eZ conference & FrOSCon
First published at Friday, 23 June 2006
Warning: This blog post is more then 18 years old – read and use with care.
eZ conference & FrOSCon
After a week at the amazing eZ conference in Skien (Norway), the city where the headquater of eZ systems is located, I will be at the FrOSCon for the next two days. I won't forget about the conference so fast - meeting all the guys, sometimes the first time in "real life", from eZ was amazing. Thanks for the nice week to all of you - the crew day was real fun and the show around the eZ awards 2006 was simply great.
I will give a talk about regular expression on the PHP track organized by the PHP Usergroup Dortmund and a talk about Rendering 3D with PHP at the main track. Together with toby there will also be an two hour talk about "PHP - best practices ".
I hope to see some of you at the FrOSCon tomorrow or on sunday.
Subscribe to updates
There are multiple ways to stay updated with new posts on my blog:
Comments
Luke Welling at Thursday, 3.8. 2006
I was not there, but I was having a skim through your slides and wondered how confident you are of the recomendation on slide 32 to use pre-increment rather than post-increment?
http://talks.php.net/show/php-best-practices/32
My quick and dirty benchmark of running your code but with the for loop incrementing 1 billion times had post-increment fractionally faster.
Are there systems where pre-increment is significantly faster?
kore at Thursday, 3.8. 2006
Yes. I gained a overall performance increase due to replacing most of the post increments with pre increments of about 10% in Image_3D.
Pre-Increment does need less operations, what will be significant in not optimized languages. I don't know if any PHP Opcode-Caches optimize this, so that you perhaps need to disable them for testing.
Just imagine, that the interpreter is able to just add 1 to a register on pre increment, but needs to allocate new memory and add 1 later on post increment. If I remember correctly these are 1 vs. 4 operations for the interpreter.