[Benchmark php frameworks]. Part 4. Laravel

Table of content.

Well, I got to benchmark the second framework according to my plan. ( More info about plan in the first part )

Some information about framework.

Site: https://laravel.com/

It's a very big community. The Community is growing as well as the framework is developing fast. It has a lot of different services. And There are many packages for different cases.

There is a good review here by @willian95.

Let's see benchmark.

Just to remind about the structure ( Full description here )

• /laravel-authors - lists of authors with posts counts.
• /laravel-categories - list of categories with counts of posts in every one.
• /laravel-posts - all posts with information about types, authors and categories.
• /laravel-index - simple "hello world" for comparison

For my benchmark I took this code and made some modifications. All code committed to Github .

framework	requests per second	relative	peak memory	relative
laravel-index	9.56	inf	6.17	1.0
laravel-authors	7.23	inf	7.38	1.2
laravel-categories	6.44	inf	7.56	1.2
laravel-posts	0.00	nan	17.73	2.9

We see that the generation time of the initial almost empty page (laravel-index) is not very different from the time of generation of other pages. For a second, ~ 6-10 pages were generated. But the generation of the largest and most complex page - laravel-posts - all 1000 posts with information about types, authors and categories - did not meet the second.

Some Graphics

Here it is visible that generation of page laravel-posts was carried out more than 6 seconds.

Later I will connect the service https://blackfire.io/ to profile the requests inside and find out where the delays occur.

Let's compare #phalcon and #laravel

framework	requests per second	relative	peak memory	relative
phalcon-index	309.99	inf	0.44	1.0
phalcon-authors	31.18	inf	0.56	1.3
phalcon-categories	16.80	inf	0.57	1.3
laravel-index	9.91	inf	6.17	14.1
laravel-categories	8.32	inf	7.56	17.3
laravel-authors	8.26	inf	7.38	16.9
phalcon-posts	0.07	inf	3.00	6.9
laravel-posts	0.00	nan	17.65	40.4

Phalcon works faster, although the page generation time has also increased, but as I suggested in a previous article, in the case of phalcon, this is similar to the complexity with mysql. Later we will find out.

Most of all I'm interested in the time of generation of a page with posts.

Phalcon-posts ~ 2.3s
Laravel-posts ~5.8s

On the one hand, the difference is only 2 times. Is this a lot or a little? It depends on the project. For a highly loaded project with millions of visitors a day, that's a lot. But in such projects, data caching is always used. In the future, I will also run tests with connected caching. And also describe the process of its inclusion.

In addition, for big queries, you need to optimize the database settings. And its updating to a newer version take effect.
But this is more difficult for me. The server part is not my strong point yet. But I'm working on it.

If there is any inaccuracy in my research - welcome to the comments.

It's also very interesting to look at the results of two more equally well-known frameworks: Yii and Symfony.

In the foreseeable future:

• Implementing a test application and running tests on Symfony
• Implementation of the test application and testing on Yii
• Try to use Cache for all frameworks
• Connection with https://blackfire.io/
• Deal with the Docker and wrap it all up in it. To be able to quickly deploy to a more powerful server for comparison.

All code is committed to github: https://github.com/semasping/php_frameworks_for_tests

Thank you for your attention and your support.