Skip to content
View in the app

A better way to browse. Learn more.

Unraid

A full-screen app on your home screen with push notifications, badges and more.

To install this app on iOS and iPadOS
  1. Tap the Share icon in Safari
  2. Scroll the menu and tap Add to Home Screen.
  3. Tap Add in the top-right corner.
To install this app on Android
  1. Tap the 3-dot menu (⋮) in the top-right corner of the browser.
  2. Tap Add to Home screen or Install app.
  3. Confirm by tapping Install.

[PLUGIN] GPU Statistics

Featured Replies

36 minutes ago, Bookingyo said:

Thank you for this plugin.

Will the above files fix my blank gpu stats? Intel Arc 380.

 

image.png.339bf85f999acc9707f4cef06349c81b.png

ARC does not show all the data like igpu. Do you have anything using the GPU i.e. a docker?

 

Pluging just uses output from intel_gpu_top

  • Replies 2.1k
  • Views 355.6k
  • Created
  • Last Reply

Top Posters In This Topic

Most Popular Posts

  • juan11perez
    juan11perez

    Good day, plugin also stopped working for me.   root@Unraid:/usr/local/emhttp/plugins/gpustat# cd /usr/local/emhttp/plugins/gpustat/ && php ./gpustatus.php Fatal error: Uncaught TypeE

  • I did some more work to get it to look like the CPU sidebar that comes with unraid 

  • I have created a PR for b3rs3rk's plugin to add MultiGPU support for 6.12 as Dashboard panels are different in 6.12. The PR works on pre realeases but does not provide Multi support. There are some ot

Posted Images

I do, Plex and Immich are both using it.

4 hours ago, Bookingyo said:

I do, Plex and Immich are both using it.

Are you really sure? Are you also really sure that a task is actually running on the dGPU since if you are using direct play in Plex nothing would use the dGPU.

If something would not work then you wouldn't see 0% at the beginning from the bars.

 

What's the output from intel_gpu_top ?

image.thumb.png.18e32e6a0d722b248c37ec5f96bddf9a.pnghere is a Plex Transcode and I do see the GPU stats flickering with data, however much of the stats remain at 0% and you are not able to see which apps have locked the GPU in.image.png.b3052277f15434cabc9adc6394af3f78.png

 

Maybe to add, I am using Unraid 7.0.0-beta.2

Edited by Bookingyo

2 hours ago, Bookingyo said:

here is a Plex Transcode and I do see the GPU stats flickering with data, however much of the stats remain at 0%

This is perfectly fine and as it should be for transcoding.

 

2 hours ago, Bookingyo said:

and you are not able to see which apps have locked the GPU in.

For Intel this is a in development feature that will take a bit of extra time to be actually released.

Is fixing the power draw display on the chart for DG2 cards still in the pipeline? Everything else appears to be working normally, as you were mentioning above.

5 hours ago, eroc1990 said:

Is fixing the power draw display on the chart for DG2 cards still in the pipeline? Everything else appears to be working normally, as you were mentioning above.

That is not something the plugin cannot resolve it is dependent on Intel updating intel_gpu_top.

On 9/16/2024 at 3:46 AM, tjb_altf4 said:

Hi Simon,

I want to create a PR for some additional apps, but I'd like to test locally first... how best to approach this?
I've already cloned the repo and made the changes I need locally.


Figured it out, will send a PR once ready, cheers!
 

Added you PR in 2024.09.26 thanks.

 

Also included is showing apps for intel gpus. Some apps may not show and will need further work.

 

The same apps as NV have been added.

 

    const SUPPORTED_APPS = [ // Order here is important because some apps use the same binaries -- order should be more specific to less
        'plex'        => ['Plex Transcoder'],
        'jellyfin'    => ['jellyfin-ffmpeg'],
        'handbrake'   => ['/usr/bin/HandBrakeCLI'],
        'emby'        => ['emby'],
        'tdarr'       => ['ffmpeg', 'HandbrakeCLI'],
        'unmanic'     => ['ffmpeg'],
        'dizquetv'    => ['ffmpeg'],
        'ersatztv'    => ['ffmpeg'],
        'fileflows'   => ['ffmpeg'],
        'frigate'     => ['ffmpeg'],
        'Threadfin'   => ['ffmpeg'],
        'codeproject' => ['python3.8'],
        'deepstack'   => ['python3'],
        'nsfminer'    => ['nsfminer'],
        'shinobipro'  => ['shinobi'],
        'foldinghome' => ['FahCore'],
        'compreface'  => ['uwsgi'],
        'ollama'      => ['ollama_llama_server'],
        'immich'      => ['immich'],
        'localai'     => ['localai'],
        'chia'        => ['chia'],
        'mmx'         => ['mmx_node'],
        'subspace'    => ['subspace'],
    ];

 

On 9/17/2024 at 6:44 AM, sdballer said:

Anyway to resolve this spacing issue?

