PTRFRLL Posted December 2, 2021 Author Share Posted December 2, 2021 (edited) @repomanz I've got the container updated to use a new CUDA version. ptrfrll/nv-docker-trex:latest Also, as @Rhomax pointed out, make sure you don't have the 'disable NVML' option set in your config Edited December 2, 2021 by PTRFRLL Quote Link to comment
letrain Posted December 2, 2021 Share Posted December 2, 2021 (edited) So I see in the settings now there is lots of nice options. I see gpu core clock, is there a way to adjust the memory? I'd much rather use this lovely docker instead of a windows vm... Edited December 3, 2021 by letrain Quote Link to comment
eagle470 Posted January 6, 2022 Share Posted January 6, 2022 I have this happily crunching away, but I'm seeing rather poor performance. I'm 10 MH/s under what I should be getting. the only thing I've done is limit the power level to 72w as that is as low as I can get it without impacting performance. I'd love to OC the card, but I co-use the card with Unmanic and Plex. 95% of the time T-Rex Miner is the ONLY thing using the card. I shut off T-Rex Miner and UnManic during peak viewing hours (4PM to midnight) hours every night. Quote Link to comment
Bolagnaise Posted January 6, 2022 Share Posted January 6, 2022 6 hours ago, eagle470 said: I shut off T-Rex Miner and UnManic during peak viewing hours (4PM to midnight) hours every night. How are you doing this? Via User scripts? Quote Link to comment
ICDeadPpl Posted January 8, 2022 Share Posted January 8, 2022 On 1/6/2022 at 11:34 PM, Bolagnaise said: How are you doing this? Via User scripts? I use this script, which checks if Plex has started transcoding and stops Trex if that is true. It starts Trex after Plex has finished transcoding. It is for Nvidia cards only. Feel free to use it and modify it to your needs. #!/bin/bash # Check if nvidia-smi daemon is running and start it if not. if [[ `ps ax | grep nvidia-smi | grep daemon` == "" ]]; then /usr/bin/nvidia-smi daemon fi sleep 300 # Wait until array is online and all dockers are started. Comment this out if you are testing the script. TREX=`docker container ls -q --filter name=trex*` while true; do if [[ `/usr/bin/nvidia-smi | grep Plex` == "" ]]; then # If Plex is not transcoding, start the trex-miner container if it is not running. if [[ `docker ps | grep $TREX` == "" ]]; then echo "No Plex, starting Trex." docker start $TREX fi else # If Plex is transcoding, stop the trex-miner container if it is running. if [[ `docker ps | grep $TREX` != "" ]]; then echo "Plex running, stopping Trex." docker stop $TREX fi fi sleep 1 done 2 2 Quote Link to comment
eagle470 Posted January 8, 2022 Share Posted January 8, 2022 On 1/6/2022 at 4:34 PM, Bolagnaise said: How are you doing this? Via User scripts? Yes, you might also look at an LHR card as that would leave some processing power for transcodes and still allow you to mine. Its not full efficiency, but it would help a lot. Quote Link to comment
eagle470 Posted January 8, 2022 Share Posted January 8, 2022 (edited) 8 hours ago, ICDeadPpl said: I use this script, which checks if Plex has started transcoding and stops Trex if that is true. It starts Trex after Plex has finished transcoding. It is for Nvidia cards only. Feel free to use it and modify it to your needs. #!/bin/bash # Check if nvidia-smi daemon is running and start it if not. if [[ `ps ax | grep nvidia-smi | grep daemon` == "" ]]; then /usr/bin/nvidia-smi daemon fi sleep 300 # Wait until array is online and all dockers are started. Comment this out if you are testing the script. TREX=`docker container ls -q --filter name=trex*` while true; do if [[ `/usr/bin/nvidia-smi | grep Plex` == "" ]]; then # If Plex is not transcoding, start the trex-miner container if it is not running. if [[ `docker ps | grep $TREX` == "" ]]; then echo "No Plex, starting Trex." docker start $TREX fi else # If Plex is transcoding, stop the trex-miner container if it is running. if [[ `docker ps | grep $TREX` != "" ]]; then echo "Plex running, stopping Trex." docker stop $TREX fi fi sleep 1 done How often do you run this? do you use a watchdog or a cron job? Edited January 8, 2022 by eagle470 Quote Link to comment
