Comparison of Speeds in PHP

[DtY]

I was wondering today just how much less efficient is double quotes than single quotes, and other such stuff. Here are the results:
$ php -v
PHP 5.2.9 (cli) (built: Apr 17 2009 03:29:12)
Copyright (c) 1997-2009 The PHP Group
Zend Engine v2.2.0, Copyright (c) 1998-2009 Zend Technologies

Single Vs. Double Quotes
Single Quotes

php:

<?php     for ($i=0; $i<10000; $i++) {         echo 'Hello, World! ';     } ?>

Double Quotes

php:

<?php     for ($i=0; $i<10000; $i++) {         echo "Hello, World! ";     } ?>

Results

code:

Single      Double user    0m0.033s    0m0.022s sys     0m0.022s    0m0.027s total   0m0.055s    0m0.049s

So, it seems that double quoted strings actually run faster (unexpected since double quoted strings support interpolation where single quoted strings don't). I am guessing this is a design flaw in PHP, and might be fixed in the future? Single quoted strings should be able to run faster than double quoted strings.

Interpolation Vs. Concatenation
Interpolation

php:

<?php     for ($i=0; $i<10000; $i++) {         echo "Hello, $i ";     } ?>

Concatenation

php:

<?php     for ($i=0; $i<10000; $i++) {         echo "Hello, " . $i;     } ?>

Results

code:

Interpolation   Concatenation user    0m0.065s        0m0.049s sys     0m0.037s        0m0.045s total   0m0.102s        0m0.094s

No surprises here. Concatenation runs faster than interpolation, but not significantly

++ Vs. +=1
++

php:

<?php     for ($i=0; $i<10000; $i++); ?>

+=1

php:

<?php     for ($i=0; $i<10000; $i+=1); ?>

total

code:

++          +=1 user    0m0.015s    0m0.024s sys     0m0.014s    0m0.010s total   0m0.029s    0m0.034s

This is good, the nicer one runs faster.

(These are just on my computer, but the trends should be the same. As far as I know, they're accurate, unless PHP is doing optimizations on some of them, that it might not do in a place that's so obvious)

[rdrake]

Try comparing i++ to ++i. PHP is full of all sorts of things like this.

[DtY]

code:

$i++        ++$i real    0m0.194s    0m0.015s user    0m0.024s    0m0.014s total   0m0.218s    0m0.029s

Oh wow
I wonder how often anyone uses the return values from pre-/post-increments. Is there any use to having both of them? (Other than c does)

[DemonWasp]

The ++i versus i++ thing is fairly frequent - even C++ has had issues with it. The biggest issue is when the increment is being done on an object with an overloaded ++ operator. Then, the ++i just does the increment. The i++ version makes a copy of the object, runs the increment on the copy it's going to keep, then uses the "backup" copy to complete the rest of the statement, even if there's nothing else to the statement.

For what it's worth, I've almost never used the return value from a ++i or i++ . You can get into trouble with that in some languages (cough C / C++).