Axion Tech.

Ubuntu 26.10 on Arrow Lake: video encoders get faster, compile jobs get slightly slower

Phoronix re-ran dozens of tests on a Core Ultra 9 285H laptop, Ubuntu 26.04 LTS versus 26.10 (Linux 7.0 to 7.3, Mesa 26.0 to 26.2). SVT-AV1, OpenAPV and x265 win; kernel, Godot and FFmpeg builds lose.

By Axion Tech Desk · · Source: Phoronix

Typographic card

Ubuntu 26.10 release is providing some nice performance gains over the Ubuntu 26.04 LTS release.

The question every LTS owner asks after a new interim release is whether the upgrade pays. Phoronix tested it on a Star Labs StarFighter Ultra with an Intel Core Ultra 9 285H (Arrow Lake H), Arc 140T integrated graphics, 64 GB of LPDDR5-7467, a 1 TB Samsung 990 EVO Plus NVMe drive and a 3840 by 2400 display, running coreboot firmware. The only variable was the stock Ubuntu release: 26.04 LTS ships Linux 7.0, Mesa 26.0 and GNOME 50; 26.10 ships Linux 7.3, Mesa 26.2 and GNOME 51.

Where 26.10 wins

According to the Phoronix chart captions, Ubuntu 26.10 was faster in SVT-AV1 at Preset 3 for both 1080p and 4K Bosphorus encodes, in OpenAPV's fastest and medium 4K 10-bit presets, in x265 1080p (described as slightly better), and in the Blender Barbershop CPU scene. The review calls the video-encoding gains nice, though the pages we read do not quantify the uplift; the percentages live in the charts, so read them there before quoting a number.

Where it loses

Ubuntu 26.04 LTS stayed ahead in the Blender BMW27 CPU scene, Linux kernel compiles (defconfig and allmodconfig), Godot Engine and FFmpeg builds, and the OpenAPV Fast preset at 1080p. Phoronix says compilation was slightly slower on 26.10 and attributes this in part to GCC 15.3 being the default compiler there rather than GCC 16. Blender overall showed nothing notable.

What to do with it

The pattern is useful: media-encoding workloads benefit from the newer kernel and libraries, while build-heavy workloads did not. That is a mixed result for a developer laptop, and a reason to be deliberate. Because 26.10 is an interim release with a shorter support window, the LTS holds its ground on stability, and the gains here are workload-specific rather than universal. Graphics results sit on later pages of the review and were not part of our reading.

Comprehensive take

Video and creative workloads on Arrow Lake laptops should consider 26.10; developers who compile all day should stay on 26.04 LTS until GCC 16 lands. Do not upgrade a work machine for a few encoder gains without testing your own build times first.

Use cases

  • Video encoders and editors

    SVT-AV1, OpenAPV and x265 results favoured 26.10 in the tested presets.

  • Developers compiling large projects

    Kernel, Godot and FFmpeg builds were slightly slower on 26.10; the LTS is the safer pick for dev laptops and build machines.

  • Star Labs StarFighter and other Arrow Lake H owners

    The test platform is a StarFighter Ultra with a Core Ultra 9 285H, so results map directly to that class of machine.

Specs

Test machineStar Labs StarFighter Ultra · Core Ultra 9 285H · Arc 140T · 64 GB LPDDR5-7467 · 1 TB 990 EVO Plus source
Ubuntu 26.04 LTS stackLinux 7.0 · Mesa 26.0 · GNOME 50 source
Ubuntu 26.10 stackLinux 7.3 · Mesa 26.2 · GNOME 51 · GCC 15.3 default source

Insights

  • Encoder workloads gained; compile workloads lost. One upgrade, opposite outcomes depending on your day job.
  • The compile slowdown is attributed in part to GCC 15.3 being the 26.10 default instead of GCC 16.
  • Only the OS stack changed between runs, so differences come from kernel, Mesa and toolchain, not hardware.

Axion Scorecard

Overall: Not yet scored

Performance 6/10

Encoder tests faster, kernel/Godot/FFmpeg compiles slightly slower; magnitudes not captured in the pages we read. source

Value Not yet scored

Not yet scored — no cited data

Efficiency Not yet scored

Not yet scored — no cited data

Innovation Not yet scored

Not yet scored — no cited data

Ecosystem Not yet scored

Not yet scored — no cited data

Availability Not yet scored

Not yet scored — no cited data

Build & Design Not yet scored

Not yet scored — no cited data

Editorial opinion; weights shown: Performance ×2, Value ×1.5, Efficiency ×1, Innovation ×1.5, Ecosystem ×1.5, Availability ×1, Build & Design ×1.

Community

Share (0): XRedditLinkedinEmail

Comments

No comments yet.

Log in or register to like or comment.