Docker: nsfminerOC (unsupported)


Recommended Posts

On 4/7/2021 at 5:29 PM, Rolucious said:

Tested it with a 3060ti and the HW transcoding works better.

The hashrates seem a bit lower than the t-rex miner, but considering no fee I think it's worth it. Thanks again for the docker.

 

Last question regarding the errors, sometimes I receive an error: GPU 0 gave incorrect result. Lower overclocking values if it happens to frequently.

I can't see the amount of shares and rejected shares, so I'm not sure what 'to frequently' is and if I really do need to lower the settings. I used to get 63,8 MH/s (+2900 mem) with my 3060ti and now I'm getting about 62 (+2650) with that error occuring once every 10 min or so. I think the fault could lie with the fans set to auto (I get an error when set to manual, so I recon it has 3 fan controllers for each of the 3 fans), but I'm not sure.

How often is okay to get the incorrect results message.

 

You should mainly not get incorrect at all. I think expected hashrates for 3060Ti is the same as 3070, so ~60-61. You rather have a bit lower hashrates and power consumption than incorrect results.

 

Fa controllers might vary, often 3 fan cards still uses only 2 controllers. There's commands to figure out that, but you might write them under "Console" in the docker container.

 

nvidia-settings -q fans

 

Link to comment

Thanks, I used the command and noticed I have 8 fans. So maybe I have to use fan ID 3/4 for GPU 1, FANID 5/6 for GPU2 etc.

Everything is working pretty well now, so I didn't bother adjusting it anymore.

 

I just upgraded to unraid 6.9.2 and noticed I got the fatal server error mentioned earlier.

I fixed it by adjusting something in the docker and saving it. After that it worked fine.

A reboot didn't seem to work, so if anyone notices something similar, it might be worth a shot.

Link to comment

Thanks for this docker, very nice. I had it run yesterday several hours with a single RTX3070 with ~59 MH. Then i shutdown the unraid server, moved it to another room and at startup i have now an error:

unable to connect to X server: Cannot assign requested address

 

it seems to have an impact to performance, as the same config has now only ~51MH. Any idea ?

Link to comment

Just try to restart the container. I have it autostarted with some wait time at 10 seconds. Maybe the drivers didn't get loaded or something. Right now your card is likely running without overclocking. If a restart doesn't help, try to reinstall it/force update. 

 

Then check in the docker log after startup if the correct nvidia drivers are downloaded and installed (must match the drivers installed in unraid)

Link to comment
26 minutes ago, olehj said:

Just try to restart the container. I have it autostarted with some wait time at 10 seconds. Maybe the drivers didn't get loaded or something. Right now your card is likely running without overclocking. If a restart doesn't help, try to reinstall it/force update. 

 

Then check in the docker log after startup if the correct nvidia drivers are downloaded and installed (must match the drivers installed in unraid)

Thanks, the driver check works and it downloads the matching driver.

I have just power and ethernet plugged in, but this was not different at the first location.

ErrorWarningSystemArrayLogin


---Trying to get Nvidia driver version---

---Successfully got driver version: 465.24.02---

---Checking Xwrapper.config---
---Configuring Xwrapper.config---
---Driver version missmatch, currently installed: v460.39, driver on Host: v465.24.02---

---Downloading and installing Nvidia Driver v465.24.02---

---Trying to get Nvidia driver version---

---Successfully got driver version: 465.24.02---

---Checking Xwrapper.config---
---Configuring Xwrapper.config---
---Driver version missmatch, currently installed: v460.39, driver on Host: v465.24.02---

---Downloading and installing Nvidia Driver v465.24.02---

---Output of nvidia-smi---
Thu Apr 15 13:16:40 2021
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 465.24.02 Driver Version: 465.24.02 CUDA Version: 11.3 |

|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. |
| | | MIG M. |
|===============================+======================+======================|
| 0 NVIDIA GeForce ... On | 00000000:01:00.0 Off | N/A |
| 0% 46C P8 16W / 135W | 0MiB / 7982MiB | 0% Default |
| | | N/A |
+-------------------------------+----------------------+----------------------+

+-----------------------------------------------------------------------------+
| Processes: |
| GPU GI CI PID Type Process name GPU Memory |
| ID ID Usage |
|=============================================================================|
| No running processes found |
+-----------------------------------------------------------------------------+
---Set persistent mode---
Persistence mode is already Enabled for GPU 00000000:01:00.0.

.
---Configuring xorg.conf---

WARNING: Unable to locate/open X configuration file.


