mfontani / perl-benchmark

What?

A benchmark for perl taint/no-taint and threaded/not threaded

How?

Using docker, on a Linux box:

# Prepare the repository:
perl ./dockerfiles.pl
./build.sh
# Run the images & produce report in bench.tsv:
make
# Analyze the report:
sort -rnk 4 < bench.tsv | column -t

Better reports

Produce multiple bench.tsv, to average them out and show error bars.

Maybe run this on various machines, or on the same machine a number of times - doable by running make -B. Example:

mkdir -p reports
# Create the first report
make
# Save it / stash it away
cp bench.tsv reports/bench1.tsv
# Create anoher...
make -B
# Stash that away...
cp bench.tsv reports/bench2.tsv
# Yet another...
make -B
# and stash that one away, too:
cp bench.tsv reports/bench3.tsv

Once done, run report.pl to produce a report on report/bench*.tsv on averages, with error bars. The report will be available in report/index.html:

perl ./report.pl
$BROWSER report/index.html