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

[fullzero]

Love this OS.   

I've been using a new mobo:  Biostar Z270GT6; got it to work all day yesterday on 5 GPUs.  Tinkered a bit today, tried to boot mobo, got pass BIOS screen, then got spammed with PCIE bus error severity=corrected.  It just spammed and spammed non stop, and would probably never stop if I hadn't powered down. 

A similar issue that is answered can be found here: https://askubuntu.com/questions/771899/pcie-bus-error-severity-corrected 

PS:  32GB USB still has some imaging issues, I've tried reformatting without the "quick" option, and tried both NTFS and FAT32. 

I had those type of errors with other biostar mobos before; they went away when I changed:

Miner Mode to enabled in the bios. 

Max TOLUD set to 3.5 GB in the bios.

Try using the disks application in linux; format in FAT slow (full format) .

[1713888258]

1713888258

[1713888258]

1713888258

[1713888258]

1713888258

[fullzero]

Hi fullzero,

great work on this distro.
My next rig will be based on a biostar tb250 btc pro with 12 PCIe Interfaces. Do you know if i can use 12 similar nvidia cards or will there be any limitations of the driver like in Windows (max. 8 GPUs of a kind)?

Cheers
tiefschwarz

I ordered one of those mobos as well; looking at the information from biostar it does seem like all the slots are in standard pcie buses so in theory it should be able to support 12 GPUs; getting PSUs for that is another issue (hard to find at the moment).  I will test with 12x when I get mine.

There is no artificial limit forced in linux; if you know how to modify the system: you are only limited by the hardware.

[fullzero]

Hi,

Trying out your OS, but first thing i'm encountering is when trying to change pools to european ones, the miner can't connect anymore (even with nanopool).
Second, when trying to use ethermine as a pool (with the switch turned to YES), it loops to read response failed end of file and cannot resolve hostname and read response failed end of file again, etc etc - no, it's not an internet or dns problem i assure, it works great.

Could you link a working onebash example with eu1.ethermine.org:4444 , so I can see what I did wrong?

are you using the Newest_oneBash linked on the OP?

I think you might be having a problem related to the way workers were being named when their host address is less than 100.

I fixed this in the Newest_oneBash.

also ensure

Code:

ETHERMINEdotORG="YES"

Let me know if this is the problem.

[salfter]

I will hold off on integrating this for now then (and wait for your changes); in the meantime I will make a link to your repo on the OP.

I've committed an update that, if it pans out, rolls everything into one Python script...no auxiliary shell scripts.  I'm testing it right now to verify that it behaves the same as the previous version.  I suspect I'll know in the morning.

Edit: Just did some accelerated testing by manually switching to a less-profitable coin first...the script killed the miner and fired up the appropriate miner.  I think the most recent update is ready for wider testing:

https://gitlab.com/salfter/nvoc-nicehash-switcher

[gyoztes]

Pls anybody explain me the difference between MANUAL_FAN="YES" + FAN_SPEED=75 and  MANUAL_FAN="NO". Stability? Consumption? Because I have used both of them and did no see difference in hash and consumption too.

I can not find the optimized setup for 1070. Do you have suggested setup? Or the newest oneBash is the suggested?

Thank you!

[bel123]

With this OS no problem to manage power with nvidias like under windows?

What tool for monitoring or managing such rigs can i use with restarts etc? (like temaviewer to manage rigs) Only ssh or any other options and maybe also auto monitoring/managing someone know?

[salfter]

Pls anybody explain me the difference between MANUAL_FAN="YES" + FAN_SPEED=75 and  MANUAL_FAN="NO". Stability? Consumption? Because I have used both of them and did no see difference in hash and consumption too.

They won't make any difference in hashrate, but automatic speed control will probably allow your GPUs to run warmer than a relatively high manual speed setting.  It allows you to trade more fan noise for cooler operating temperatures.

I can not find the optimized setup for 1070. Do you have suggested setup? Or the newest oneBash is the suggested?

I don't know that I'd consider it fully optimized, but you could use the settings in my auto-profitability switcher as a start (see the link a few posts ago).  The Equihash and DaggerHashimoto settings are probably the most tested at this point, as those have been the most profitable lately.

[cvitanihc0011]

Pls anybody explain me the difference between MANUAL_FAN="YES" + FAN_SPEED=75 and  MANUAL_FAN="NO". Stability? Consumption? Because I have used both of them and did no see difference in hash and consumption too.

I can not find the optimized setup for 1070. Do you have suggested setup? Or the newest oneBash is the suggested?

Thank you!

zec - pl 125 , cc 125 , mc 650 = > 460 sols . Rig running v solid - 8 GPU
eth - pl 85  , cc 150 , mc 1100= > 30.3mh  . Solid .. no crashes ..

You can try different settings and see how it behaves. I don't like to watch them all day for crashes .

