[PLUGIN] GPU Statistics

[SimonF]

19 hours ago, WingmanNZ said:

Unraid Version: 6.12.0-rc2.9

Plugin Version: 2023.03.27

Created 2023.04.12a with this fix. Thanks for the code change.

[WingmanNZ]

9 hours ago, SimonF said:

Created 2023.04.12a with this fix. Thanks for the code change.

Thank you. I have installed that from https://raw.githubusercontent.com/SimonFair/gpustat-unraid/master/gpustat.plg and can confirm it resolves the issue

[mattalat]

Any update when an 6.12 RC compatible version will be available on the community apps page?

[tjb_altf4]

Any chance we can see the config for supported apps moved from being hardcoded to a config file ?

I have some containers I've put together that use GPU which are unlikely to be used/useful for most others, and would be ideal if I could add without those without needing to modify/fork the plugin.

Thanks!

[SimonF]

On 4/15/2023 at 4:35 AM, mattalat said:

Any update when an 6.12 RC compatible version will be available on the community apps page?

I have added my version to CA until my PR is merged.

[SimonF]

On 4/24/2023 at 4:51 AM, tjb_altf4 said:

Any chance we can see the config for supported apps moved from being hardcoded to a config file ?

I have some containers I've put together that use GPU which are unlikely to be used/useful for most others, and would be ideal if I could add without those without needing to modify/fork the plugin.

Thanks!

Would this be for Nvidia, do you know the commands to look for? These are the current one for NV

 

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'],

'deepstack' => ['python3'],

'nsfminer' => ['nsfminer'],

'shinobipro' => ['shinobi'],

'foldinghome' => ['FahCore'],

];

[husqnz]

More of a suggestion than an issue - for those of us who use the Dark theme, it would be good for there to be a light/white background for logos of apps which are also dark (Frigate is the main one which comes to mind).

 

Obviously not a priority but a nice to have

[tjb_altf4]

On 4/26/2023 at 2:28 AM, SimonF said:

Would this be for Nvidia, do you know the commands to look for? These are the current one for NV

Yeah all Nvidia and commands I can see in nvidia-smi... as far as I could find there are 3 areas to update:

• apps array in gpustatus.page
• nvidia class in Nvidia.php
• app icons in images folder

Should I make a pull request instead and just add what I need?

 

[SimonF]

2 hours ago, tjb_altf4 said:

Yeah all Nvidia and commands I can see in nvidia-smi... as far as I could find there are 3 areas to update:

• apps array in gpustatus.page
• nvidia class in Nvidia.php
• app icons in images folder

Should I make a pull request instead and just add what I need?

 

It would be achievable to just don't know if I have the time to action at present. There are two versions of the status page there is the mult.page one also. Would be easier with a PR or if you send me the info I can update mine repo Multi GPU PR.

 

On 4/27/2023 at 9:17 PM, husqnz said:

More of a suggestion than an issue - for those of us who use the Dark theme, it would be good for there to be a light/white background for logos of apps which are also dark (Frigate is the main one which comes to mind).

 

Obviously not a priority but a nice to have

Can you provide a screen print of your output?

[LeBlackNight]

I really appreciate this plugin as it is super useful for monitoring the number of encode operations that I have going on at once with Plex. It is spamming my sys.log file with the following. Is this something that can be fixed I hope?

 

Let me know what information I can provide.

 

May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)
May  1 04:55:31 Tower kernel: ACPI BIOS Error (bug): Failure creating named object [\_SB.PC00.PEG1.PEGP._DSM.USRG], AE_ALREADY_EXISTS (20220331/dsfield-184)
May  1 04:55:31 Tower kernel: ACPI Error: AE_ALREADY_EXISTS, CreateBufferField failure (20220331/dswload2-477)
May  1 04:55:31 Tower kernel: ACPI Error: Aborting method \_SB.PC00.PEG1.PEGP._DSM due to previous error (AE_ALREADY_EXISTS) (20220331/psparse-529)

 

[sage2050]

I tried to change the refresh rate to 30s (30000) and it's still pinging my cpu every second. I also turned it off entirely but it's still going every second. Do i need to restart the server for this setting to take effect?

 

edit: nevermind - I had the dash still open on a second machine. once I navigated away the setting took effect.

Edited May 1, 2023 by sage2050

[husqnz]

On 4/29/2023 at 5:59 PM, SimonF said:

It would be achievable to just don't know if I have the time to action at present. There are two versions of the status page there is the mult.page one also. Would be easier with a PR or if you send me the info I can update mine repo Multi GPU PR.

