Such benchmarks only test the default configuration, right? And ignores any optimizations one would do.
Yes, since with optimizations on OS level and with custom kernels you can pretty much tune it to any other result.
Differences are inconsequential aside from image manipulation, asyncio_websockets, librsvg, and AV1.
the yolo distro, using their users as guinea pigs to test patches that haven’t even gone upstream yet
I’m always interested in Micheal’s comparisons but rarely see anything more than an illustration of newer libraries etc showing natural improvements. The trade off of Arch distros is the increased workload of managing a constant change & inevitable instability. Arch devs are notoriously for kicking out capricious system borking changes and the Pacman package manager is rather weak at dealing with cumulative changes. 2% or 3 % potential ephemeral improvements in speed vs hovering over the cli ‘fixing’ things seems a poor bargain to me.
