Open tom-- opened 8 years ago
Here are preliminary results comparing php56 vs php70: https://github.com/motin/php-framework-benchmark/blob/multiple-stacks/OTHER_RESULTS.md#motin
Interesting!
Do the zeros represent failed tests?
The No-framework number changed little between 5.6 and 7.0. Is that because that case is bound by something other than PHP script execution, e.g. fpm?
@motin How many times do you run the benchmarks? Did you confirm the results are roughly the same all the time?
I don't understand why tipsy is faster than phalcon, codeigniter is so slow, on PHP-FPM 5.6.4 with opcode cache.
Do the zeros represent failed tests?
Yes indeed
The No-framework number changed little between 5.6 and 7.0. Is that because that case is bound by something other than PHP script execution, e.g. fpm?
The environments are identical in set-up, hardware, OS and configuration except for the fact that php7.0-fpm is used instead of php5-fpm. I'd say it is pretty fair to assume that PHP 7 is simply a bit faster even in the hello world case.
How many times do you run the benchmarks? Did you confirm the results are roughly the same all the time?
I ran them a couple more times, and the results vary quite a bit. Sometimes tipsy is twice as fast as phalcon. Will run this in production environment and report back the results.
I don't understand why tipsy is faster than phalcon, codeigniter is so slow, on PHP-FPM 5.6.4 with opcode cache.
Haven't checked, but maybe because it is extremely light-weight and more similar to a the no-framework set-up than a full-fledged framework.
It's good idea to add column for php5.6 and another column for php7 in final results picture.
Also please add ZF3
Enhancement request: It would be very interesting to be able to compare PHP 5.6 against 7.0 using your benchmarks. I mean: update the readme to show both.