Bug: Resolve seemingly uses software rendering for the UI

[TibiIius]

This issue is NOT about missing codecs

• I understand

Describe the bug

I installed Davincibox using Docker as my container engine of choice. I run a slightly modified version of Fedora Silverblue 43, using the uBlue Silverblue NVIDIA image as base.

When using Resolve, the GUI is seemingly rendered using software rendering. It's painfully slow to operate, and the timeline preview does not work. Also, every pane resize causes the CPU usage to spike up to 80+%.

Rendering using the GPU works fine (tested with a small AV1 clip using NVIDIA's hardware transcoding).

Steps to reproduce

No response

Installation method used

Manual with Distrobox

OS Version

Fedora Silverblue 43 (custom image based on uBlue Silverblue NVIDIA)

Desktop Environment

GNOME 49

GPU Vendor

NVIDIA

GPU Model

RTX 5050

CPU

Ryzen 3 3100

CPU w/ integrated GPU?

No iGPU

DaVinci Resolve Free or DaVinci Resolve Studio

Free

DaVinci Resolve version

20

Davincibox build (NOT the GitHub release number)

ghcr.io/zelikos/davincibox:latest 2e8908cb0213 2.23GB 584MB U

Extra context

Davinci Resolve frequently reports that my system is out of video memory, even though this is not the case (using just under 2GB out 8GB VRAM).

[zelikos]

I installed Davincibox using Docker as my container engine of choice.

davincibox is developed and tested only with Podman as the container engine; any others are completely untested and unsupported by me.

That being said, we've had a persistent issue with NVIDIA GPUs, so if this issue isn't due to using Docker instead of Podman, it's probably related to the aforementioned issue.

[TibiIius]

I installed Davincibox using Docker as my container engine of choice.

davincibox is developed and tested only with Podman as the container engine; any others are completely untested and unsupported by me.

That being said, we've had a persistent issue with NVIDIA GPUs, so if this issue isn't due to using Docker instead of Podman, it's probably related to the aforementioned issue.

I tried setting up a Podman container instead, and also tried the fixes mentioned in the issue above, but that sadly did not help. I instead opted to just use an unofficial Flatpak build, which seems to work just fine. Feel free to close the issue if you want, but we may want to leave it open in case other users have the same problem as I do.