Performance test comparison between PHP framework jsnpp and thinkphp6

Studying recently Jsnpp frame , This framework is not only syntactically simple and easy to use , And the initial feeling is that the execution speed is quite fast , To confirm , So I did this test .

Obviously , The test has a frame of reference , Testing without reference is meaningless , Because the hardware environment is different , Then the test results are different . So first you have to select a reference system , because Jsnpp The frame is PHP The framework of language implementation , So use c The framework of language extension cannot be used as a frame of reference , You need to choose one that is also PHP The framework of language implementation .Laravel It's more popular PHP frame , But its performance is known to be relatively low , The results should be known without comparison . and thinkphp relative Laravel Better performance , So I plan to take it first thinkphp Let's compare .

meanwhile , This test will also use native PHP As a reference , Take a look at the framework implementation , How much performance loss .

The test framework ：

Jsnpp 4.7.0 （ Official website ：http://www.jsnpp.com/）

thinkphp 6.0 （ Official website ：https://www.thinkphp.cn/）

Test environment ：

PHP7.1

The test environment was intended to use PHP8.1, But use composer download thinkphp6 in the future , in PHP8.1 The following cannot be used , So we used PHP7 To test .

Testing software ：ab Software

Test method ： Simulate 10000 visits ,200 Concurrent situation .

Test process record ：

1、 Download the latest version ：

use first composer Download the latest version of Jsnpp and thinkphp, Here I set up folders in the same environment to store the downloaded source code ,Jsnpp Frame on “jsnpp” In the folder ,thinkphp Frame on “tp6” In the folder .

2、 Prepare the test code ：

Because both frameworks use MVC Architecture design , Generally, the template engine will be used to display the page , So both test frameworks use templates . The content displayed on the page is “Hello World!”.

Jsnpp After frame Download , The default is to use the template engine , So there's no need to modify Jsnpp Controller file of the framework , Only need to “app/view/index/index.html” The content of this file is changed to “Hello World!” That's all right. .

thinkphp6 The framework is troublesome , You need to modify the controller file to the following code first ：

namespace app\controller;

use app\BaseController;
use think\facade\View;

class Index extends BaseController
{
    public function index()
    {
        return View::fetch('index');
    }

    public function hello($name = 'ThinkPHP6')
    {
        return 'hello,' . $name;
    }
}

Then add “app/view/index/index.html” This file , And the content is changed to “Hello World!”.

3、 To test ：

Here are Jsnpp Test results of the framework ：

Jsnpp The number of requests per second of the framework is basically maintained at 3600 about .

Here are thinkphp6 Test results ：

 

  Above is thinkphp6 The best test result , Others are basically in 3000 following , On average 2800 about . Back analysis , I use the best results of this time to illustrate , namely 3000 This data .

That's right thinkphp6 Another test of

4、 Native PHP Code testing ：

I created a new one under the same level directory “php” Folder , Then put... In it “index.php” file , The contents of the document are as follows ：

 

<?php
echo 'Hello World!';

This is the use of PHP Simple output of native code “Hello World!”.

Here are the test results ：

Multiple tests ,PHP The native code is basically 4000 about .

5、 Conclusion ：

Pass the above test , We can see ,Jsnpp The performance of the frame is obviously better than thinkphp6,Jsnpp The framework can achieve 3600, and thinkphp6 It can only reach 3000, It is equivalent to that under concurrent conditions ,Jsnpp The framework can be in thinkphp6 Improve the performance base 20% Performance of ,(3600 - 3000) / 3000 = 0.2.

Relative to native PHP Code for ,Jsnpp Less loss of frame performance , Only 10% Performance loss of ,(4000 - 3600) / 4000 = 0.1.

and thinkphp6 The performance loss reaches 25%,(4000 - 3000) / 4000 = 0.25.

Obviously , In terms of performance loss , What is the gap between the two 2.5 times , obvious Jsnpp The framework is much better than thinkphp6.

The above tests are based on desktop computers , Not a test on the server , The data in different hardware environments will change , But the relative data will not change much , The difference of test performance mainly depends on the relative data , The relative data under the same conditions can reflect the performance differences of different frameworks , This test is conducted in strict accordance with the same conditions , Only in this way can we get the correct conclusion .