 

Can you provide a screen print of your output?

 

Depending on your screen contrast/brightness settings, you should be able to see a faint black frigate bird - in dark mode, hard to see. Fine in light mode obviously!

 

 

[jakeisrollin]

On 4/11/2023 at 6:04 AM, ich777 said:

Because this card doesn't report the power draw on Linux.

is there somewhere i can find a list of gpus that report the power draw? i have a quadro p1000 and i am getting 0w so i want to see if mine is does

[ich777]

3 hours ago, jakeisrollin said:

i have a quadro p1000 and i am getting 0w so i want to see if mine is does

Please read this post:

 

3 hours ago, jakeisrollin said:

is there somewhere i can find a list of gpus that report the power draw?

No, because even if it does now that doesn't mean that Nvidia will break it in the future, this was the case for many card over time.

[SimonF]

On 5/9/2023 at 9:21 PM, husqnz said:

 

Depending on your screen contrast/brightness settings, you should be able to see a faint black frigate bird - in dark mode, hard to see. Fine in light mode obviously!

 

 

If you can provide a png file I can look to replace the existing one. I can only find all black.

[ConnerVT]

Well, Frigate, it's not that important.

 

😉

(Sorry - I just couldn't resist.)  

[ezek1el3000]

Recently Power and Throttling wont show up. I only got N/A - The GPU is a nVidia P2000. Unraid 6.11.5. Any idea why it's not displayed? I recently updated my bios and while i was at it i updated my Intel Management Engine Firmware. Everything works just fine except this.

diagnostics-20230601-2049.zip

[SimonF]

20 hours ago, ezek1el3000 said:

Recently Power and Throttling wont show up. I only got N/A - The GPU is a nVidia P2000. Unraid 6.11.5. Any idea why it's not displayed? I recently updated my bios and while i was at it i updated my Intel Management Engine Firmware. Everything works just fine except this.

diagnostics-20230601-2049.zip 196.39 kB · 0 downloads

what output do you get from nvidia-smi -q -x -g 0

[ezek1el3000]

nvidia_smi_log.txt

[SimonF]

32 minutes ago, ezek1el3000 said:

nvidia_smi_log.txt 42.23 kB · 2 downloads

Looks like the driver may have changed output. GPU Stat is looking for power_readings but now seem to be module + gpu.

 

Throttle does not seem to be reported @ich777 is 535.43.02 a newish update?

 

                <gpu_power_readings>
                        <power_state>P8</power_state>
                        <power_draw>6.97 W</power_draw>
                        <current_power_limit>75.00 W</current_power_limit>
                        <requested_power_limit>75.00 W</requested_power_limit>
                        <default_power_limit>75.00 W</default_power_limit>
                        <min_power_limit>75.00 W</min_power_limit>
                        <max_power_limit>75.00 W</max_power_limit>
                </gpu_power_readings>
                <module_power_readings>
                        <power_state>P8</power_state>
                        <power_draw>N/A</power_draw>
                        <current_power_limit>N/A</current_power_limit>
                        <requested_power_limit>N/A</requested_power_limit>
                        <default_power_limit>N/A</default_power_limit>
                        <min_power_limit>N/A</min_power_limit>
                        <max_power_limit>N/A</max_power_limit>
                </module_power_readings>

[ich777]

1 hour ago, SimonF said:

is 535.43.02 a newish update?

Yes, this is the newest beta driver released a few days ago.

It is possible that this is an oversight from Nvidia but in my opinion it is more likeley that they drop this feature when its not reported anymore.

[SimonF]

12 hours ago, ich777 said:

Yes, this is the newest beta driver released a few days ago.

It is possible that this is an oversight from Nvidia but in my opinion it is more likeley that they drop this feature when its not reported anymore.

Can you config the output you get for power readings with nvidia-smi -q -x -g 0 my GPU uses the really old driver. I could look to make a change if driver > 535.43.02

[ich777]

1 hour ago, SimonF said:

I could look to make a change if driver > 535.43.02

Sure thing, give me a bit, I‘m currently not at home and have to put the Nvidia card into my test server.

[ich777]

On 6/3/2023 at 8:32 AM, SimonF said:

I could look to make a change if driver > 535.43.02

Is this what you where after?

535.43.02.txt

[SimonF]

1 hour ago, ich777 said:

Is this what you where after?

535.43.02.txt 25.67 kB · 1 download

Yes thanks looks like they have change the name of power readings to gpu power readings