Some numbers with the microbenchmark:
$ for i in $(seq 1 100); do dd if=/dev/zero of=/dev/null bs=16K count=10000 2>&1 | grep GB; done > /tmp/w.log
$ awk 'BEGIN { sum = 0 } { sum +=$10 } END { print sum/100 }' /tmp/w.log
on AMD zen3
original: 20.11 Gb/s
rep_good: 34.662 Gb/s
erms: 36.378 Gb/s
fsrm: 36.398 Gb/s
I'll run some real benchmarks later but these numbers kinda speak for
themselves so it would be unlikely that it would show different perf
with a real, process-intensive benchmark...
--
Regards/Gruss,
Boris.
https://people.kernel.org/tglx/notes-about-netiquette
