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

[fullzero]

Thanks for your effort FullZero love your work.

I tried your NVOC on my 13 card 1070 rig, but an into a few issues.

- I couldn't get all 13 to be recognised. claymore continues to report only 6 GPUs (mobo is ASROCK H110 BTC)
- I tried to set powerlimit both global and individual to 105W, and it only holds the 105W limit for about 3-5 seconds, after that it quickly creeps up to 125W and holds at 125W. I have tried with the maximus script on and off no difference.

Anyone have any issues like this?

How are you powering this rig?  Have you joined the PSUs (using an adapter to connect them) if you are using multiple ones?  If not the system not recognizing all your GPUs is most likely related to this.


With the powerlimit:

Most likely you aren't setting: 

# Note you MUST set individual powerlimits if using Maxximus007_AUTO_TEMPERATURE_CONTROL

INDIVIDUAL_POWERLIMIT_0=75

INDIVIDUAL_POWERLIMIT_1=75

INDIVIDUAL_POWERLIMIT_2=75

INDIVIDUAL_POWERLIMIT_3=75

INDIVIDUAL_POWERLIMIT_4=75

INDIVIDUAL_POWERLIMIT_5=75

INDIVIDUAL_POWERLIMIT_6=75

INDIVIDUAL_POWERLIMIT_7=75

INDIVIDUAL_POWERLIMIT_8=75

INDIVIDUAL_POWERLIMIT_9=75

INDIVIDUAL_POWERLIMIT_10=75

INDIVIDUAL_POWERLIMIT_11=75

INDIVIDUAL_POWERLIMIT_12=75

INDIVIDUAL_POWERLIMIT_13=75

Also I recommend updating your 1bash and the other files with the ones from the bux fix download on the Original Post.  I know you aren't using the newest 1bash because your individual powerlimit is setting to 125 and not 75.

[fullzero]

Hi, guys!

I have a problem. My rig is 8 Nvidia 1080Ti on motherboard MSI Z170 M5, works on win 10 with nicehash ok all cards working, but windows is eating too much and time to time there are problems exactly with system. Anyway, I decided to try nvOC. Used to play with Linux in my childhood and respect it for stability and convenience.

With nvOC I have few issues: 1. It boots too long I think. 3-5 minutes on my rig.

2. It reboots every 5 minutes.

3. I'm trying to mine LBC, but every time it connects I get the message: Stratum identification failed.


What am I doing wrong? What to look at?

Thanks in advance. I enjoy using nvOC even for short time ))

Booting will take longer than Windows; eventually I will make a lite build that will boot faster: but this is not currently a priority vs other more useful improvements.

This is what is most likely happening:

You are not connecting successfully to the pool because you are using a pool where the workername must be predefined at the pool to be authorized.

Because it is not; when you try to connect the pool rejects the connection.  

This causes your GPU utilization to be below 90%

This causes the watchdog to restart 1bash; and eventually reboot your rig.


Go to your pool / login and make a workername, then change the workername on your rig to match.

Also if you are rebooting every 5 mins; You haven't downloaded the newest 1bash and files ( bug fix download link on the original post).


Let me know if this is; what is, in fact the case: or if the problem persists.

[fullzero]

So interesting (scary) discovery last night. I built a brand new 7 GPU rig just like my other ones I have ...identical components. Brand new cards from newegg GTX 1060 3GB SC version. -I like these cards so I stick with them. Anyway my other rigs are using both the 1060 SC and regular 1060 (non SC) version without issue. The SC just run cooler but hash the same as their non SC counterpart. So my normal hashrates on my existing rigs is 23MH/s and I am quite pleased with that.

I did the very few bios updates on my Asus Z270A motherboard and slapped a pre-built nvOC USB Stick in ...away she went. However this is where it got bad. I was only getting about 18MH/s with the same exact OC ratings as my other rigs. I was getting lag and ERRORS as well. So OK ...thought maybe I have a bad riser card. I start disconnecting and removing everything to start with the first card only. Just 1 card in the first x16 slot closest to the CPU as it should be. Same thing.

Still getting 18MH/s and some hanging. So I change that riser card and cable out completely. Still the same thing. I reduce the OC to 0/0 and my MH/s drops to about 15MH/s now. But the system seems to be stable.

I'm really scratching my head on this guys. I futzed with this for over 3 hours last night trying all sorts of things. I'm starting to wonder if the Nvidia driver install doesn't update the bios on the video cards somehow. So the fact that the driver was already installed on this image, the cards never got "flashed"? My other rigs were part of windows builds from the ground up and cards were in the rig when I updated the video drivers. So if my theory is correct those GPU's got flashed at that time and hence they are fine with nvOC0018.

