There is nothing wrong with FuelPHP's performance. You always have to be careful with interpreting results like this. Every framework is designed different, and that design can have an influence on the test results in a specific test case. Same for the way the "Hello World" example is coded.
For example, FuelPHP 1.x loads quite a few core files on startup, which might be noticable in benchmarks on systems with relatively slow I/O. But this number doesn't really increase when you build your app.
Another example: FuelPHP's "Hello World" example contains a fully functional controller, including loading css, images, and twitter bootstrap. It also uses a ViewModel. Compare that to CI's version, which only echo's a few lines of text.
Having said that, while working on the 2.0 kernel and core we run performance tests on every change, both on a Windows and a Linux platform. So far, the 2.0 core is about 7% faster then 1.1.
And finally, performance isn't everything. By definition every layer of abstraction adds complexity and makes things a bit slower. We feel that is acceptable (within limits off course) if there are enough gains elsewhere. FuelPHP is a framework designed for large/corporate applications. In such environments, the speed of development is a big part of the total cost, so that the cost of a slightly faster server isn't really significant.
Donate
