Since php scripts (like CMSs) are written for a general audience, they have lots of comments between */ */ tags to explain what the php codes do this. Obviously this increases the sizes of the php files.

I'm just curious whether the presence of such comments affect the load of files? I mean removing the unnecessary comments will make the php files lighters (lesser size); does this have a significant influence on the script performance when reading the php files?

No.

lines inside comment blocks /* ... */ are simply skipped by the interpreter as it goes into "read ahead" on the open delimiter until it encounter the end delimiter.

lines with "//" in are processed to the comment marker and anything after the marker is discarded.

Originally Posted by chrishirst

No.

lines inside comment blocks /* ... */ are simply skipped by the interpreter as it goes into "read ahead" on the open delimiter until it encounter the end delimiter.

lines with "//" in are processed to the comment marker and anything after the marker is discarded.

But it should have an influence on the CPU usage, as the system should open the whole file; like opening a large file in windows.

i didnt think it opened the whole file?
I thought it streamed it..

But it should have an influence on the CPU usage, as the system should open the whole file; like opening a large file in windows.

And what do you think the PHP engine does on each access to a PHP file ?
Php is a dynamic, scripting language. As it is not compiled in a binary form, each access to it cause the parsing of the file.
Of course it have an incidence, but the comment are so light, that they are negligible.

Last time I did a trace into PHP files, 98% of the execution time where sending SQL query to the db and waiting for the result.
Optimize your code logic first, then your sql queries, and don't bother going further.

Originally Posted by tripy

Last time I did a trace into PHP files, 98% of the execution time where sending SQL query to the db and waiting for the result.

Is that typical, or is it just the last time you did the test?

Well, the db operations where heavy in that specific page.
The typical time I saw on more conventional pages is around 60~70%.

Significantly less, but still a lot.
At the time, I used Xdebug and KCacheGrind on linux to visualize the trace.
You can use wincachegrind on windows for that too: http://sourceforge.net/projects/wincachegrind/

http://xdebug.org/docs/profiler