nicksphone

Using the AIO build of the docker?

you using the nvida build of unraid? Using the AIO build of the docker?

your missing Container Variable: NVIDIA_VISIBLE_DEVICES with the gpu id

Tdarr_Plugin_MC93_Migz1FFMPEG Migz-Transcode Using Nvidia GPU & FFMPEG

6 trans codes Every 0.5s: nvidia-smi Tower: Wed May 13 10:43:47 2020 Wed May 13 10:43:47 2020 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 440.59 Driver Version: 440.59 CUDA Version: 10.2 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GT 710 Off | 00000000:08:00.0 N/A | N/A | | N/A 29C P8 N/A / N/A | 44MiB / 980MiB | N/A Default | +-------------------------------+----------------------+----------------------+ | 1 GeForce GTX 105... Off | 00000000:09:00.0 Off | N/A | | 38% 44C P0 N/A / 75W | 1577MiB / 4040MiB | 43% Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 Not Supported | | 1 6218 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 236MiB | | 1 11010 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 382MiB | | 1 27130 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 236MiB | | 1 27200 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 236MiB | | 1 27343 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 236MiB | | 1 31088 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 236MiB | +-----------------------------------------------------------------------------+ 2 Trans Codes Wed May 13 10:48:31 2020 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 440.59 Driver Version: 440.59 CUDA Version: 10.2 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GT 710 Off | 00000000:08:00.0 N/A | N/A | | N/A 29C P8 N/A / N/A | 44MiB / 980MiB | N/A Default | +-------------------------------+----------------------+----------------------+ | 1 GeForce GTX 105... Off | 00000000:09:00.0 Off | N/A | | 39% 46C P0 N/A / 75W | 866MiB / 4040MiB | 60% Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 Not Supported | | 1 11010 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 382MiB | | 1 18842 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 470MiB | +-----------------------------------------------------------------------------+ single transcode Every 0.5s: nvidia-smi Tower: Wed May 13 10:49:41 2020 Wed May 13 10:49:41 2020 +-----------------------------------------------------------------------------+ | NVIDIA-SMI 440.59 Driver Version: 440.59 CUDA Version: 10.2 | |-------------------------------+----------------------+----------------------+ | GPU Name Persistence-M| Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util Compute M. | |===============================+======================+======================| | 0 GeForce GT 710 Off | 00000000:08:00.0 N/A | N/A | | N/A 29C P8 N/A / N/A | 44MiB / 980MiB | N/A Default | +-------------------------------+----------------------+----------------------+ | 1 GeForce GTX 105... Off | 00000000:09:00.0 Off | N/A | | 39% 45C P0 N/A / 75W | 484MiB / 4040MiB | 38% Default | +-------------------------------+----------------------+----------------------+ +-----------------------------------------------------------------------------+ | Processes: GPU Memory | | GPU PID Type Process name Usage | |=============================================================================| | 0 Not Supported | | 1 18842 C ...rver/assets/app/ffmpeg/ffmpeg345/ffmpeg 470MiB | +-----------------------------------------------------------------------------+

on my driver modded 1050ti no matter how any trans codes i have running i get a max of 60% usage of gpu usually 40-50% and its not a cpu bottle neck got 12 out of my 16 threads pinned to the docker (matching physical core with thread) using watch nvidia-smi while the trans code is running

Handbrake NVEC 264 configurable lets you pick container.

im using Tdarr_Plugin_MC93_Migz1FFMPEG Migz-Transcode Using Nvidia GPU & FFMPEG

I upped the allowed CPU cores and still only uses 50 to 58 % of the GPU everything else is as far as I know set to default. It's the only docker assigned the GPU so not sure what's wrong if people are getting more utilisation.

@HaveAGitGat using the driver mod no matter how many transcodes I set the GPU is only using about 50% and my CPU usage is low as well. Is this the max or is there a setting I missed

i looks like an issue i had before we had backups, goto backups see if you have one before the update. if so click the clear button. restart tdarr, restore the backup, restart again. that should fix it if you have not changed any locations in the settings for file locations if its the same issue i had. if thats not it tag haveagit on your reply

are you trying to pass through the same gpu to all 3? if so as soon as one grabs it the others think is not there anymore its not like windows where multiple programs can use it at once. disable the plex and emby gpu pass though and try it again. also turn on the Linux FFmpeg NVENC binary in options see if that helps.

Have you considered having your temp files in a folder not scanned my what ever tool is picking it up only takes making another share and pointing it there. alternatively if you have enough ram to you do it in /tmp

its using 2.5gigs of the 4 gigs its saved its self i would not worry about it unless your docker.img is on a small cache drive krusadar a simple file manager uses 2.4 gigs of docker space. ive been running it for months and im using 2.5 gigs still. if your running low on docker space change the size a few gigs. if your docker is filling up something else might be the hog.

That was 4k reencodes i max out the load and stress. With 20 encodes you might want to get some test files made to see the transcode times. I noticed 3 to 4 times longer when in had more transcodes going so your not saving any time and adding more wear on the drives.

Save your money and get a 1050ti and use the driver mod to let it do more than 2 transcodes. you will run out of cpu before you run out of gpu. i have a 2700x pinned on 6 cores and i get maxed out on the cpu at 6 transcodes i dropped it to 3 so less failed and my system did not lag the gpu was reading 35% usage. it churns through the files fast enough when i was just using 2 GPU encoding is soooooooo much faster than cpu as it uses both.

@HaveAGitGat first let me thank you again your app has saved me tons of time and space. Correct me if im wrong but in the early days one of your hopes was integration into sonarr and radarr so it would try to reque failed files am i remembering that correctly or am i thinking of another app i tried? ether way is that in the pipeline? as it would make this amazing app even better.

1.07 the last update I did not change anything. I changed db location now and it's starting now not a big deal redoing the settings but thought you might want to know as all I did was upgrade from the last version

Update broke the AIO container Starting database mongodb ...fail!

wow thank you so much i know im a pain in the ass but im trying to learn

doing it this way i would have to make separate libs so it still strips the meta from all the files?

my real concern is the 720p and 1080p files that were in 264 growing in size usally 1.5x on 720 and 2-3 times the size on 1080p no point re-encoding them if its going to take up more space using the plugin is there a way for it to only do 4k files if this is the norm?

using the handbrake plugin im not realy saving any space with the nvec h265 what im saving on the 4k files im loosing on half my 1080's doubling in size. im not great with setting these things up but i was saving more space with unmanic and not seeing a huge quality drop any idea what settings he was using?

i did not add a random name for the DB that was the issue i was unsure if it made a file or not Edit 1: Another silly question. with the aio docker do i need mongodb docker anymore if im not using it for anything else?

pointed it to an appdata folder and get * Starting database mongodb ...fail! * Starting database mongodb ...fail!