search

IR0NSIGHT / fasteffekt

A benchmarking tool for performance of effekt-language based
Other
0 stars 0 forks source link

use sensible problem sizes #40

Open IR0NSIGHT opened 9 months ago

IR0NSIGHT commented 9 months ago

IN ORDER TO have benchmarks that all take approx. the same amount of time WE WANT TO use larger problem sizes so that a single run of benchmark.normalRun takes ~2 seconds

discussion point: if we loop a benchmark over and over to increase its benchTime, does that impact the data we get? wouldnt it be better to just run once and leave it at that?

option: zu schnelle benchmarks in JS künstlich aufblasen durch inner iterations ODER neue verified ergebnisse finden

IR0NSIGHT commented 9 months ago

Decision with @phischu: leave "innerIterations" at one, accept that some microbenchmarks return 0 or 1 millis as benchTime that way we dont "verfälschen" the ergebnisse of the benchmarks