Nice, thanks for the quick turnaround!

Dude! do you have a coffee link or something, What a legend!!

8 minutes ago, kerbys said:

Dude! do you have a coffee link or something, What a legend!!with

In CA within the info section for my plugins is a donate link.

I do notice the compute bar seems inverted. Like higher compute makes the bar shrink and less compute makes it grow

21 hours ago, Maximo101 said:

Is the Fan Power (watts) the total draw of power from my gpu, or is there more power used over the fan?

They are two settings fan is one bar rpm and power is the other for the GPU. So it is the total power used by the GPU.

12 minutes ago, bmrowe said:

I do notice the compute bar seems inverted. Like higher compute makes the bar shrink and less compute makes it grow

Thanks will look into that.

Will be fixed in next release.

On 5/26/2025 at 9:36 PM, Akagami said:

CodeProject.AI docker (codeproject/ai-server:cuda12_2) with Object Detection (YOLOv5 6.2) is incorrectly reporting in the nvidia gpu plugin (Nvidia Driver Version: 575.51.02 - plugin up to date 2025.03.25) dashboard as beeing in use by "immich (python 3.8)" as i do not have immich installed i had to trial and error restaring all dockers with gpu to find it. I just recently noticed this as before i had never seen the codeproject docker reported in the plugin

Can you run php /usr/local/emhttp/plugins/gpustat/lib/Pidcheck 3459075 where the number is the PID from nvida-smi and post the results.

@SimonF Updated to the latest plugin just now and still no power stat ☹️

amdgpu-pci-0100

Adapter: PCI adapter

vddgfx: 718.00 mV

fan1: N/A (min = 1600 RPM, max = 6000 RPM)

edge: +41.0°C (crit = +99.0°C, hyst = -273.1°C)

PPT: 6.24 W (cap = 28.00 W)

5 hours ago, SimonF said:

Can you run php /usr/local/emhttp/plugins/gpustat/lib/Pidcheck 3459075 where the number is the PID from nvida-smi and post the results.

PID is:2590993

PID Info

python3.8/app/preinstalled-modules/ObjectDetectionYOLOv5-6.2/detect_adapter.py

Parent for PID Info

ps j 2590993 | awk '{ $1=$1 };NR>1' | cut -d ' ' -f 1

Parent PID:2590593

./CodeProject.AI.Server

On 5/27/2025 at 6:32 PM, bernieJo said:

  On 5/27/2025 at 12:30 AM, yuzufilm said:

Hi, I'm having an issue in which the dashboard component of the plugin will not stay open/collapses itself pretty much instantly.

Same Problem here ! I have only an Alder Lake GPU (N100) and did not explicitly bound it to VFIO, I only use the it in a vm as virtual Graphic Card, as introduced with Unraid 7.1. With the "Pre-Release" test received from Simon all worked well. It's the same behavior whether the vm is activ or not. What changed ??

Still same Problem. If others don't see that, did I misconfigure something ?

Edited May 31, 20251 yr by bernieJo

7 hours ago, bernieJo said:

Still same Problem. If others don't see that, did I misconfigure something ?

Ohhhh, I See that it is ok now, does not collapse anymore. I opened it, it collapsed and then I started a VM that is using VGPU. Now it is open and displays the use of the iGPU.

Looks that before opening and using a VGPU something in the plugin was not initialized properly ??

3 hours ago, bernieJo said:

Ohhhh, I See that it is ok now, does not collapse anymore. I opened it, it collapsed and then I started a VM that is using VGPU. Now it is open and displays the use of the iGPU.

Looks that before opening and using a VGPU something in the plugin was not initialized properly ??

It could also be the cache, clearing the site cache fixed it for me

###2025.06.02

- Fix AMD Power readings for different card types.

- Fix Compute Bar

- Fix CodeProject detection

On 6/2/2025 at 7:33 PM, SimonF said:

###2025.06.02

- Fix AMD Power readings for different card types.

- Fix Compute Bar

- Fix CodeProject detection

works like a charm

On 5/30/2025 at 5:55 AM, SimonF said:

They are two settings fan is one bar rpm and power is the other for the GPU. So it is the total power used by the GPU.

