Measured runs

Every run, with its status stated up front. A number here is only as good as the benchmarks that produced it, so open a run to see what they covered.

REGRESSEDfaster on the benchmarks, slower elsewhere CAVEATEDa win with a condition attached FIXTUREour own demo code, written slow on purpose
  1. GOdustin/go-humanizeCAVEATED geomean over the declared benchmarks · one of three keeps was spurious −40.2% ns/opfirst commit slower on NaN, ±Inf
  2. JAVAstleary/JSON-javaREGRESSED on four JMH benchmarks · not sent upstream −25.1% ns/opslower on other inputs
  3. RSunicode-rs/unicode-segmentationREGRESSED on three criterion benchmarks · not sent upstream −57.1%slower on other inputs
  4. SWapple/swift-asn1CAVEATED unreviewed unsafe code · nothing contributed upstream 10.98× fasterunsafe, unreviewed
  5. TSautor3search-typescript demoFIXTURE own demo code, written slow on purpose −92.55%
  6. JSautor3search-javascript demoFIXTURE own demo code, written slow on purpose −54.00%

Results hold only for the benchmarks a run measures. Inputs they do not cover can get slower. Check your hot paths before you merge.