Aura Framework AB results.

[janek kos]

I have noticed that Aura Framework Apache Benchmark tests are ... well, I expected a higher results. With -c10 and -n10 it showed me 28 requests per second. 

This is hello world page, no database connection or other common features (Framework_Demo). When I add them I assume that AB tests will go down. 

Is there any guide how to make it faster? I know it's a hello world benchmark, but I expected Aura to have noticable more req/s than "big buddy" Symfony2.

Maybe there is something with my enviroment? It's Windows 8 with XAMPP.

I have to check at micro framework version of Aura, wonder if it will be noticable faster.

[Brandon Savage]

Janek,

Performance benchmarks are only one piece of a small puzzle. There are a number of ways your marks can be off: lack of an opcode cache, fact that you're running on local hardware, etc. For example, I just executed AB against my own install and I got 309 requests per second.

Instead of focusing on RPS you can focus on bottlenecks in your code, and identify what specific components of Aura take the longest. I recommend XDebug and cachegrind. Once you know this information you can determine how best to optimize (if optimization is even necessary).

Brandon

--
You received this message because you are subscribed to the Google Groups "The Aura Project for PHP" group.
To unsubscribe from this group and stop receiving emails from it, send an email to auraphp+u...@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

[janek kos]

I understand that AB is just a piece - however I expected more req/s from vanilla install of such modular framework. Maybe it's really in my local configuration, I'll dig into this later. Your 309 req/s looks amazing, if I had 80 or 100 I would call it quite fast. But 28 is a bit disappointing for me.

[Paul M. Jones]

On Mar 3, 2014, at 10:34 AM, janek kos <the...@gmail.com> wrote:

> I understand that AB is just a piece - however I expected more req/s from vanilla install of such modular framework. Maybe it's really in my local configuration, I'll dig into this later. Your 309 req/s looks amazing, if I had 80 or 100 I would call it quite fast. But 28 is a bit disappointing for me.

It's pretty easy to set up benchmarking incorrectly. Indeed, it may be that AB is being weird; you might want to try Siege or http_perf as an alternative.

Just for reference: if I recall correctly, my own limited benching showed it operating at about 8% or so of a PHP `echo "hello world"` (whereas Symfony, for example, runs somewhat lower than that).


--
Paul M. Jones
pmjo...@gmail.com
http://paul-m-jones.com

Modernizing Legacy Applications in PHP
http://mlaphp.com

[Hari K T]

Hey,

So I was looking at siege and the post http://blog.remarkablelabs.com/2012/11/benchmarking-and-load-testing-with-siege .

I am not an expert in the benchmarking, so I leave things to the masters.

Thanks

Hari K T

You can ring me : +91 9388 75 8821

http://harikt.com , https://github.com/auraphp , http://www.linkedin.com/in/harikt , http://www.xing.com/profile/Hari_KT

Skype  : kthari85

Twitter : harikt

[janek kos]

I have downloaded siege for windows and I'll try to make some additional tests in this week.