Thank you, so my gpu is using 180w in this image (ollama running a local ai model responding to a complex task, as it idles at 10watts)

3 hours ago, Maximo101 said:

Thank you, so my gpu is using 180w in this image (ollama running a local ai model responding to a complex task, as it idles at 10watts)

Yes that is correct.

Hi I just got my nvidia 5060 gpu working with opensource nvidia plugin driver. However, now the GPU Statistics plugin is not working on my dashboard. Is this GPU supported? Is there anything I can do to get it to work?

I forgot to select the GPU ID in the plugin settings. But its not reporting the GPU device model. For the record device model is not shown in the opensource nvidia driver plugin either. I have seen other screen shots with 5070 ti's posted above. Any advice is appreciated.

Edited June 11, 20251 yr by CaptainTrap
added additional detail

Hello, for some reason my GPU has all 0's now and its not showing any information, previously worked flawlessly.

Edited June 23, 20251 yr by itbemark

5 minutes ago, itbemark said:

Hello, for some reason my GPU has all 0's now and its not showing any information, previously worked flawlessly.

What does nvidia-smi show?

2 minutes ago, SimonF said:

What does nvidia-smi show?

Shit just went and looked, said, No GPU Installed, FML... okay it's a me issue not a plugin issue. Thank you!

Okay this perplexes me, GPU has been working fine, I did add a 10 gig NIC, but fan is running on GPU, lights turning on, but does not recognize it. I event removed the new NIC still same issue.

3 minutes ago, itbemark said:

Okay this perplexes me, GPU has been working fine, I did add a 10 gig NIC, but fan is running on GPU, lights turning on, but does not recognize it. I event removed the new NIC still same issue.

Does it show in lspci and the ID is the same?

1 hour ago, SimonF said:

Does it show in lspci and the ID is the same?

GPU Officially DEAD, tried it in another system.. RIP 1080ti, you were a loyal and long standing solider!

Title: Severe CPU spike and system instability after installing GPU Statistics plugin

Description:
After installing the GPU Statistics plugin on my system, everything becomes unstable. CPU usage spikes, the system becomes sluggish or freezes intermittently, and GPU passthrough behavior becomes unpredictable.

These issues occur even though:

• Both GPUs are in separate IOMMU groups,

• Proper drivers (including open-source NVIDIA drivers) are installed,

• No VM is using both GPUs simultaneously,

• PCIe slots and device assignments are configured correctly.

No changes to drivers, device bindings, or PCIe ports fix the issue.
The problem immediately disappears after uninstalling the GPU Statistics plugin.

Setup:

• Gigabyte X870E AORUS ELITE WIFI7

• AMD Ryzen 9 9950X3D

• GPU 1: RTX 5080 (passed through to VM)

• GPU 2: RTX A2000 (idle or used by host)

• Unraid 6.x (7.1.4)

Please investigate this – it’s 100% reproducible and tied directly to the plugin being installed, regardless of active GPU usage.

Edited June 29, 20251 yr by RBK-Serwer

On 6/30/2025 at 12:13 AM, RBK-Serwer said:

Title: Severe CPU spike and system instability after installing GPU Statistics plugin

Description:
After installing the GPU Statistics plugin on my system, everything becomes unstable. CPU usage spikes, the system becomes sluggish or freezes intermittently, and GPU passthrough behavior becomes unpredictable.

These issues occur even though:

• Both GPUs are in separate IOMMU groups,

• Proper drivers (including open-source NVIDIA drivers) are installed,

• No VM is using both GPUs simultaneously,

• PCIe slots and device assignments are configured correctly.

No changes to drivers, device bindings, or PCIe ports fix the issue.
The problem immediately disappears after uninstalling the GPU Statistics plugin.

Setup:

• Gigabyte X870E AORUS ELITE WIFI7

• AMD Ryzen 9 9950X3D

• GPU 1: RTX 5080 (passed through to VM)

• GPU 2: RTX A2000 (idle or used by host)

• Unraid 6.x (7.1.4)

Please investigate this – it’s 100% reproducible and tied directly to the plugin being installed, regardless of active GPU usage.

Do you get the same effect running nvidia-smi? Is 5080 bound to vfio?