gadwinn / benchmark
Tiny development tool comparing the execution time and memory use of pieces of code.
v0.1.0
2026-11-17 15:00 UTC
Requires
- php: >=8.3
Requires (Dev)
- phpdocumentor/shim: *
- phpstan/phpstan: *
- phpunit/phpunit: *
- povils/phpmnd: *
- roave/security-advisories: dev-latest@dev
- squizlabs/php_codesniffer: *
Suggests
None
Provides
None
Conflicts
None
Replaces
None
This package is auto-updated.
Last update: 2026-11-18 11:12:20 UTC
README
A tiny development tool to compare the execution time and memory use of pieces of code.
Requirements
- PHP >= 8.3
Installation
composer require --dev gadwinn/benchmark
Usage
use G\Benchmark\Benchmark; $data = range(1, 10_000); (new Benchmark(iterations: 1_000)) ->addCallable('array_sum', 'array_sum', $data) ->addCallable('foreach', static function (array $numbers): int { $sum = 0; foreach ($numbers as $number) { $sum += $number; } return $sum; }, $data) ->display();
TOP 1 - Benchmarking "array_sum":
Total time (ns): 6.2147e+6
Total time (s): 0.0062147
Iterations: 1000
Time per iteration (ns): 6214.7
Time per iteration (s): 6.2147e-6
Memory peak usage (B): 0
Memory peak usage (MiB): 0
--------------------------------------------------------------------------------
TOP 2 - Benchmarking "foreach":
...
Each callable is called iterations times (10 by default) with the arguments given after it. The report lists the callables from the fastest to the slowest; it is coloured in a terminal and rendered as HTML in a browser.
| Method | Description |
|---|---|
addCallable(string $name, callable $callable, mixed ...$args): self |
Registers a piece of code and its arguments |
display(): void |
Runs the benchmark and prints the report |
report(): string |
Runs the benchmark and returns the report |
run(): Generator |
Runs the benchmark and yields one raw result per callable, in the order they were added |
getCallables(): array |
The registered callables |
Raw results
foreach ($benchmark->run() as $result) { // [ // 'callable_name' => 'array_sum', // 'total_time_ns' => 6214700, // time spent inside the callable, all iterations // 'iterations' => 1000, // 'time_per_iteration_ns' => 6214.7, // 'memory_peak_usage_b' => 0, // ] }
Reading the figures
- Time only counts what happens inside the callable; the loop around it is not measured.
- Memory is the highest amount of memory a single call needed on top of what was already allocated when it started. It is the maximum over the iterations, not a sum or an average.
- Timings vary with the machine load, OPcache and the JIT: compare results obtained in the same run, and raise
iterationsfor short pieces of code.
Development
composer install composer test-all
License
MIT