But this new rig only ever booted into a fully built nvOC0018. -Anyone run into this themselves? -I have another motherboard so I'm going to try that on the off-chance that I have a bad motherboard somehow? Help and suggestions welcome. Other similar experiences would be great.

I don't think nvidia automatically updates bios firmware.  This would cause a ton of GPUs to be bricked, by users losing power / hard shutting down their computers while the bios updated.

Most likely you have lost the silicon lottery this time around:

The earlier GPUs most likely have better memory than the new ones.  I would guess the new ones have micron memory; and thus due to the combination of these GPUs having only 3gb and significantly worse memory they hash significantly lower.

You can see if the GPUs are of a different build model by using:

Code:

sudo lspci -v

on one of your older rigs and then the new one and comparing.

[sertix_T]

is not launching teamviewer please let me know. 
 

Yes, it works! And I'm an old man blind! Did not notice this setting! Went the other way! I do not know much English and worked too long under Windows...

[DJ ACK]

Not really v0018 related.  However, interesting to learn.  Received a batch of Gigabyte 1050Ti 1506 Boost Clock edition cards from Amazon last week.  Installed them in a rig and they hashed at an insanely low rate (single digits).  I swapped out risers, swapped cards between rigs, and they just would not hash.  You can tell these cards have issues because the DAG file % complete lags way behind the other cards during the DAG file build.  Running the card as a stand alone the DAG file builds quickly and the card hashes fine.  The minute you add it with another card the hash rate tanks.  Even with the same manufactured batch (serial number off by a couple digits).  After comparing the serial number to the 48 other Gigabyte 1050Ti cards I have, I noticed that the serial number is a very low/early edition number compared to all the rest of my cards.  My hunch is that these cards were sitting in a warehouse corner for years and they were a failed production batch with crap memory.  Sent them back to Amazon as defective.  

Anyone else had crap card inventory show up since May/June time frame?  

[fogcity]

Hey there --

I'm using v0018 with 6 gtx 1070s and an MSI Z-270A PRO to mine eth and all is working just fine.

I've searched the forum but can't seem to find an answer to what is probably a simple question.

Where do I assign my worker email address for my pool?  I'm using genoil (not Claymore) and nanopool. Perhaps it should be in the ETH_EXTENTION_ARGUMENT?  I would really appreciate it someone could tell me where it goes and provide me an example of how to do so.

Regards!!!

[alko67bi]

I have NV 1060. How I can change Performance Level from 2 to 3 in Nvidia X server settings?

[fullzero]

Hey there --

I'm using v0018 with 6 gtx 1070s and an MSI Z-270A PRO to mine eth and all is working just fine.

I've searched the forum but can't seem to find an answer to what is probably a simple question.

Where do I assign my worker email address for my pool?  I'm using genoil (not Claymore) and nanopool. Perhaps it should be in the ETH_EXTENTION_ARGUMENT?  I would really appreciate it someone could tell me where it goes and provide me an example of how to do so.

Regards!!!

For nanopool:

append / followed by your email address to the end of the worker; for example if my worker is nvOC and my email is bob@aol.com, I would use:

Code:

ETH_WORKER="nvOC/bob@aol.com"

[fullzero]

I have NV 1060. How I can change Performance Level from 2 to 3 in Nvidia X server settings?

in 1bash set:

Code:

GPUPowerMizerMode_Adjust="YES"

GPUPowerMizerMode=3

or open a new guake tab and enter:

Code:

sudo nvidia-settings -a [gpu:0]/GPUPowerMizerMode=3

changing the number in (gpu:0) for the GPU you are adjusting.

[fogcity]

Hey there --

I'm using v0018 with 6 gtx 1070s and an MSI Z-270A PRO to mine eth and all is working just fine.

I've searched the forum but can't seem to find an answer to what is probably a simple question.

Where do I assign my worker email address for my pool?  I'm using genoil (not Claymore) and nanopool. Perhaps it should be in the ETH_EXTENTION_ARGUMENT?  I would really appreciate it someone could tell me where it goes and provide me an example of how to do so.

Regards!!!

For nanopool:

append / followed by your email address to the end of the worker; for example if my worker is nvOC and my email is bob@aol.com, I would use:

Code:

ETH_WORKER="nvOC/bob@aol.com"

Full zero --

Thanks for the prompt reply!

So if I understand correctly, and if my email is bob@aol.com, I need to go into my 1bash file to where it's collecting my eth info (right beneath where you choose to use genoil or claymore) and replace the line that says:

     ETH_WORKER="nv$IP_AS_WORKER"

with

     ETH_WORKER="nv$IP_AS_WORKER/bob@aol.com"

Is that correct?