Package xorg-server was not found in the pkg-config search path.
Perhaps you should add the directory containing `xorg-server.pc'
to the PKG_CONFIG_PATH environment variable
No package 'xorg-server' found
Option "AllowEmptyInitialConfiguration" "True" added to Screen "Screen0".
Option "AllowEmptyInitialConfiguration" "True" added to Screen "Screen0 (1)".
Option "AllowEmptyInitialConfiguration" "True" added to Screen "Screen0 (2)".
Option "AllowEmptyInitialConfiguration" "True" added to Screen "Screen0 (3)".
New X configuration file written to '/etc/X11/xorg.conf'

---Starting X server---


X.Org X Server 1.19.6
Release Date: 2017-12-20
X Protocol Version 11, Revision 0

Build Operating System: Linux 4.15.0-124-generic x86_64 Ubuntu
Current Operating System: Linux a972730727e2 5.10.28-Unraid #1 SMP Wed Apr 7 08:23:18 PDT 2021 x86_64
Kernel command line: BOOT_IMAGE=/bzimage initrd=/bzroot

Build Date: 30 November 2020 08:01:56PM
xorg-server 2:1.19.6-1ubuntu4.8 (For technical support please see http://www.ubuntu.com/support)
Current version of pixman: 0.34.0

Before reporting problems, check http://wiki.x.org
to make sure that you have the latest version.

Markers: (--) probed, (**) from config file, (==) default setting,
(++) from command line, (!!) notice, (II) informational,
(WW) warning, (EE) error, (NI) not implemented, (??) unknown.

(==) Log file: "/var/log/Xorg.0.log", Time: Thu Apr 15 13:16:51 2021
(==) Using config file: "/etc/X11/xorg.conf"
(==) Using system config directory "/usr/share/X11/xorg.conf.d"
vesa: Ignoring device with a bound kernel driver

(EE)
Fatal server error:

(EE) no screens found(EE)
(EE)
Please consult the The X.Org Foundation support
at http://wiki.x.org
for help.
(EE) Please also check the log file at "/var/log/Xorg.0.log" for additional information.
(EE)
(EE) Server terminated with error (1). Closing log file.

---Adjusting GPU values---
Power limit for GPU 00000000:01:00.0 was set to 135.00 W from 135.00 W.

.
Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help` for usage information.



Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help` for usage information.



Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help` for usage information.



Unable to init server: Could not connect: Connection refused

ERROR: The control display is undefined; please run `nvidia-settings --help` for usage information.



---Starting worker---
13:17:01 miner nsfminer 1.3.12 (No stinkin' fees edition)
13:17:01 miner Copyright 2021 Jean M. Cyr, Licensed under the terms
13:17:01 miner of the GNU General Public License Version 3

13:17:01 miner https://github.com/no-fee-ethereum-mining/nsfminer
13:17:01 miner Build: linux/release/gnu
13:17:01 miner 3rd Party: GCC 7.5.0, CUDA 11.2, Boost 1.75.0
13:17:01 miner 3rd Party: OpenSSL 1.1.1i, Ethash 0.5.0
13:17:01 miner Configured pool eth.2miners.com:2020
13:17:01 miner Configured pool eth.2miners.com:2020
13:17:01 miner Selected pool eth.2miners.com:2020
13:17:01 miner Stratum mode : Stratum
13:17:01 miner Established connection to eth.2miners.com:2020
13:17:01 miner Spinning up miners...
13:17:01 cu-0 Using Pci 0000:01:00.0: NVIDIA GeForce RTX 3070 (Compute 8.6) Memory : 7.80 GB
13:17:01 miner Authorized worker 0x??????????????????.rtx3070
13:17:01 miner Epoch : 408 Difficulty : 8.73 Gh
13:17:01 miner Job: 904999b5 eth.2miners.com:2020
xinit: giving up
xinit: unable to connect to X server: Cannot assign requested address
xinit: server error

xinit: giving up
xinit: unable to connect to X server: Cannot assign requested address
xinit: server error

