They recently released a huge bump of their AI package, they probably wanted to differentiate between the older and newer version...still they could have renamed the plugin, or at very least, mark older comments as "for an older version"
I actually pushed a fix for the soundcard upstream back in kernel 5.13, it had been working great. And yeah, don't expect a Wacom level pen, it doesn't come with tilt. I never trust/tried sleep on Linux(could corrupt the FS), but everything is working great on my side.
The RTX 4080 should be capable of ~40 TFLOPS, yet they only report 2,160 billion operations per second. Shouldn't this be enough to reconsider the benchmark?
They probably made some serious error in measuring FLOPS.
Regarding the fact that CPU beats NPU is possible but they should benchmark many matrix multiplications without any application synchronization in order to have a decent comparison.
Both works pretty well