Benchmark: Pimple vs Symfony/DependencyInjection

composer.json

{
    "require": {
        "pimple/pimple": "1.0.*",
        "symfony/dependency-injection": "2.1.*"
    }
}

RESULTS

$ ./run-tests.sh 
Pimple
265K
579K
0.00086s
DependencyInjection
267K
931K
0.006124s
DependencyInjection (dumped)
688K
738K
0.000495s

run-tests.sh

#!/bin/bash

FLAGS="-d xdebug.max_nesting_level=300"

echo "Pimple"
php $FLAGS test-pimple.php

echo "DependencyInjection"
php $FLAGS test-di.php

echo "DependencyInjection (dumped)"
php $FLAGS test-di-dumped.php

service.php

<?php

class Service {
    private $service;
    private $param;

    public function __construct(Service $service, $param)
    {
        $this->service = $service;
        $this->param = $param;
    }
}

test-di-dumped.php

<?php

require 'vendor/autoload.php';
require 'service.php';
require 'dumped.php';

$start = microtime(true);
echo intval(memory_get_usage() / 1024).'K'."\n";

$container = new ProjectServiceContainer();

echo intval(memory_get_usage() / 1024).'K'."\n";
echo round(microtime(true) - $start, 6).'s'."\n";

test-di.php

<?php

require 'vendor/autoload.php';
require 'service.php';

use Symfony\Component\DependencyInjection\ContainerBuilder;
use Symfony\Component\DependencyInjection\Reference;

$start = microtime(true);
echo intval(memory_get_usage() / 1024).'K'."\n";

$container = new ContainerBuilder();
$container->setParameter('param', 'foo');
$container
        ->register('service_1', 'Service')
        ->addArgument(null)
        ->addArgument('%param%');

foreach (range(2, 200) as $i) {
    $container
        ->register('service_'.$i, 'Service')
        ->addArgument(new Reference('service_'.($i - 1)))
        ->addArgument('%param%');
}

echo intval(memory_get_usage() / 1024).'K'."\n";
echo round(microtime(true) - $start, 6).'s'."\n";

use Symfony\Component\DependencyInjection\Compiler\Compiler;

$compiler = new Compiler();
$compiler->compile($container);

use Symfony\Component\DependencyInjection\Dumper\PhpDumper;

$dumper = new PhpDumper($container);
file_put_contents('dumped.php', $dumper->dump());

test-pimple.php

<?php

require 'vendor/autoload.php';
require 'service.php';

$start = microtime(true);
echo intval(memory_get_usage() / 1024).'K'."\n";

$container = new Pimple();
$container['param'] = 'foo';
$container['service_0'] = null;

foreach (range(1, 200) as $i) {
    $container['service_'.$i] = $container->share(function ($container) use ($i) {
        return new Service($container['service_'.($i - 1)], $container['param']);
    });
}

echo intval(memory_get_usage() / 1024).'K'."\n";
echo round(microtime(true) - $start, 6).'s'."\n";

The code for the DI test is wrong. It does not run the compiler passes in the benchmark code, thus taking into account only part of the building

and this does not benchmark the use of the container, only its creation

Compilation is an optional step. I did not want to benchmark it in this case. But fair enough, benchmarking with compilation and actually accessing a service would be interesting things to benchmark as well.

which container should I choose?