13:17:06 miner 0:00 A0 0.00 h - cu0 0.00 46C 0% 16.30W
13:17:07 cu-0 Required GPU mem: Total 4.25 GB, Cache 67.00 MB, DAG 4.19 GB, Miscellaneous 56.00 B
13:17:08 miner Job: d90d0acd eth.2miners.com:2020
13:17:08 miner Job: d90d0acd eth.2miners.com:2020
13:17:11 miner 0:00 A0 0.00 h - cu0 0.00 54C 0% 134.78W
13:17:15 miner Job: 862c2da2 eth.2miners.com:2020
13:17:16 miner 0:00 A0 0.00 h - cu0 0.00 56C 0% 134.59W
13:17:19 cu-0 4.19 GB of DAG data generated in 12.1 seconds
13:17:21 miner 0:00 A0 14.75 Kh - cu0 14.75 58C 0% 134.84W
13:17:22 miner Job: ce0400e7 eth.2miners.com:2020
13:17:22 miner Job: ce0400e7 eth.2miners.com:2020
13:17:26 miner 0:00 A0 51.55 Mh - cu0 51.55 59C 0% 134.79W
13:17:26 miner 0:00 A0 51.55 Mh - cu0 51.55 59C 0% 134.79W
13:17:29 miner Job: f34ef082 eth.2miners.com:2020
13:17:31 miner 0:00 A0 51.54 Mh - cu0 51.54 60C 0% 134.87W
13:17:36 miner Job: 20685c30 eth.2miners.com:2020
13:17:36 miner 0:00 A0 51.53 Mh - cu0 51.53 60C 0% 134.88W
13:17:41 miner 0:00 A0 51.54 Mh - cu0 51.54 60C 50% 134.66W
13:17:43 miner Job: d04fcd12 eth.2miners.com:2020
13:17:44 miner Job: 173a1c51 eth.2miners.com:2020
13:17:46 miner 0:00 A0 51.55 Mh - cu0 51.55 61C 50% 134.65W

 

Link to comment

No, and unraid is fresh from last week, nothing installed yet, beside of fileshare.

i just deleted nvidia driver and the docker and reinstalled ... tried different driver versions between. Nothing changed.

 

how can i check the gpu frequency and memory frequency, check that the settings are applied ? sorry, just mining on windows the last weeks ...

Link to comment
2 minutes ago, Grobalt said:

I have an idea but i am not able to test tonight - is it possible that the overclocking settings are not applied as i have no more a display connected ?

It should not matter, I have 3 GPU's running without display, the 4th GPU is running VM with mining inside it as I use it for multiple things.

 

The configuration should ignore connected screens and create fake screens. But you can try with one connected, maybe a HDMI-dummy will fix the problem in that case.

Link to comment

i created a windows 10 VM. was a bit difficult to get that running well with 6.9.2 - always crashed the whole unraid server during nvidia driver installation. setup.exe -s was succesful, a silent driver installation.

then with afterburner settings applied i was able to mine with nearly 64 MH/s which is quite fast for a RTX3070 🙂

 

 

 

 

rtx3070.JPG

Link to comment
8 minutes ago, Grobalt said:

i created a windows 10 VM. was a bit difficult to get that running well with 6.9.2 - always crashed the whole unraid server during nvidia driver installation. setup.exe -s was succesful, a silent driver installation.

then with afterburner settings applied i was able to mine with nearly 64 MH/s which is quite fast for a RTX3070 🙂

 

 

 

 

rtx3070.JPG

Might work fine with one GPU, i didn't get any luck by trying to pass through more than one to a VM. Dunno why, but docker worked for me 😛

Link to comment
6 hours ago, truthfulie said:

Been able to get 60mh with 3070 just on 130watt through Windows. Any idea why the container needs extra 20w to get the same hashrate?

I have no idea, different drivers perhaps (windows and Linux drivers might behave differently for all i know). I haven't tried it myself, but ensure that the gpu and memory clock is exactly the same (the input is different from windows and Linux drivers)

Link to comment
4 hours ago, olehj said:

I have no idea, different drivers perhaps (windows and Linux drivers might behave differently for all i know). I haven't tried it myself, but ensure that the gpu and memory clock is exactly the same (the input is different from windows and Linux drivers)

And also, do you use the same tool to check the actual power usage, nvidia-smi? Some 3rd party tools might show different numbers, I dunno.. just a thought.

Link to comment

Interesting. I have not checked the power usage with the same tool. I've checked with nvidia-smi with unraid and I've only tested nicehash's own miner's (when the card was plugged into a windows machine) reporting which was kind of nice. It gives wattage and efficiency rating based on wattage vs hashrate. Anyway now I mine ETH directly with a pool. 20w ain't nothing but it's also not big enough difference in power bill for me to worry about it too much. I'll fiddle with OC settings a bit more.

 

PS. which specific 3070 model did you write those settings? I am running FE model.

