Mastodon Feed
dysfun@treehouse.systems ("gaytabase") wrote:
holy shit, i just doubled perf of my benchmark!
this isn't because it got cheaper (it's already bordering on optimal!), it's because the benchmark was adding too much overhead (it was a realistic benchmark rather than synthetic, thus stuff was having to be saved and i wasn't measuring what i was trying to measure!)
i had a rough target of 'under 10ns on this machine'. well i haven't quite gotten there, but i'm under 11 ns, so not too bad.