Home  >  Article  >  Backend Development  >  Use XHProf to find instances of PHP performance bottlenecks

Use XHProf to find instances of PHP performance bottlenecks

小云云
小云云Original
2017-12-14 10:02:121652browse

Performance bottlenecks generally refer to newly developed applications by developers.

For example,

An application developed in Java or C and deployed on an application server for processing user transaction requests.

For example,

A developer developed a payment processing program. During testing, it was discovered that

this payment processing program was unable to process concurrent payment requests sent by users. , can only be processed serially and cannot be processed in parallel, resulting in a very long processing response time for payment transactions. At this time, it can be considered that a performance bottleneck has occurred in the application.

This article will share with you an example of using XHProf to find PHP performance bottlenecks. It has a good reference value and I hope it can help everyone.

XHProf is an extension developed by Facebook to test PHP performance. This article records the method of using XHProf to optimize PHP performance and find performance bottlenecks in PHP applications.

1. Install Xhprof extension

//github上下载https://github.com/facebook/xhprof
unzip xhprof-master.zip 
cd xhprof-master/extension/
/usr/local/php/bin/phpize
./configure --with-php-config=/usr/local/php/bin/php-config --enable-xhprof
make && make install

2. Modify php.ini

[xhprof]
extension=xhprof.so
xhprof.output_dir=/tmp

In the configuration, xhprof.output_dir specifies the location where the generated profile file is stored, we will specify it for /tmp.

3. Move the relevant files into the project

//xhprof下载压缩包中的xhprof_html和xhprof_lib
cp -r xhprof-master/xhprof_html /usr/local/nginx/html/xhprof/
cp -r xhprof-master/xhprof_lib /usr/local/nginx/html/xhprof/

Configure a domain name, and the browser can access http://will.com/xhprof/xhprof_html/index.php

server{
 listen 80;
 server_name will.com;
 location / {
  root /usr/local/nginx/html;
  index index.html;
 }
 location ~ \.php$ {
  root html;
  fastcgi_pass 127.0.0.1:9000;
  fastcgi_index index.php;
  fastcgi_param SCRIPT_FILENAME $document_root$fastcgi_script_name;
  include  fastcgi_params;
 }
 }

4. Install graphivz

//需要安装graphviz否则查看性能图片时候会报failed to execute cmd: " dot -Tpng". stderr: `sh: dot: command not found '
yum -y install graphviz

5. Write test files

//入口文件的开始位置
xhprof_enable(XHPROF_FLAGS_MEMORY | XHPROF_FLAGS_CPU);

业务逻辑...

//业务逻辑结束后
$xhprof_data = xhprof_disable();
include_once "/usr/local/nginx/html/xhprof/xhprof_lib/utils/xhprof_lib.php"; 
include_once "/usr/local/nginx/html/xhprof/xhprof_lib/utils/xhprof_runs.php"; 
$objXhprofRun = new XHProfRuns_Default();//数据会保存在php.ini中xhprof.output_dir设置的目录去中 
$run_id = $objXhprofRun->save_run($xhprof_data, "test");

Complete code example (random red envelope demo)

<?php
xhprof_enable(XHPROF_FLAGS_MEMORY | XHPROF_FLAGS_CPU);
function show($info)
{
 echo "<pre class="brush:php;toolbar:false">";
 print_r($info);
}

