利用curl_multi_init并行多线程处理网络数据

for ($i=0; $i < 100; $i++)
{
      $urls[] = "http://blog.youkuaiyun.com/lg_lin?wd=".mt_rand(10000,20000);
}


$t = microtime(true);
//单线程
foreach ($urls as $key => $value)
{

    $ch = curl_init();
    curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, FALSE);
    curl_setopt($ch, CURLOPT_RETURNTRANSFER, TRUE);
    curl_setopt($ch, CURLOPT_TIMEOUT, 30);
    curl_setopt($ch, CURLOPT_URL, $value);
    $response =  curl_exec($ch);

    curl_close($ch);


    $result_new[$key] = $response;
}
var_dump($result_new);

$e = microtime(true);
echo "For循环：".($e-$t)."\n";

/**
 **** 多线程****
 * 由于受网络带宽限制，一次性不能并发太多，可以分批处理
 * 总量超过100以后，效果就比单线程明显提高
*/

$p = 5;//每次执行多少条
$ring = ceil(count($urls) / $p);

for($n = 0; $n < $ring; $n++)
{
    $temp_url = array();
    $star = $n * $p;
    $end = ($n+1) * $p;
    for($ii = $star; $ii < $end; $ii++)
    {
        if (isset($urls[$ii]))
        {
            $temp_url[] = $urls[$ii];
        }

    }

    $mh = curl_multi_init();
    foreach ($temp_url as $i => $url) {
        $conn[$i] = curl_init($url);
        curl_setopt($conn[$i], CURLOPT_USERAGENT, $_SERVER['HTTP_USER_AGENT']);
        curl_setopt($conn[$i], CURLOPT_SSL_VERIFYPEER, FALSE);
        curl_setopt($conn[$i], CURLOPT_RETURNTRANSFER, TRUE);
        curl_setopt($conn[$i], CURLOPT_TIMEOUT, 30);
        curl_setopt($conn[$i], CURLOPT_URL, $url);
        curl_multi_add_handle($mh,$conn[$i]);
    }

    do{
        curl_multi_exec($mh, $active);
    } while ($active);

    $active = null;

    foreach ($temp_url as $i => $url) {
        $data[$i] = curl_multi_getcontent($conn[$i]); // 获得爬取的代码字符串

    }

    foreach ($temp_url as $i => $url) {
        curl_multi_remove_handle($mh, $conn[$i]);
        curl_close($conn[$i]);
    }

    curl_multi_close($mh);


    var_dump($data);

}

$t = microtime(true);

echo "多线程：".($t-$e)."\n";


exit();