davincibox

This project aims to provide a ready-to-go container with all of the needed dependencies to install and run DaVinci Resolve on Linux, based on information compiled by bluesabre in his GitHub Gist. Davincibox is primarily intended for users of image-based systems such as Fedora Atomic Desktops and Universal Blue images, but it can be used on any Linux distro that distrobox is available on.

Disclaimers

This project and its maintainer are not affiliated nor associated with BlackMagicDesign.

OS Support

The following distros are considered "officially" supported:

• Fedora Atomic (i.e. Silverblue, Kinoite)
• Bluefin
• Bazzite

Support for other distros such as Arch Linux, NixOS, Debian, and any derivatives are done on a best effort basis, but considered community supported. The priority for davincibox is for image-based/atomic systems due to those benefiting most from the container-based approach.

For Debian, Ubuntu, and Linux Mint users, if you have difficulties with using davincibox, another great option for using DaVinci Resolve is MakeResolveDeb by Daniel Tufvesson. Do note that if using MakeResolveDeb, if using an AMD GPU, you will need to install ROCm (or rusticl) yourself; for Intel, you will need intel-compute-runtime.

GPU Support

My own testing for davincibox has been with an AMD RX 6600XT GPU, an AMD RX 9070 GPU, and an Intel Core Ultra 5 125H (i.e. Intel Arc integrated graphics). Other users have contributed test results here: #21

AMD GPUs use ROCm (AMD's official compute package). There have previously been stability issues with using ROCm, typically in Fusion, but otherwise most everything does work. Note that AMD's official GPU support for ROCm is limited, and users of older AMD GPUs (such as the RX 500 series) may not have success with ROCm.

Intel GPUs use intel-compute-runtime. No issues have been found so far.

Previously, rusticl (via the mesa-libOpenCL package) was included as the default for AMD GPUs, but this was found to break ROCm entirely, and so is no longer included in davincibox-opencl.

I do not have an NVIDIA GPU at this time and cannot directly troubleshoot issues with using them. Support and troubleshooting for NVIDIA-related issues is very reliant on members of the community.

DaVinci Resolve Studio

Davincibox has had limited testing with DaVinci Resolve Studio. Use at your own risk. See #26

Codecs

DaVinci Resolve on Linux, especially the free version, has limited codec support. Unless you can show that certain codecs are available when using DaVinci Resolve outside of davincibox, but not when using davincibox, do not report issues about missing codecs here.

Requirements & GPU Information

You will need Podman, as well as distrobox or toolbox.

You will also need the latest release of DaVinci Resolve from Blackmagic's website.

Your user in the host system may need to be a member of the groups 'render' and 'video'. You can add these with usermod -aG render,video USERNAME in the console of the host system.

If you're less comfortable in the CLI, I recommend using the setup.sh script from this repository to help simplify the setup process, but ultimately use of the CLI is a requirement.

Important: Ensure lshw is installed on your system, as setup.sh uses it to detect whether you are using an Nvidia GPU or not. This is pre-installed on Universal Blue images, but you may need to install it yourself on other Linux distributions.

AMD

AMD uses ROCm, and should generally work out of the box on modern AMD GPUs (RX 6000 series and above). However, some distros such as Debian may additionally require the rocm-podman-support package to be installed (on the host, not in davincibox).

GPU support in ROCm is very limited and it may not work on older AMD GPUs (RX 5000 series and prior). If the standard setup instructions do not result in a functional DaVinci Resolve, you can instead try using rusticl:

• Distrobox: distrobox enter davincibox
• Toolbox: toolbox enter davincibox

Then, from inside davincibox:

sudo dnf install mesa-libOpenCL

# If you exclude the path below, run-davinci defaults to launching DaVinci Resolve.
# If you want to launch one of the other programs included with Resolve,
# use the path to its binary as shown below.
run-davinci -c /path/to/binary

If rusticl does work better for you and you would like to launch Resolve with rusticl via the desktop shortcut, you can adjust the DaVinciResolve.desktop file in $HOME/.local/share/applications.