//不作数据校验
$rules = array(
 2=>array('min'=>1, 'max'=>10, 'chance'=>30),//金额:分 概率:百分之(默认为100%,不足100%按第一档计算)
 array('min'=>11, 'max'=>25, 'chance'=>60),
 array('min'=>26, 'max'=>50, 'chance'=>10),
 array('min'=>50, 'max'=>80, 'chance'=>0),
 array('min'=>80, 'max'=>100, 'chance'=>0),
);
$total_money = 10000;//红包总金额
$res = array();
while($total_money>0)
{
 $index = getLevel($rules);
 $money = setMoney($rules, $index);
 if ($money > $total_money)//金额不足
 {
 $money = $total_money;
 $total_money = 0;
 } else {
 $total_money -= $money;
 }
 $res[] = ($index+1)."---".$money;
}
echo show($res);
echo $total_money . "<br/>";
//1.先确定档次
function getLevel($rules)
{
 $level = array();
 $chance = 0;
 foreach($rules as $k=>$v)
 {
 if ($v['chance']>0)
 {
  $chance += $v['chance']*100;//扩大100倍
  $level[$k] = $chance;
 }
 }
 $index = 0;
 $rand_num = mt_rand(1, 10000);
 foreach($level as $k=>$v)
 {
 if ($rand_num <= $v)
 {
  $index = $k;
  break;
 }
 }
 return $index;
}
//2.确定档次之后,再确定金额
function setMoney($rules, $index)
{
 $money = mt_rand($rules[$index][&#39;min&#39;]*10000, $rules[$index][&#39;max&#39;]*10000)/10000;
 $money = ceil($money);
 $money > 1 && $money = $money -1;//防止出现免单情况
 return $money;
}
$xhprof_data = xhprof_disable();
include_once "/usr/local/nginx/html/xhprof/xhprof_lib/utils/xhprof_lib.php"; 
include_once "/usr/local/nginx/html/xhprof/xhprof_lib/utils/xhprof_runs.php"; 
$objXhprofRun = new XHProfRuns_Default();//数据会保存在php.ini中xhprof.output_dir设置的目录去中 
$run_id = $objXhprofRun->save_run($xhprof_data, "test");
echo "http://will.com/xhprof/xhprof_html/index.php?run=$run_id&source=test";//变量$runId是本次请求生成分析结果的id,最后我们输出了一个链接地址,使用改地址就可以看到本次请求的分析结果。

6. View analysis results

Run the business code first;

Then the browser opens http://will.com/xhprof/xhprof_html/index.php, click the last time to generate the xhprof file

Note the View Full Callgraph link in the middle, through which we can see the graphical analysis results

The red part in the figure indicates low performance , the part that takes a long time, we can optimize the system code according to which functions are marked in red

In addition, the meaning of the xhprof report field:

Function Name: method name .

Calls: The number of times the method has been called.

Calls%: The number of method calls as a percentage of the total number of method calls at the same level.

Incl.Wall Time(microsec): The time it takes for method execution, including the execution time of sub-methods. (Unit: microseconds)

IWall%: The percentage of time spent in method execution.

Excl. Wall Time (microsec): The time it takes to execute the method itself, excluding the execution time of sub-methods. (Unit: microseconds)

EWall%: The percentage of time spent executing the method itself.

Incl. CPU(microsecs): The CPU time spent on method execution, including the execution time of sub-methods. (Unit: microseconds)

ICpu%: The percentage of CPU time spent in method execution.

Excl. CPU (microsec): The CPU time spent executing the method itself, excluding the execution time of sub-methods. (Unit: microseconds)

ECPU%: The percentage of CPU time spent executing the method itself.

Incl.MemUse(bytes): The memory occupied by method execution, including the memory occupied by sub-method execution. (Unit: bytes)

IMemUse%: The percentage of memory occupied by method execution.

Excl.MemUse(bytes): The memory occupied by the execution of the method itself, excluding the memory occupied by the execution of sub-methods. (Unit: bytes)

EMemUse%: The percentage of memory occupied by the method itself.

Incl.PeakMemUse(bytes): Incl.MemUse peak value. (Unit: Bytes)

IPeakMemUse%: Incl.MemUse peak percentage.

Excl.PeakMemUse(bytes): Excl.MemUse peak value. Unit: (byte)

EPeakMemUse%: Excl.MemUse peak percentage.

Related recommendations:

JavaScript startup performance bottleneck analysis and solutions

Use xdebug to analyze the PHP program to find out the performance Bottleneck

Tracking PHP code performance bottleneck

The above is the detailed content of Use XHProf to find instances of PHP performance bottlenecks. For more information, please follow other related articles on the PHP Chinese website!

Statement:
The content of this article is voluntarily contributed by netizens, and the copyright belongs to the original author. This site does not assume corresponding legal responsibility. If you find any content suspected of plagiarism or infringement, please contact admin@php.cn