ICDeadPpl Posted January 8, 2022 Share Posted January 8, 2022 3 hours ago, eagle470 said: How often do you run this? do you use a watchdog or a cron job? I have set it to run at first array start only. The script itself runs in a infinite loop and has a 1 second long pause between checks (the "sleep 1" part at the end). 1 Quote Link to comment
eagle470 Posted January 9, 2022 Share Posted January 9, 2022 On 1/8/2022 at 1:35 PM, ICDeadPpl said: I have set it to run at first array start only. The script itself runs in a infinite loop and has a 1 second long pause between checks (the "sleep 1" part at the end). thanks, I cloned it and modded it to work for unmanic as well. super helpful! Quote Link to comment
eagle470 Posted January 10, 2022 Share Posted January 10, 2022 @PTRFRLL would you be interested in adding the following libraries to the package? it would enable people to use the following script inside the container. You could also add the script to allow for limited OC capabilities: https://askubuntu.com/questions/1340946/nvidia-rtx-3070-eth-overclock-in-ubuntu20-04-nvidia-settings-a-gpugraphicscloc Quote Link to comment
PTRFRLL Posted January 10, 2022 Author Share Posted January 10, 2022 3 hours ago, eagle470 said: @PTRFRLL would you be interested in adding the following libraries to the package? it would enable people to use the following script inside the container. You could also add the script to allow for limited OC capabilities: https://askubuntu.com/questions/1340946/nvidia-rtx-3070-eth-overclock-in-ubuntu20-04-nvidia-settings-a-gpugraphicscloc I have not found a good way to include nvidia-settings and X11 into the container without requiring a specific version of nvidia drivers. I hesitate to do that as it would cause compatibility issues with many who use this container. I'm open to suggestions though. Quote Link to comment
eagle470 Posted January 10, 2022 Share Posted January 10, 2022 4 minutes ago, PTRFRLL said: I have not found a good way to include nvidia-settings and X11 into the container without requiring a specific version of nvidia drivers. I hesitate to do that as it would cause compatibility issues with many who use this container. I'm open to suggestions though. I'll take a beta fork, I have no problem running a secondary container, especially since it would only impact 1 gpu. Quote Link to comment
kizer Posted January 11, 2022 Share Posted January 11, 2022 MovieMatch Is there a way to sort the movies by last added vs Random? I ask this because when I add new movies to Plex the 4 of us normally have a little argument to which movie we want to watch that was recently added. Quote Link to comment
PTRFRLL Posted January 11, 2022 Author Share Posted January 11, 2022 11 hours ago, kizer said: MovieMatch Is there a way to sort the movies by last added vs Random? I ask this because when I add new movies to Plex the 4 of us normally have a little argument to which movie we want to watch that was recently added. I don't believe there is a way to sort at present. You could submit an issue on the project for consideration: https://github.com/LukeChannings/moviematch Quote Link to comment
Bolagnaise Posted January 12, 2022 Share Posted January 12, 2022 Just got to say, i found scrypted thanks to you, its a godsend for getting my unifi cameras into homekit secure video. 1 Quote Link to comment
Vatoe Posted January 18, 2022 Share Posted January 18, 2022 (edited) Can anyone tell me how to set power level for a specific GPU, via unraid console? I have two gf cards - a 2060 super - which I managed to get to 125 watts, but I would like to set the power level of the 1660 super I have, and Im not sure how to get the command to work against/for this card - I thinking I need to select/nominate it somehow first? Edited January 18, 2022 by Vatoe Quote Link to comment
