Install Intel® NPU firmware
================================

Introduction
:::::::::::::::::::::::::::

Intel® NPU device is an AI inference accelerator integrated with Intel client CPUs,
starting from Intel® Core™ Ultra generation of CPUs (formerly known as Meteor Lake).
It enables energy-efficient execution of artificial neural network tasks.

To make sure that your system has an NPU available, please follow
`the steps <https://www.intel.com/content/www/us/en/support/articles/000097597/processors.html>`_.

The full device name is Neural Processing Unit, but the Linux kernel driver uses
the older name - Versatile Processing Unit (VPU).

Installation
:::::::::::::::::::::::::::

1. Remove old packages

.. code-block:: bash

    sudo dpkg --purge --force-remove-reinstreq intel-driver-compiler-npu intel-fw-npu intel-level-zero-npu

2. Download all debs package

.. code-block:: bash

    wget https://github.com/intel/linux-npu-driver/releases/download/v1.13.0/intel-driver-compiler-npu_1.13.0.20250131-13074932693_ubuntu22.04_amd64.deb --no-check-certificate
    wget https://github.com/intel/linux-npu-driver/releases/download/v1.13.0/intel-fw-npu_1.13.0.20250131-13074932693_ubuntu22.04_amd64.deb --no-check-certificate
    wget https://github.com/intel/linux-npu-driver/releases/download/v1.13.0/intel-level-zero-npu_1.13.0.20250131-13074932693_ubuntu22.04_amd64.deb --no-check-certificate

3. Install ``libtbb12`` which is a dependency for ``intel-driver-compiler-npu``

.. code-block:: bash

    sudo apt update
    sudo apt install libtbb12

4. Install all packages

.. code-block:: bash

    sudo dpkg -i *.deb

5. Install Level Zero if it is not in the system

.. code-block:: bash

    # Check if Level Zero is installed
    dpkg -l level-zero

Download and install package if Level Zero is missing

.. code-block:: bash

    wget https://github.com/oneapi-src/level-zero/releases/download/v1.18.5/level-zero_1.18.5+u22.04_amd64.deb --no-check-certificate
    sudo dpkg -i level-zero*.deb

6. Reboot

.. code-block:: bash

    reboot
    # if everything works, we should see /dev/accel/accel0 device
    ls /dev/accel/accel0
    /dev/accel/accel0
    # to receive intel_vpu state
    dmesg | grep intel_vpu

7. User access to the device

As a root user, this step can be skipped.

The new device ``/dev/accel/accel0`` requires manual setting of permissions access.
The ``accel`` devices should be in the "render" group in Ubuntu:

.. code-block:: bash

    # set the render group for accel device
    sudo chown root:render /dev/accel/accel0
    sudo chmod g+rw /dev/accel/accel0
    # add user to the render group
    sudo usermod -a -G render <user-name>
    # user needs to restart the session to use the new group (log out and log in)

The above steps must be repeated each time module is reloaded or on every reboot.
To avoid manual setup of the group for ``accel`` device, the udev rules can be used:

.. code-block:: bash

    sudo bash -c "echo 'SUBSYSTEM==\"accel\", KERNEL==\"accel*\", GROUP=\"render\", MODE=\"0660\"' > /etc/udev/rules.d/10-intel-vpu.rules"
    sudo udevadm control --reload-rules
    sudo udevadm trigger --subsystem-match=accel

.. tip:: In case of NPU is not visible, always check the access to the device with following command:

    .. code-block:: console

        ls -lah /dev/accel/accel0
        crw-rw---- 1 root render 261, 0 Mar 22 13:22 /dev/accel/accel0

    If render is missing, or ``crw-rw----`` is not set, please repeat the steps to set the access to the device.