All other dashboard items have a gap at the bottom.

image.png.7c5f57e6dec0555dca2828cae320ad16.png

Space added image.png

If apps are in list above are not showing please provide the following:

 

Run intel_gpu_top

image.png

 

Replace pid in this line with pid from above.

/proc/pid/cmdline

 

and for the parent.

 

ps j %0d | awk '{ $1=$1 };NR>1' | cut -d ' ' -f 1 where %0d is pid

 

With output of 2nd pid from the above line run /proc/2nd pid/cmdline with 2nd pid.

4 hours ago, SimonF said:

That is not something the plugin cannot resolve it is dependent on Intel updating intel_gpu_top.

Fair point.

image.png.49f0f814bf66f268e4ffd03d8ca4b6a3.pngintel igpu doesent show active apps, i try plex and emby, didnot work. Anything i can do?

6 hours ago, ReddotCleaner said:

image.png.49f0f814bf66f268e4ffd03d8ca4b6a3.pngintel igpu doesent show active apps, i try plex and emby, didnot work. Anything i can do?

As per the post above. Can you run intel_gpu_top and provide a screen print and the output from the commands below the pid valie has to be put into the command where it states pid or %0d. 2nd pid is from the ps command. I will look at at creating a file in the furture with this info.

 

Replace pid in this line with pid from above.

/proc/pid/cmdline

 

and for the parent.

 

ps j %0d | awk '{ $1=$1 };NR>1' | cut -d ' ' -f 1 where %0d is pid

 

With output of 2nd pid from the above line run /proc/2nd pid/cmdline with 2nd pid.

 

Here is an example

intel-gpu-top: Intel Jasperlake (Gen11) @ /dev/dri/card0 -    8/   8 MHz;  97% RC6;      144 irqs/s

         ENGINES     BUSY                                                                                                           MI_SEMA MI_WAIT
       Render/3D    0.68% |▊                                                                                                      |      0%      0%
         Blitter    0.00% |                                                                                                       |      0%      0%
           Video    1.41% |█▌                                                                                                     |      0%      0%
    VideoEnhance    0.29% |▍                                                                                                      |      0%      0%

    PID             Render/3D                         Blitter                           Video                        VideoEnhance            NAME  
2124418 |                               ||                               ||▏                              ||▏                              | ffmpeg
2124421 |▏                              ||                               ||▏                              ||                               | ffmpeg
2206315 |▏                              ||                               ||▏                              ||                               | ffmpeg

^C
root@LincstationN1:~# cat /proc/2124418/cmdline
ffmpeg-hide_banner-loglevelwarning-threads2-hwaccel_flagsallow_profile_mismatch-hwaccelvaapi-hwaccel_device/dev/dri/renderD128-hwaccel_output_formatvaapi-user_agentFFmpeg Frigate/0.14.1-f4f3cfa-avoid_negative_tsmake_zero-fflags+genpts+discardcorrupt-rtsp_transporttcp-timeout5000000-use_wallclock_as_timestamps1-irtsp://admin:[email protected]/h264Preview_01_sub-r5-vffps=5,scale_vaapi=w=640:h=480:format=nv12,hwdownload,format=nv12,format=yuv420p-threads2-frawvideo-pix_fmtyuv420ppipe:root@LincstationN1:~# 
root@LincstationN1:~# ps j 2124418 | awk '{ $1=$1 };NR>1' | cut -d ' ' -f 1 
2124407
root@LincstationN1:~# cat /proc/2124407/cmdline
frigate.capture:house1root@LincstationN1:~# 

image.thumb.png.3517f328dc435128c66146d876a6946c.pngimage.thumb.png.4410fc2a39fa2427b6c12d3b9b77b6fa.pnghere, tell me if u need anything more

13 hours ago, ReddotCleaner said:

image.thumb.png.3517f328dc435128c66146d876a6946c.pngimage.thumb.png.4410fc2a39fa2427b6c12d3b9b77b6fa.pnghere, tell me if u need anything more