PTRFRLL Posted January 18, 2022 Author Share Posted January 18, 2022 (edited) 6 hours ago, Vatoe said: Can anyone tell me how to set power level for a specific GPU, via unraid console? I have two gf cards - a 2060 super - which I managed to get to 125 watts, but I would like to set the power level of the 1660 super I have, and Im not sure how to get the command to work against/for this card - I thinking I need to select/nominate it somehow first? I believe you add the -i flag to the command to specify the id or index of the GPU you want. Use nvidia-smi to print all cards and grab the GPU ID: nvidia-smi -i 0 -pl 125 #assuming card 0 is the one you want Edited January 18, 2022 by PTRFRLL Quote Link to comment
Vatoe Posted January 18, 2022 Share Posted January 18, 2022 2 hours ago, PTRFRLL said: I believe you add the -i flag to the command to specify the id or index of the GPU you want. Use nvidia-smi to print all cards and grab the GPU ID: nvidia-smi -i 0 -pl 125 #assuming card 0 is the one you want Awesome that did the trick.! Thanks for our help. Quote Link to comment
squirrellydw Posted January 25, 2022 Share Posted January 25, 2022 (edited) On 1/11/2022 at 8:10 PM, Bolagnaise said: Just got to say, i found scrypted thanks to you, its a godsend for getting my unifi cameras into homekit secure video. Can you tell me what settings you used for Scrypted, mine says it's started but the page isn't loading Edited January 25, 2022 by squirrellydw Quote Link to comment
PTRFRLL Posted January 26, 2022 Author Share Posted January 26, 2022 On 1/25/2022 at 2:34 PM, squirrellydw said: Can you tell me what settings you used for Scrypted, mine says it's started but the page isn't loading Try this link (substitute your IP and PORT): https://IP:PORT/endpoint/@scrypted/core/public/#/ Quote Link to comment
squirrellydw Posted January 26, 2022 Share Posted January 26, 2022 I figured it out. The docker doesn’t open to the web ui. Had to go to ip plus 10443 Quote Link to comment
BigHairyWookie Posted February 12, 2022 Share Posted February 12, 2022 The latest version on T-Rex seems to be giving me some errors and not starting Version 3.6.1 works fine I checked all my configs and they haven't changed, so I'm not quite sure what's going on. Switching to the cuda10-latest tag does work, but I noticed that it was using v0.24.2 of t-rex nvidia-smi output: Any thoughts to what the problem might be? Quote Link to comment
PTRFRLL Posted February 14, 2022 Author Share Posted February 14, 2022 On 2/12/2022 at 9:38 AM, BigHairyWookie said: The latest version on T-Rex seems to be giving me some errors and not starting Version 3.6.1 works fine I checked all my configs and they haven't changed, so I'm not quite sure what's going on. Switching to the cuda10-latest tag does work, but I noticed that it was using v0.24.2 of t-rex nvidia-smi output: Any thoughts to what the problem might be? My only guess would be something related to that latest driver (510.39.01). You might try downgrading that OR seeing if the new container build (3.8) helps at all Quote Link to comment
BigHairyWookie Posted February 15, 2022 Share Posted February 15, 2022 12 hours ago, PTRFRLL said: My only guess would be something related to that latest driver (510.39.01). You might try downgrading that OR seeing if the new container build (3.8) helps at all Thanks, 3.8 seems to have fixed the issue, must've been a driver thing like you say. Quote Link to comment
Ystebad Posted February 18, 2022 Share Posted February 18, 2022 Need help with Trex docker. I have to manually enter my pool information each time it starts or it won't connect/mine. In the unraid docker edit there is only one field named "server" and here I have entered eth-us-east.flexpool.io However it won't connect unless I go into the actual GUI after it's started and select SSL then enter the pool above and enter the port 5555. Then it works fine, until the docker is restarted at which time it won't connect again. I tried editing the config.json under the tree-miner folder of appdata with the following: "pools" : [ { "pass" : "x", "url" : "stratum+ssl://eth-us-east.flexpool.io:5555", However this doesn't seem to work. What am I missing/doing wrong? advTHANKSance! Quote Link to comment
Recommended Posts
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.