Link to comment
1 minute ago, truthfulie said:

Interesting. I have not checked the power usage with the same tool. I've checked with nvidia-smi with unraid and I've only tested nicehash's own miner's (when the card was plugged into a windows machine) reporting which was kind of nice. It gives wattage and efficiency rating based on wattage vs hashrate. Anyway now I mine ETH directly with a pool. 20w ain't nothing but it's also not big enough difference in power bill for me to worry about it too much. I'll fiddle with OC settings a bit more.

 

PS. which specific 3070 model did you write those settings? I am running FE model.

 

I have a RTX 3070 Gigabyte Gaming OC 8GB, but it sits inside a relatively hot server, so I expect some hashes might be lost just there.

 

But it's important to compare with the same tools, maybe nicehash calculates hasrates differently as well. A fair comparison would be using nsfminer for windows and the one in linux/docker container, and then use nvidia-smi on both windows and linux to determine power usage (maybe also nvidia-smi shows the same clocks as well across platforms?).

Link to comment

Just an update. Been playing around with the settings a bit over the weekend and I managed to get the 3070 FE to 60.45MH with lower power usage.

 

Settings

pl: 135

clock offset: -503

memory offset: 2300

fan: 70

 

130w can get to 60 but was struggling to keep it locked at 60. Would dip to 59 here and there. So I just gave it 5 more watts. More efficient than 150 I was using before.

Link to comment
Posted (edited)
On 4/26/2021 at 5:07 AM, truthfulie said:

Just an update. Been playing around with the settings a bit over the weekend and I managed to get the 3070 FE to 60.45MH with lower power usage.

 

Settings

pl: 135

clock offset: -503

memory offset: 2300

fan: 70

 

130w can get to 60 but was struggling to keep it locked at 60. Would dip to 59 here and there. So I just gave it 5 more watts. More efficient than 150 I was using before.

Nice, decided to fiddle around with your numbers.. and should have done a bit more fiddling earlier.. ended up with these results on the RTX 3070:

  • Power limit: 120W (seems to be the minimum allowed for my card)
  • Clock offset: -550
  • Memory offset: 2300
  • Fan: auto

Hashrate: 60.2 MH/s - so apparently very effective!

 

Edited by olehj
Link to comment

I updated the miner, but the logs state that the driver version could no be found.

Log:
---Trying to get Nvidia driver version---
---Something went wrong, can't get driver version, putting server into sleep mode---

 

Do I need to update all the drivers to the latest version? Currently running 460.67

Link to comment
9 minutes ago, Rolucious said:

I updated the miner, but the logs state that the driver version could no be found.

Log:
---Trying to get Nvidia driver version---
---Something went wrong, can't get driver version, putting server into sleep mode---

 

Do I need to update all the drivers to the latest version? Currently running 460.67

It should auto detect and install correct drivers. Maybe the location it download drivers from are down for maintenance or something. I can't check it now, but it worked for me yesterday. 

 

Link to comment
1 hour ago, olehj said:

It should auto detect and install correct drivers. Maybe the location it download drivers from are down for maintenance or something. I can't check it now, but it worked for me yesterday. 

 

I updated the driver in the nvidia pluging to 465.27, but that caused nvidia-smi to not properly report the GPU's for some reason, resulting that the cards weren't able to overclock.

In the end the only nvidia driver I found working was 460.73.

 

It's currently working correctly

Link to comment
On 5/2/2021 at 5:56 PM, olehj said:

Nice, decided to fiddle around with your numbers.. and should have done a bit more fiddling earlier.. ended up with these results on the RTX 3070:

  • Power limit: 120W (seems to be the minimum allowed for my card)
  • Clock offset: -550
  • Memory offset: 2300
  • Fan: auto

Hashrate: 60.2 MH/s - so apparently very effective!

 

 

Tried your settings and I'm only getting 50 MH/s. I need to give it more power. Managed to get 60.3 MH/s with 120W (like I was originally getting on Windows) but it wouldn't hold. I am now back to 135W power limit and getting consistent hashrate of 60.4. Probably because mine is FE not AIB card.

 

In any case, I am having the same issue as Rolucious. Updated the docker and the miner would not work. I was on older v460.67 driver (just because I was reading the older post on this thread when I first set it up) Updated to the latest driver to see if that was the cause, it wasn't. Set it to 460.73.01 like Rolucious said and it is up and running. No real benefit for me to be on the latest driver so I'll just use it but might be worth a look.

Link to comment
  • olehj changed the title to Docker: nsfminerOC (unsupported)

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.