Hopefully the failover pool features too soon.

All of my 1070s are still all Windows and smOS Nvidia Alpha v1, and would like to migrate all of them to nvOC ETH dwarfpool (-esm 0) or ethpool with esm settings.

While all my 1080tis are already on ZEC nvOC with very good results using EWBF.

Thanks

To add those commands just scroll down to the bottom of the onebash and add them to the claymores command line settings 

hmmm... I finally see the light!

Fullzero pointed me to that part too... got a syntax error but after looking them again in a bigger screen (LOL) ... I think I finally got it - will revert later.

I will had more customization options for Claymore in the next version based on your request.

[fullzero]

Thank you for your very good work :-).

For nVidia mining with Ubuntu, I use the nvidia-smi intrface with the driver.
With nvidia-smi, I can know the frequencies that I can set up for GPu and memory, but that is never +500 like you do...

How can you set up yours GPUs at so high value with Ubuntu (with Windows, it is very simple via AfterBurner, for example) ?

And other question : with nvOC, can you undervolt nVidia GPUs ?

I recommend watching one of the demo videos on the OP to get an idea of how nvOC works.

I have not tested undervolting and I'm not sure if it is supported by the 378.13 Driver. 

I will add this to my list.

[fullzero]

You guys seeing NVIDIA cards going to 0 hash often?  I'm seeing this very often on Zotac Mini and 1080TI Gigabyte rigs

With Claymore 9.5 yes (on AMD rigs), back to 9.4 for me. Have had it with too much OC on Gigabyte mini, running for 20-40 minutes then back to 0.

I will switch one of my 1070 mini rigs to ETH and see if this happens for me as well.  What OC settings are you using when this happens?

For Zotac Mini it is 200 Core and 450 Mem.  PL is default at 125.  For Gigabyte 1080Ti  both Default with No PowerLimit or 100 Core and 900 Mem and PL 225.  My brother in law reported this issue first and I never saw it until later on the Zotac mini but was seeing it on the 1080Ti.  It has become more often the more rigs I have so he's creating a program to detect 0 hash and auto reboot the rig

If you lower the cc to -100 on both the 1070 and the 1080 ti; does this still happen?

I will work on a 0 hash state detection and restart method.

[fullzero]

In the onebash file does it monitor the mining software  and restart them if they fail?

In any of the blocks that call the selected COIN; take ZEC for example:

Code:

if [ $COIN == "ZEC" ]
then
HCD='/home/m1/zec/miner'
ZECADDR="$ZEC_ADDRESS.$ZEC_WORKER"
until $HCD --eexit 3 --fee $EWBF_PERCENT --pec --server $ZEC_POOL --user $ZECADDR --pass z --port $ZEC_PORT;
   do
   echo "FAILURE; reinit in 5" >&2
   sleep 5
done
fi

this part is essentially a loop:

Code:

until $HCD --eexit 3 --fee $EWBF_PERCENT --pec --server $ZEC_POOL --user $ZECADDR --pass z --port $ZEC_PORT;
   do
   echo "FAILURE; reinit in 5" >&2
   sleep 5
done

Code:

--eexit 3

this argument tells EWBF to close on error

when it closes this line is executed:

Code:

echo "FAILURE; reinit in 5" >&2  

then: 

Code:

sleep 5

then this line:

Code:

until $HCD --eexit 3 --fee $EWBF_PERCENT --pec --server $ZEC_POOL --user $ZECADDR --pass z --port $ZEC_PORT;