Change the Exec= line as follows:

• Distrobox: Exec=distrobox-enter -n davincibox -- /usr/bin/run-davinci -c /opt/resolve/bin/resolve %u
• Toolbox: Exec=/usr/bin/toolbox run -c davincibox /usr/bin/run-davinci -c /opt/resolve/bin/resolve %u

You can do the same for the other programs that come with DaVinci Resolve, too; just add the -c flag after /usr/bin/run-davinci.

Intel

For Intel GPUs, the intel-compute-runtime package is included and used by default.

If needed, rusticl can also be used via the same instructions in the AMD section above.

NVIDIA

NVIDIA users will need to install the nvidia-container-toolkit on their host OS. If you are using a Universal Blue image such as Bluefin, this will already be installed. Otherwise, see NVIDIA's installation guide for instructions for your distribution.

You may additionally need to add a SELinux policy to allow containers to use your GPU. See the "Adding the SELinux policy module" in this guide. Otherwise, you may get an "Unsupported GPU processing mode" error when launching Resolve.

Setup

CLI:

Open a terminal, then run chmod +x /path/to/setup.sh

Then, /path/to/setup.sh /path/to/DaVinci_Resolve_versionnumber_Linux.run

GUI:

If you're more comfortable in a GUI:

If you're using GNOME, open Files and navigate to where you downloaded the script to. In the example below, the script is in the same folder that I extracted the DaVinci Resolve download to. I recommend you do the same for ease of use, as the rest of the instructions will assume you have done so.

Right-click, and select Properties.

Then, make sure "Executable as Program" is toggled on.

Right-click on an empty spot in the folder. You should see either "Open in Console" as in the screenshot, or "Open in Terminal." Either will be fine.

In the newly-opened terminal window, enter the command below. Replace 'version' with the version of DaVinci Resolve that you are installing (see screenshot for example):

./setup.sh ./DaVinci_Resolve_version_Linux.run

Then, follow any further prompts in the installation script.

Manual

Setup Davincibox

First, get davincibox set up. There are two different builds of davincibox, depending on whether you use an NVIDIA GPU or not:

NVIDIA Users

Distrobox:

distrobox create -i ghcr.io/zelikos/davincibox:latest --additional-flags "--device nvidia.com/gpu=all" -n davincibox

Toolbox:

toolbox create -i ghcr.io/zelikos/davincibox:latest -c davincibox

Intel & AMD Users

Distrobox:

distrobox create -i ghcr.io/zelikos/davincibox-opencl:latest -n davincibox

Toolbox:

toolbox create -i ghcr.io/zelikos/davincibox-opencl:latest -c davincibox

Install DaVinci Resolve

On the host, run --appimage-extract on your DaVinci Resolve installer

/path/to/DaVinci_Resolve_version_Linux.run --appimage-extract

Then, run setup-davinci squashfs-root/AppRun distrobox/toolbox from within the container

e.g.

Distrobox:

distrobox enter davincibox -- setup-davinci squashfs-root/AppRun distrobox

Toolbox:

toolbox run --container davincibox setup-davinci squashfs-root/AppRun toolbox

The suffix at the end is for the add-davinci-launcher script. If omitted, setup will still run, but adding the launcher to your application menu won't work.

You can still run add-davinci-launcher separately, as either add-davinci-launcher distrobox or add-davinci-launcher toolbox, depending on what you're using.

After installation completes, you can remove the squashfs-root directory.

After setup, run sudo dnf update in the container to ensure drivers are up to date:

Distrobox:

distrobox enter davincibox -- sudo dnf update

Toolbox:

toolbox run -c davincibox sudo dnf update

Upgrading

Usually, you'll only need to update the packages in the container (i.e. driver updates), not the container image itself. In this case, simply run sudo dnf update in davincibox as shown above.

Otherwise, a full upgrade requires re-creating the davincibox container with the newest version of the image. This should only needed if there has been a new GitHub release for davincibox, as this may include