Benchmark Comparison & Statistics Builder — Free
Runs 100% in your browser — your files never leave your device.
Raw benchmark samples do not make a useful report until they are summarized consistently. Add any number of named suites, paste measurements one per line or comma-separated, and compare mean, population variance, sample count, delta, and ratio against the fastest suite. Export the ranked result as Markdown.
How it works
- 1
Add benchmark suites
Name each contender and paste its measurement samples.
- 2
Choose a unit
Milliseconds, operations per second, or any label works.
- 3
Compare and export
Review ranked statistics and copy the Markdown table.
About this tool
What the tool compares
Enter two measured durations and the tool reports the difference, the ratio, and the percentage change — the honest arithmetic behind "how much faster is it". Benchmark claims live or die on exactly these three numbers.
Reading benchmark results without fooling yourself
A 2x ratio means half the time; a 95 percent reduction is not "nearly the same" — percentages hide scale. Report the raw times alongside the ratio, and measure with the stopwatch rather than remembering: human time estimates skew optimistic by large factors.
Fair benchmarks share everything else
The comparison is only as fair as its controls: same input data, same machine load, same measurement method. A ratio computed across different inputs measures the inputs as much as the implementations — control the variables before quoting the number.
Frequently asked questions
How many runs should I benchmark before comparing?
Several, and compare the minimum or median rather than a single run — caches, background load, and warm-up make first runs slower. The tool compares the durations you measured.
What does the percentage change mean for speed?
A 50 percent reduction in time equals a 2x speedup — percentage-of-time and ratio are different descriptions of the same result. Report both to avoid confusion.
Can I compare durations measured in different units?
Enter both in the same unit — convert first so the ratio and difference compare like with like.