Thanks in advance for the help-- keep up the great work !!!

[DJ ACK]

Witnessed an odd issue with v0018 today.  All 8 of my 1050Tis only report ~13 MH/s.   Thought it was risers and swapped a couple out, no change.  Re-imaged the SSD and upon initial boot up, everything was fine, reporting 14.7 MH/s.  As soon as I rebooted it dropped the hash back down to 13 MH/s.  Swapped out cards and booted with just a single card same result.   Re-imaged again, same result.  Any idea what would cause this?  

I am running ASUS Z270 Prime-A boards with Gigabyte 1050Ti Boost Clock 1506 cards.  I don't think it is an ETH difficulty issue as my other rigs have the same configuration and I am not seeing the same issue.  My power is set at 60watts, temp settings at 60, my fan setting 50, core OC is 0 and mem OC is 1450.  I changed around power and mem settings, but no difference.

Could one of the modules be adjusting this to match one of the above variables?  My cards never get hot enough for the fan to go above 50.

Autotemp could be lowering the powerlimit; but only after your fans go to max.  If they don't go above 50% then that is unlikely.  The current powerlimit is displayed by the autotemp readout, so you can verify if it has changed when this is the case.

If it has been; I would raise your temp target to 70.

Watchdog only restarts 1bash or the rig; it shouldn't make an OC changes.

What kind of risers are you using? Are you powering them direct with Pcie 6pin?  If not this may be the problem.

Also you can probably get the same MH/s with the minimum powerlimit and -100 or -200 cc.

Well amazing.. that was it.. bumped the temp target up to 70, dropped the cc down to -200 and I found my missing hash. Now to tweak my other rigs.  Looks like you help me find another 65 MH/s for my entire setup.  Thank you very much.

That one rig has mixed molex and 6 pin type risers.  I will change out the molex when my new stock of risers comes in.

[rayd89]

Thanks for your effort FullZero love your work.

I tried your NVOC on my 13 card 1070 rig, but an into a few issues.

- I couldn't get all 13 to be recognised. claymore continues to report only 6 GPUs (mobo is ASROCK H110 BTC)
- I tried to set powerlimit both global and individual to 105W, and it only holds the 105W limit for about 3-5 seconds, after that it quickly creeps up to 125W and holds at 125W. I have tried with the maximus script on and off no difference.

Anyone have any issues like this?

How are you powering this rig?  Have you joined the PSUs (using an adapter to connect them) if you are using multiple ones?  If not the system not recognizing all your GPUs is most likely related to this.


With the powerlimit:

Most likely you aren't setting:  

# Note you MUST set individual powerlimits if using Maxximus007_AUTO_TEMPERATURE_CONTROL

INDIVIDUAL_POWERLIMIT_0=75

INDIVIDUAL_POWERLIMIT_1=75

INDIVIDUAL_POWERLIMIT_2=75

INDIVIDUAL_POWERLIMIT_3=75

INDIVIDUAL_POWERLIMIT_4=75

INDIVIDUAL_POWERLIMIT_5=75

INDIVIDUAL_POWERLIMIT_6=75

INDIVIDUAL_POWERLIMIT_7=75

INDIVIDUAL_POWERLIMIT_8=75

INDIVIDUAL_POWERLIMIT_9=75

INDIVIDUAL_POWERLIMIT_10=75

INDIVIDUAL_POWERLIMIT_11=75

INDIVIDUAL_POWERLIMIT_12=75

INDIVIDUAL_POWERLIMIT_13=75

Also I recommend updating your 1bash and the other files with the ones from the bux fix download on the Original Post.  I know you aren't using the newest 1bash because your individual powerlimit is setting to 125 and not 75.

Thanks for your swift reply!

I am connecting two PSU's together using a Add2PSU adaptor. Windows is able to see all 13 (but only hash with 8 )
Ubuntu is only seeing 6

I had changed all the individual settings to 105 before but i have not tried the latest 1bash file, i will give that a shot soon and return back with results! hopefully it fixes it

Thanks

[alko67bi]

I have NV 1060. How I can change Performance Level from 2 to 3 in Nvidia X server settings?

in 1bash set:

Code:

GPUPowerMizerMode_Adjust="YES"

GPUPowerMizerMode=3

or open a new guake tab and enter:

Code:

sudo nvidia-settings -a [gpu:0]/GPUPowerMizerMode=3

changing the number in (gpu:0) for the GPU you are adjusting.

