Is it worth it to optimize?

I

Ilya Beloborodov2015-10-30 00:41:51

Yii

Ilya Beloborodov, 2015-10-30 00:41:51

In general, I've been working on a yii project for a year.
browser game.
And then I thought: why do I need this yii? I don't use even 10% of the possibilities from this framework in this project.
Yii out of the box consumes approximately 5MB of RAM.
I decided to plan a quick framework. made, tested, consumes 28kb of RAM.
Is it worth finishing your framework and transferring all the achievements for the year there?
How much will I win?

Reply

Answer the question

In order to leave comments, you need to log in

3 answer(s)

I

Igor Kalashnikov, 2015-10-30
@kowap

Rewriting for your bike is a bad decision.
A good solution might be to tweak yii, or profile your code to find bottlenecks.
The main question: is the project somehow in production? you have at least 100 users online, what are you worried about 5mb? 1GB of RAM costs 10 bucks a month on DigitalOcean, if anything.
Your bike will almost 100% come out sideways. If you really want to optimize, look for some more productive framework than Yii and crawl to it. I used to like Phalcon.

A

Andrew, 2015-10-30
@Ashlst

I decided to plan a quick framework

Can look towards microframeworks????

V

v- death, 2015-10-30
@vGrabko99

Of course it's worth it! You will win big % of productivity.
Simple with yii, pull out those 10% that you use and make your own micro framework from them.
Once I did something like Heroes of Might and Magic on phoneGap + api
api wrote initially on laravel4 everything was ok. But over time, about 30k people installed my craft. And online was constantly about 3k.
I profiled my code for a long time. Nothing helped (compared to rewriting from scratch to php native + thoughtful architecture). As a result, I received an acceleration of responses almost twice. Having danced with php (brought it all to stable), I began to rewrite bits of code on GO (microservices. For example, all web sockets were transferred to it in the first place) now I have more than 40% of all API methods on GO. There are plans to completely switch to Go. (now in some go services I use c++ libs)