For the specific RX 6400 versus Intel Arc A310 choice in this ZimaOS/Jellyfin thread, the Arc A310 is the stronger media-transcoding fit. The reason is not that AMD GPUs are universally unsupported on ZimaOS; that claim appeared in an early reply and was challenged in the same thread. The more durable difference is codec capability and Jellyfin’s current Linux acceleration support.
The user ultimately installed the A310, went through VA-API/QSV configuration, and later showed real intel_gpu_top video-engine activity while FFmpeg was transcoding. That is much stronger evidence than whether the ZimaOS dashboard displayed a GPU widget.
Why the Arc A310 Made More Sense Than the RX 6400
Jellyfin’s current Intel GPU acceleration guide says QSV is preferred on mainstream Intel GPUs and explicitly supports Arc A-series hardware on Linux. Arc A-series also supports AV1 encoding, which is valuable for a dedicated media-transcoding card.
The current ZimaOS GPU compatibility guide lists Intel A310, A380, A580, A750 and A770 in its Intel A-series compatibility table.
Direct Play Can Make a Working GPU Look Idle
One of the most useful corrections in the thread was that local playback may Direct Play the source file. In that case, Jellyfin has little reason to engage the GPU heavily. To validate hardware transcoding, deliberately trigger a lower bitrate, incompatible codec path, subtitle burn-in or another condition that forces conversion.
QSV and VA-API Are Both Relevant on Linux
The community moved from VA-API toward Intel Quick Sync. That matches current Jellyfin guidance: QSV is generally preferred on supported mainstream Intel hardware, while VA-API remains available and is important for older or compatibility-focused paths.
Do not select every codec checkbox blindly. Enable only formats the hardware and driver stack actually expose.
Restart Jellyfin After Changing Acceleration Settings
Several settings in Jellyfin are applied only after the server process restarts. The user also enabled hardware acceleration for trickplay, which can reduce CPU load while generating preview images but adds another GPU workload during library processing.
A Blank QSV Device Field Does Not Prove Failure
The user worried because Jellyfin did not display the QSV device name the same way another system did. Community replies suggested leaving the field blank and testing actual transcoding behavior instead of diagnosing from the UI label alone.
Verify the GPU at the Host and During a Real Transcode
The strongest evidence in the thread was runtime activity: intel_gpu_top showed the Video engine busy while an FFmpeg process was active. That proves more than a static device list because it shows the media engine doing work at the moment of the transcode.
The hardware acceleration guide explains the same Direct Play versus transcode boundary, while the Jellyfin hardware guide helps decide whether a dedicated GPU is needed at all.
App Source and Version Still Matter
The thread also touched on App Store source categories and later Jellyfin versions. A working host GPU does not guarantee every container image exposes /dev/dri or includes the same FFmpeg/media libraries. When acceleration fails after an app update, compare the container definition and device mappings before blaming the card.
Browser Playback Is a Separate Compatibility Layer
The discussion about DTS and Dolby audio is separate from GPU detection. A browser may force transcoding or fail to pass through an audio format that a dedicated Jellyfin client can handle directly. Use the client type and playback information when deciding whether a “GPU problem” is actually a media-client compatibility problem.
What the User Eventually Achieved
By the end of the thread, the user reported that transcoding was working well overall. The configuration journey still exposed rough edges—missing UI labels, older media behaving differently and some uncertainty around settings—but the A310 was doing real video work.
Bottom Line
For a compact Jellyfin transcoding card on current Linux/ZimaOS, Intel Arc A310 is a defensible choice because Jellyfin supports QSV/VA-API on Arc and ZimaOS currently lists the A310 in its Intel compatibility table. Verify success with an actual forced transcode and GPU-engine activity, not with the ZimaOS dashboard widget alone.