[gigi8114]

Hi,

Trying out your OS, but first thing i'm encountering is when trying to change pools to european ones, the miner can't connect anymore (even with nanopool).
Second, when trying to use ethermine as a pool (with the switch turned to YES), it loops to read response failed end of file and cannot resolve hostname and read response failed end of file again, etc etc - no, it's not an internet or dns problem i assure, it works great.

Could you link a working onebash example with eu1.ethermine.org:4444 , so I can see what I did wrong?

are you using the Newest_oneBash linked on the OP?

I think you might be having a problem related to the way workers were being named when their host address is less than 100.

I fixed this in the Newest_oneBash.

also ensure

Code:

ETHERMINEdotORG="YES"

Let me know if this is the problem.

Hi,
I found the solution, i think it was because i tried to put a custom worker name - maybe you can include such a feature in the future?
Also, I couldn't find how I can see the current mining process. I did see the screen -r commands, but that implies killing the current process and restarting it. I'd like to be able to see, from SSH, the current mining process without killing it. Is this possible?

[lbrasi]

Hi,

Trying out your OS, but first thing i'm encountering is when trying to change pools to european ones, the miner can't connect anymore (even with nanopool).
Second, when trying to use ethermine as a pool (with the switch turned to YES), it loops to read response failed end of file and cannot resolve hostname and read response failed end of file again, etc etc - no, it's not an internet or dns problem i assure, it works great.

Could you link a working onebash example with eu1.ethermine.org:4444 , so I can see what I did wrong?

are you using the Newest_oneBash linked on the OP?

I think you might be having a problem related to the way workers were being named when their host address is less than 100.

I fixed this in the Newest_oneBash.

also ensure

Code:

ETHERMINEdotORG="YES"

Let me know if this is the problem.

Hi,
I found the solution, i think it was because i tried to put a custom worker name - maybe you can include such a feature in the future?
Also, I couldn't find how I can see the current mining process. I did see the screen -r commands, but that implies killing the current process and restarting it. I'd like to be able to see, from SSH, the current mining process without killing it. Is this possible?

I believe you must kill the process first. When the mining process starts from boot up, you can SSH in and type

Code:

ps aux | grep gnome-terminal  (find the PID for gnome-terminal-server)
kill PID
export DISPLAY=:0
screen -dmS rig1  (rig1 can be named whatever you like)
screen -x rig1
bash '/media/m1/1263-A96E/oneBash'

[S9k]

Hi,

Please help!
I have got stuck on this problems  
My configuration:

-ASUS PRIME Z270-P - 2 . I tried both, results are similar.
-EVGA GeForce GTX 1080 GAMING ACX 3.0 - 2
-MSI Geforce GTX 1080 Gaming X-  2
-The Gigabyte power supply unit on 1200 watts


Three video cards work perfectly in any any combinations,

m1@m1-desktop:~$ nvidia-smi -L
GPU 0: GeForce GTX 1080 (UUID: GPU-43453088-0fca-9442-106d-7594d157ebf2)
GPU 1: GeForce GTX 1080 (UUID: GPU-d099b67e-f204-66fa-96dc-365a6b559a7e)
GPU 2: GeForce GTX 1080 (UUID: GPU-5aacd4db-f68b-917e-8ac2-84caf68d6cac)
m1@m1-desktop:~$

m1@m1-desktop:~$ lspci |grep VGA
01:00.0 VGA compatible controller: NVIDIA Corporation Device 1b80 (rev a1)
03:00.0 VGA compatible controller: NVIDIA Corporation Device 1b80 (rev a1)
05:00.0 VGA compatible controller: NVIDIA Corporation Device 1b80 (rev a1)
m1@m1-desktop:~$


but if I add the fourth (in this case the ID GPU-5aacd4db-f68b-917e-8ac2-84caf68d6cac ), then the system falls. Here what I see in dmesg


