23 Feb 2014
HHVM, facebook hiphop virtual machine gets a lot of popularity recently due to good performance results. There are some benchmarks made by people showing how good hhvm is as a result of non-blocking I/O and better caching. It is also result of hiphop bytecode, the caching mahanism that is implement that translates eventually everything to c++ but in this case hhvm goes one step further instead of just keeping the php scripts in memory as apc does.
Continue reading...
27 Nov 2013
Part I: Installation of VMWare Workstation
Continue reading...
27 Nov 2013
Tests are performed on CentOS VM with 4GB RAM. I am running 10 queries from application w/o binding parameters in the sql statements in order to verify if i would get less time for bind rather than without. It turns out i am not getting much performance gain using the sql example and data set below.
Continue reading...
27 Nov 2013
Graphics show the results I got from measuring response time using mysql and mongo with php yii framework. First I took Yii so I could get an environment up and running quickly and concentrate on the results. I inserted 110000 records in both databases and did simple selects to print data out. I print out 6 sqls with measuring time for each of them with incrementing the total.
Continue reading...
21 Aug 2013
So, I started using NodeJS with express framework, so I could first figure out how it does things and then make some performance tests to see if it is worth the hustle. I’ve combined it with mongodb and my plan was first to insert data in database and see the behaving as I increase the volume of it. I wanted to call multiple queries with one page call and the way nodejs handles them is like fire them sequentially, continue executing the script and after the db finishes with the queries it will print out the results. It means this:
Continue reading...