If you do the following change

 

nano /usr/local/emhttp/plugins/gpustat/lib/Intel.php

 

scroll down to 

    const SUPPORTED_APPS = [ // Order here is important because some apps use the same binaries -- order should be more specific to less
        'plex'        => ['Plex Transcoder'],
        'jellyfin'    => ['jellyfin-ffmpeg'],
        'handbrake'   => ['/usr/bin/HandBrakeCLI'],
        'emby'        => ['emby'],

and change to

    const SUPPORTED_APPS = [ // Order here is important because some apps use the same binaries -- order should be more specific to less
        'plex'        => ['Plex Transcoder'],
        'jellyfin'    => ['jellyfin-ffmpeg'],
        'handbrake'   => ['/usr/bin/HandBrakeCLI'],
        'emby'        => ['ffmpeg','emby'],

save and exit(ctrl s and ctrl x)

 

Can confirm latest GPU Stats from regular updates works on my Arc 750 as well in Plex.

image.png.0bac084be1b144cd13208aad844b492e.png

I am running a i5-13500t and when transcoding, gpu stats shows everything at 0. i have intel_gpu_top installed as well. wattage seems to be reporting correctly in intel_gpu_top, but thats about it. tested by transcoding a video file in plex. image.thumb.png.5d812a4b709277cf804c53381934181b.png

image.thumb.png.cb322b5161a8129710f7dc442898c361.png

 

EDIT: i figured it out, i had to add /dev/dri as a device path and added the iGPU within plex settings. i used this guide here, my only issue now is that IMC Bus Utilization is not populating with transmit and receive showing as N/A. Anyone know how to fix this? Is it related to cache pools? because i dont have that setup. I thought it was related to RAM cache and added an extra parameter for that in the plex docker settings but that didnt fix it

image.thumb.jpeg.c9d339414c455eb8852428e4b7a7f6ae.jpeg

 

  

On 2/26/2024 at 4:58 AM, ich777 said:

 

It is a known issue that no IMC Bus utilization is shown on 13th Gen+

 

oh... i did not know this

Edited by cheese24

Good day all. Would this plugin show 2 GPUs? I have a second one connected but I don't see it in the dropdown where you select the GPU you want to show. Maybe I did something wrong, so want to ask if that's the case.

19 minutes ago, NotHere said:

Good day all. Would this plugin show 2 GPUs? I have a second one connected but I don't see it in the dropdown where you select the GPU you want to show. Maybe I did something wrong, so want to ask if that's the case.

Which os ver are you running. Multi was only added in 6.12+

 

What does lspci | grep VGA show

37 minutes ago, SimonF said:

Which os ver are you running. Multi was only added in 6.12+

 

What does lspci | grep VGA show

Thanks so much for the reply SimonF. User error :(. Having issues with both GPUs in Unraid but now they are both showing :). 

Hi, I have three GPUs in my 6.12.13 system and two of them are Nvidia P40s (04:00:0 and 06:00:0) added a few weeks ago and the other one is the integrated Intel 770 (00:02:0). When I select any one GPU, they all show up on the dashboard, When I select any one P40 and the Intel 770, they both show up normally on the dashboard.

 

When I select all three GPUs, the Intel info is displayed on the dashboad but only one Nvidia is displayed (06:00:0). To make things a bit more confusing, when I select the two P40s, only one is displayed on the dashboard and it's the 04:00:0.

 

Note: Both Nvidia GPUs show up using nvidia-smi and they both work as I can isolated them and run workloads on them individually.

 

Is this a bug or is there something under my control?

Edited by simpleServer

On 9/12/2024 at 6:52 PM, Sptz87 said:

 

 

Hi all,

 

Today, out of nowhere this started to happen with this plugin. Any idea how to get it to stick? Notice the error that shows at the bottom? I have no VMs at all.

I am having this problem now as well.  The plugin worked fine until recently, when it started exhibiting the behavior displayed in Sptz87's video.  The GPU is still functioning the same as it always has.  I'm not running any VMs, either.

 

EDIT:  Seems to be a cookies issue, as it works when I log in via incognito mode.  For now I'll just run Unraid that way rather than clear my cookies and log myself out of everything.

Edited by MrMastodonFarm

Running unRAID 7.0-beta3 (Beta3 is when the issue started as it was fine with Beta2) with an AMD RX6400 GPU

 

Dashboard shows 100% load on the GPU when idle

image.png.990a0be803529ccfa1e65b32bb919267.png

 

If a docker (Such as Plex) starts using the GPU it will display accurately until it becomes idle again and it will jump back to 100%

 

radeontop also shows 100% usage

image.png.4818fc7a08a8edfb0bcbe3c2d5634524.png

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...

Account

Navigation

Search

Search

Configure browser push notifications

Chrome (Android)
  1. Tap the lock icon next to the address bar.
  2. Tap Permissions → Notifications.
  3. Adjust your preference.
Chrome (Desktop)
  1. Click the padlock icon in the address bar.
  2. Select Site settings.
  3. Find Notifications and adjust your preference.