[   98.722227] nvidia-modeset: Allocated GPU:0 (GPU-43453088-0fca-9442-106d-7594d157ebf2) @ PCI:0000:01:00.0
[   98.769072] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769117] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769144] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769169] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769193] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769217] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   98.769241] ACPI Warning: \_SB_.PCI0.RP04.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.359255] nvidia-modeset: Allocated GPU:1 (GPU-5c9c8e29-a088-90a6-2a20-b2b2b971d1fb) @ PCI:0000:05:00.0
[   99.398991] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399035] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399063] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399087] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399112] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399136] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buff er], ACPI requires [Package] (20150930/nsarguments-95)
[   99.399160] ACPI Warning: \_SB_.PCI0.RP05.PXSX._DSM: Argument #4 type mismatch - Found [Buffer], ACPI requires [Package] (20150930/nsarguments-95)
[   99.984670] nvidia-modeset: Allocated GPU:2 (GPU-5aacd4db-f68b-917e-8ac2-84caf68d6cac) @ PCI:0000:06:00.0
[  100.619118] nvidia-modeset: Allocated GPU:3 (GPU-d099b67e-f204-66fa-96dc-365a6b559a7e) @ PCI:0000:03:00.0
[  100.743159] NVRM: GPU at PCI:0000:01:00: GPU-43453088-0fca-9442-106d-7594d157ebf2
[  100.743162] NVRM: GPU Board Serial Number:
[  100.743164] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 000001e0 00000801 00000004 00000005
[  100.743649] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000080 00000004 00000005 00000004
[  102.432593] r8169 0000:07:00.0 enp7s0: link up
[  102.432600] IPv6: ADDRCONF(NETDEV_CHANGE): enp7s0: link becomes ready
[  103.743306] nvidia-modeset: WARNING: GPU:0: Lost display notification (0:0x00000000); continuing.
[  103.773941] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000080 00000000 00000005 00000004
[  105.501795] Bluetooth: BNEP (Ethernet Emulation) ver 1.3
[  105.501798] Bluetooth: BNEP filters: protocol multicast
[  105.501802] Bluetooth: BNEP socket layer initialized
[  105.613048] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000080 00000000 00000005 00000004
[  105.613106] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000080 00000000 00000005 00000004
[  105.704570] NVRM: Xid (PCI:0000:01:00): 56, CMDre 00000000 00000080 00000000 00000005 00000004
[  105.704972] BUG: unable to handle kernel paging request at ffff88167153d830
[  105.704974] IP: [<ffffffffc0262880>] _nv008171rm+0x620/0x780 [nvidia]
[  105.705052] PGD 220c067 PUD 0
[  105.705053] Oops: 0000 [#1] SMP

Three days I try to solve a problem.
I changed versions of BIOS (0325,0608,0610) and risers, control 4G is included, has updated NVIDIA drivers to 381.22 - nothing helps.
Maybe somebody will have ideas?

[tempgoga]

If anyone is interested my 7 card rig with all GTX1070 FE is finally stable running for the past 72 hours with the following clocks and pl:

__CORE_OVERCLOCK_0=000
MEMORY_OVERCLOCK_0=1400

__CORE_OVERCLOCK_1=000
MEMORY_OVERCLOCK_1=1400

__CORE_OVERCLOCK_2=000
MEMORY_OVERCLOCK_2=1300

__CORE_OVERCLOCK_3=000
MEMORY_OVERCLOCK_3=1600  

__CORE_OVERCLOCK_4=000
MEMORY_OVERCLOCK_4=1400

__CORE_OVERCLOCK_5=000
MEMORY_OVERCLOCK_5=1300

__CORE_OVERCLOCK_6=000
MEMORY_OVERCLOCK_6=1200


INDIVIDUAL_POWERLIMIT_0=125

INDIVIDUAL_POWERLIMIT_1=125

INDIVIDUAL_POWERLIMIT_2=120

INDIVIDUAL_POWERLIMIT_3=130

INDIVIDUAL_POWERLIMIT_4=125

INDIVIDUAL_POWERLIMIT_5=120

INDIVIDUAL_POWERLIMIT_6=120

i determined the power limit for the cards by their individual temperature and lowered/increased their clocks accordingly, if i give them all the same power limit GPU 3 is 10 degrees C cooler then GPU 2/5/6 while GPU 0/1/4 are right in the middle, at the current power limits all the cards are pretty much the same temp plus or minus 1 degree.

i might be able to push GPUs 1-5 further, i suspect GPU 6 was causing my crashes since the last crash when it was clocked @ 1300 it was the only GPU to lose its sensor connections, since lowering it to 1200 everything is stable so there might be room for higher clocks on the rest of the cards.

wish there was a way to determine which are samsung and which are micron..

[xleejohnx]

I will hold off on integrating this for now then (and wait for your changes); in the meantime I will make a link to your repo on the OP.

I've committed an update that, if it pans out, rolls everything into one Python script...no auxiliary shell scripts.  I'm testing it right now to verify that it behaves the same as the previous version.  I suspect I'll know in the morning.

Edit: Just did some accelerated testing by manually switching to a less-profitable coin first...the script killed the miner and fired up the appropriate miner.  I think the most recent update is ready for wider testing:

https://gitlab.com/salfter/nvoc-nicehash-switcher

If I could make my nvOS stable I would like to test this
I blame my biostar z170 board for all of my problems lol
Maybe tomorrow I can give it a go

[philipma1957]

I will hold off on integrating this for now then (and wait for your changes); in the meantime I will make a link to your repo on the OP.

I've committed an update that, if it pans out, rolls everything into one Python script...no auxiliary shell scripts.  I'm testing it right now to verify that it behaves the same as the previous version.  I suspect I'll know in the morning.

