Intel® Distribution of OpenVINO™ Toolkit
Community assistance about the Intel® Distribution of OpenVINO™ toolkit, OpenCV, and all aspects of computer vision-related on Intel® platforms.
6401 Discussions

How to attach iGPU into OVMS service in docker-compose configuration file

Aqil-ITXOTIC
Beginner
1,484 Views

Hi,

I try to run OVMS as a service in docker-compose, but i don't know how to attach iGPU. So far i run OVMS as a separate service outside from docker-compose. It would be great if my docker-compose can start all its service without need to run one specific server outside from it.

 

Here is the screenshot of the OVMS part in my docker-compose file configuration.

Screenshot from 2023-02-22 09-54-04.png

0 Kudos
5 Replies
Wan_Intel
Moderator
1,439 Views

Hi Aqil-ITXOTIC,

Thanks for reaching out to us.

Are you trying to attach iGPU into OpenVINO™ Model Server in docker-compose configuration file on Windows?

 

On another note, if you are trying to attach iGPU into OpenVINO™ Model Server in docker-compose configuration file on Ubuntu, please use the following command:

devices:
  - "/dev/dri"

 

 

Regards,

Wan

 

0 Kudos
Aqil-ITXOTIC
Beginner
1,430 Views

Hi Wan,

 

Thank you for your response. Unfortunately i tried what you said, and it doesn't work. That's why it was commented out in my screenshot. I don't know what happen, the OVMS doesn't detect my iGPU, only CPU and GNA, when i enable OVMS inside docker-compose.yaml file.

Hence this force me to disable it from my docker-compose and run it separately.

 

Here is the information showed when OVMS is integrated in docker-compose (ofc by putting what you said, attach dev/dri) :

Screenshot from 2023-02-23 15-56-17.png

 

And this is the separate script i wrote to up OVMS (and put it in same network of other containers i'm running) :

Screenshot from 2023-02-23 15-57-53.png

By running this way, it does detect iGPU.

For your further information, i am running Ubuntu 22.04 with Kernel 5.17-oem (with latest Intel Graphic Driver), processor Intel® Core™ i5-1135G7.

The models i'm running, i run it on target-device "AUTO" mode, so that Openvino will pick the best devices for inferencing a video stream that i feed into.

For me, it just an inconvenient way to run my application with this way, docker-compose up for all containers necessary except OVMS, which have to run separately.

 

Looking forward for your response and your assistance.

0 Kudos
Wan_Intel
Moderator
1,426 Views

Hi Aqil-ITXOTIC,

Thanks for your information.

For your information, I've validated the command above as shown as follows:

ubuntu_gpu.png

We regret that the command above doesn't work on your machine. I've escalated this to our next level and we'll update you as soon as possible.

 

 

Regards,

Wan

 

0 Kudos
Wan_Intel
Moderator
1,350 Views

Hi Aqil-ITXOTIC,

Thanks for your patience.

We regret to inform you that we're unable to replicate your issue because we're able to attach iGPU in OpenVINO™ Model Server with configuration file via docker compose. Sorry for the inconvenience.

 

 

Regards,

Wan

 

0 Kudos
Wan_Intel
Moderator
1,275 Views

Hi Aqil-ITXOTIC,

If you need additional information from Intel, please submit a new question as this thread will no longer be monitored.

 

 

Regards,

Wan

 

0 Kudos
Reply