According to this, Laravel doesn't look too good, but I fail to see the relevance of a benchmark that does nothing other than show the "Hello World" page.
Speed is great... if you need a very speedy app. However for most apps, you just want a response in a reasonable time. e.g. under 100ms. If my app is averaging 150ms, then it's simple - I spend a few more £/$ a month on a faster box...
...this is a far cheaper solution than spending hours learning different frameworks or spending lots of time on optimisation. Cheaper/faster hardware is only a few month's wait.
Also factor in that PHP7 when it arrives should easily half the request time for most PHP scripts/frameworks if you check out their recent benchmarks (and they haven't finished yet!) https://wiki.php.net/phpng#performance_evaluation
I've seen lots of those benchmarks recently. I'm keen to see how the rest of those frameworks would do with a fairly decent app the size of say, Laracasts. A nigh impossible benchmark of course as that would require building 10 different sites to service each framework. My hunch is that L5 would be impressively scalable in comparison to the others.
Worth mentioning that Ruby on Rails is pretty darn unimpressive in those kind of benchmarks too yet the number of large sites that use that as a framework is vast!
@JeffreyWay would love to know how your response times changed as Laracasts grew bigger and bigger.
I guess most of the things stripped out in the neomerx tweaking are things you want to keep for a real world application. Very interesting read though, and it shows how useful a profiling tool like Blackfire can be.
My personal experience with Laravel is that it is destined to be slower than most other frameworks due to design decisions. Framework tries to do a lot of things behind the scenes; a lot of which can be taken care of by a bit of an advanced developer easily and efficiently. I've implemented a small CRM project with Laravel as part of my self-learning and every time an exception was thrown, looking at the stack trace, I knew it is not the right framework for me; since most of the projects I've developed eventually comes to a point where I have to do anything in the book to optimize the full stack.
https://github.com/neomerx/rhw-l5/blob/master/readme.md is a good resource if you want to stick with Laravel, however, any framework can benefit from those optimizations done at server level, hence when it comes to comparison with other frameworks, Laravel might look even slower.
I think it would be great if @JeffreyWay could do a video on optimizations that can be done for laravel. I'm building an API right now that will be used by many of our clients, and I'm a bit worried about performance during peak hours. Lumen looks really promising but came out to late for us to use for this project.