Edit: Just did some accelerated testing by manually switching to a less-profitable coin first...the script killed the miner and fired up the appropriate miner.  I think the most recent update is ready for wider testing:

https://gitlab.com/salfter/nvoc-nicehash-switcher

If I could make my nvOS stable I would like to test this
I blame my biostar z170 board for all of my problems lol
Maybe tomorrow I can give it a go

my biostar z170 is flawless up to 4 cards  after that not so good.

since I have six of them  I use  3 cards a board easy peasy that way.

[philipma1957]

I think you are using too low of a powerlimit and this is preventing higher OC from making a difference.

What powerlimit are you using?

When dual mining you need to use a higher powerlimit; ensure:

Code:

POWERLIMIT="YES"            

POWERLIMIT_WATTS=125

 or higher maybe up to 135 .  I don't have any ASUS GTX 1070 Turbos so I don't know what is OPT for them.

########

Interesting when I first set up rig 7x cards used 1150 watts.  

When I had PL set to 125  the power use dropped to 985w at the wall.

Now if I turn PL off or use 135 or 140 there is zero change at the wall it still uses 985w.

When terminal boots up I can see the PL come up but I have a theory that it is stuck at 125 and not changing.

Is that possible?

I have another rig that I just set up with 4 cards and it uses 785w regardless of the power setting.  On that rig the correct setting shows when terminal spools up but it is not limiting power. With 125 on each card I thought it should be 700 w max.

Any ideas on if PL can get 'sticky' ?

do not use the master setting  use individual.  for each card  and try
124
125
126
124
125
126

[DJ ACK]

Does anyone know of a way to detect GPU memory vendor information in Linux?  Similar to how GPU-Z can display the info in Winblows.  I want to group my mining rig clusters based on GPU memory vendors.  I think this will dramatically help in improving stability when tweaking each card in each rig. 

[fullzero]

I will hold off on integrating this for now then (and wait for your changes); in the meantime I will make a link to your repo on the OP.

I've committed an update that, if it pans out, rolls everything into one Python script...no auxiliary shell scripts.  I'm testing it right now to verify that it behaves the same as the previous version.  I suspect I'll know in the morning.

Edit: Just did some accelerated testing by manually switching to a less-profitable coin first...the script killed the miner and fired up the appropriate miner.  I think the most recent update is ready for wider testing:

https://gitlab.com/salfter/nvoc-nicehash-switcher

Looks good; I will probably change this py a little when I integrate it to make it read all the variables from oneBash.

[fullzero]

With this OS no problem to manage power with nvidias like under windows?

What tool for monitoring or managing such rigs can i use with restarts etc? (like temaviewer to manage rigs) Only ssh or any other options and maybe also auto monitoring/managing someone know?

You can use individual powerlimits for each GPU.

I am making a monitor / push / update / dashboard app now; keep getting sidetracked.

Right now you can use teamviewer, RDP, SSH in + screen; or any other third party solution that works with Ubuntu.

[fullzero]

Hi,

Trying out your OS, but first thing i'm encountering is when trying to change pools to european ones, the miner can't connect anymore (even with nanopool).
Second, when trying to use ethermine as a pool (with the switch turned to YES), it loops to read response failed end of file and cannot resolve hostname and read response failed end of file again, etc etc - no, it's not an internet or dns problem i assure, it works great.

Could you link a working onebash example with eu1.ethermine.org:4444 , so I can see what I did wrong?

are you using the Newest_oneBash linked on the OP?

I think you might be having a problem related to the way workers were being named when their host address is less than 100.

I fixed this in the Newest_oneBash.

also ensure

Code:

ETHERMINEdotORG="YES"

Let me know if this is the problem.

Hi,
I found the solution, i think it was because i tried to put a custom worker name - maybe you can include such a feature in the future?
Also, I couldn't find how I can see the current mining process. I did see the screen -r commands, but that implies killing the current process and restarting it. I'd like to be able to see, from SSH, the current mining process without killing it. Is this possible?

you should be able to use a custom workername so long as it doesn't include illegal characters

For SSH / screen info see:

https://bitcointalk.org/index.php?topic=1854250.msg19503300#msg19503300

[lbrasi]

I will hold off on integrating this for now then (and wait for your changes); in the meantime I will make a link to your repo on the OP.

I've committed an update that, if it pans out, rolls everything into one Python script...no auxiliary shell scripts.  I'm testing it right now to verify that it behaves the same as the previous version.  I suspect I'll know in the morning.

Edit: Just did some accelerated testing by manually switching to a less-profitable coin first...the script killed the miner and fired up the appropriate miner.  I think the most recent update is ready for wider testing:

https://gitlab.com/salfter/nvoc-nicehash-switcher

For some odd reason since you consolidated all the code to the switch.py, now the only miner that runs is equihash.