We have benchmarks for the floating-point formatting algorithms themselves, but not for the surrounding machinery like Formatter and translating to the flt2dec::Part slices.
And libcore/benches