Getting started¶

Requirements¶

• Python 3.12 or newer
• No required runtime dependencies

Install the core package:

pip install benchcore

Install the optional pytest dependency:

pip install "benchcore[pytest]"

Run a configured benchmark¶

from benchcore.core.models import BenchmarkConfig
from benchcore import Bench

bench = Bench(
    BenchmarkConfig(
        rounds=10,
        warmup_rounds=2,
    )
)
result = bench.run(sorted, [5, 4, 3, 2, 1])

print(result.value)
print(result.iterations)
print(result.statistics.mean_ns)

Automatic calibration is enabled when iterations is None. Set a positive integer to reproduce a fixed workload:

config = BenchmarkConfig(rounds=20, iterations=100, warmup_rounds=3)

Create and render a report¶

from benchcore.reporting import TerminalReporter, create_report

report = create_report("sorted-small", result)
print(TerminalReporter().render(report))

Reporting happens after measurement and never inside the timed region.