[Cake] Test results

[Dave Taht]

up until a few moments ago I had assumed that cake was peeling...

and I was composing an email talking about how all cake's optimizations
seemed not to be any better, and in some cases, worse, than sqm-scripts
in all the test series with offloads on I ran last week. Now I know....

I was prepping to rip out a lot of the stuff that we have not tested at
scale - like the invsqrt cache, most of the other mods to codel, and then
take a hard look at the target and quantum calculations.

one thing I think would be helpful to everyone is to start publishing
test results, as well as perf results for arm, mips, and x86_64.

But I am not sure what people would find useful.