Some Quick Graphics/Game Tests With GNOME 3.32 On Clear Linux
Using a Radeon RX Vega 56 graphics card, I tested Clear Linux between its releases having GNOME Shell 3.30 and the move to GNOME Shell 3.30.2. On both builds of Clear Linux, Linux 5.0.1 was in use along with X.Org Server 1.20.4 (they aren't yet defaulting to a Wayland session), and Mesa 19.1-devel.
I ran just some quick graphics/gaming benchmarks both windowed and full-screen to see how the performance has changed under GNOME 3.32. GNOME 3.32 does have responsiveness improvements and other performance enhancements that are difficult to quantify in benchmarks while today's tests are looking at just any raw performance changes to these games/apps. In using GNOME 3.32 already for several days on Clear Linux, from a qualitative side it's much better than GNOME 3.30 due to the optimizations and other fixes that went in over the past half year.
Personally I remain quite happy with GNOME 3.32 and it's a nice evolutionary improvement over 3.30.
Some OpenGL games are performing slightly faster under GNOME 3.32.
In demanding cases though the results are virtually unchanged.
The ParaView workstation software was edging higher but regressing in the Wavlet Volume sub-test.
There were only minor changes with the Java graphics performance.
From the limited tests done, there wasn't much to note but that isn't unexpected since the optimizations going into GNOME 3.32 don't translate into raw gaming/graphics throughput performance. From using GNOME 3.32, users though should feel a better experience if you're a heavy GNOME user and have run into various bottlenecks and shortcomings with the previous releases. The GNOME 3.32 experience on Clear Linux has so far been great and without any problems in my encounters.
When Ubuntu 19.04 / Fedora 30 approaches, I'll be back with some cross-desktop benchmarks and other comparison fun.
4 Comments