"Valid values for 'GPUPowerMizerMode' are: 0, 1 and 2."
But in Nvidia X server settings displayed 0, 1, 2 and 3

[darkfortedx]

my rigs keep restarting after few minutes. Some after 5 mins some after 30. Im not sure why this keeps happening.

im using all 1070's with core 100 / 1050 also i tried 150-200 / 1100-1600. No luck

[ducksauce88]

I'm trying to make changes to my xorg.conf, and when I do after reboot the file is always wiped. Why does this file keep getting overwritten after reboot?

[Nexillus]

my rigs keep restarting after few minutes. Some after 5 mins some after 30. Im not sure why this keeps happening.

im using all 1070's with core 100 / 1050 also i tried 150-200 / 1100-1600. No luck

what is your power limit on them? for 1070s I would push around 120-125 with a moderate/heavy OC. Anything less can cause stability problems.

[darkfortedx]

my rigs keep restarting after few minutes. Some after 5 mins some after 30. Im not sure why this keeps happening.

im using all 1070's with core 100 / 1050 also i tried 150-200 / 1100-1600. No luck

what is your power limit on them? for 1070s I would push around 120-125 with a moderate/heavy OC. Anything less can cause stability problems.

my power limits are all at 125

[salfter]

I've updated my switcher to separate out the configuration from the rest of the code:

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

Configuration now resides in a JSON file that is easily edited, whether manually or automatically.  I'm looking at the latest 1bash to have it generate this file at runtime.

I also think I've found and eliminated a source of instability in my mining rig:



After moving from a tower case to a Spotswood frame, I noticed that the power cable feeding one of the GPUs was noticeably warm, and so were the Molex connectors on that cable that plugged into the adapter shown above.  I suspect the wiring on the Molex-connector cable on this power supply isn't as beefy as the others, on the theory that it'll only have some fans and (very rarely) a floppy drive hanging off of it, not a hard drive (since those are all SATA now) or anything more power-hungry.  daggerhashimoto was the only algo running with any stability, and a voltmeter on one of the unused Molex connectors showed a drop on the +12V line down to 11.3V.  Nearly any other algo would cause the GPU to drop off the bus.

I picked up different adapters at Fry's this afternoon that go from two SATA connectors to one PCIe connector.  (I already had one of these in use powering another GPU with an 8-pin connector.  I bought two more for the 8- and 6-pin connectors on the problematic GPU.)  I swapped them in when I got home, fired up the rig, and saw that the drop on +12V had improved to 11.65V, which I think is in spec.  Not only that, but I benchmarked all algos supported by the switcher with the new three-GPU configuration, and not once did I have to reboot.

Now I just need my PCIe risers to arrive...and maybe also to replace the 650W ATX power supply I'm currently using with one of those server power supplies with an adapter that provides lots of PCIe outputs, as this one's nearly maxed out for connectivity.  (Three 1070s and an Asus Prime Z270-AR also draw about 400W with the cards underclocked.)  I remember seeing one company bundling a picoPSU with theirs so you could run the motherboard off a PCIe cable.

[dbolivar]

I've updated my switcher to separate out the configuration from the rest of the code:

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

Configuration now resides in a JSON file that is easily edited, whether manually or automatically.  I'm looking at the latest 1bash to have it generate this file at runtime.

Hey salfter / Scott, is it possible to add individual OC settings to the JSON file? For instance:

Original:
"power_limit": [115,115,95], "gpu_oc": -200, "mem_oc": 600, "fan": [70,75,70]

Individual OC:
"power_limit": [115,115,95], "gpu_oc": [-200,0,50], "mem_oc": [600,1000,800], "fan": [70,75,70]

From what I could see in the "for i in range(0, cards)" loop in switch.py, you already implemented that. So, just checking if my understanding is correct -- otherwise, I'd like to suggest it as a new feature.

And thanks for your great work (very elegant approach with the JSON file, btw).

EDIT: ooops, checking the SIA example in the README, it seems my question is already answered, lol.

[salfter]

Hey salfter / Scott, is it possible to add individual OC settings to the JSON file? For instance:

Original:
"power_limit": [115,115,95], "gpu_oc": -200, "mem_oc": 600, "fan": [70,75,70]

Individual OC:
"power_limit": [115,115,95], "gpu_oc": [-200,0,50], "mem_oc": [600,1000,800], "fan": [70,75,70]

From what I could see in the "for i in range(0, cards)" loop in switch.py, you already implemented that. So, just checking if my understanding is correct -- otherwise, I'd like to suggest it as a new feature.

That's exactly how it works.  I left in examples configured both ways.