* Thus wrote Justin Palmer:
> Why is substr all over the board in how fast it processes the same
> string? Is it the server?
Benchmarking can be a funny thing. Usually benchmarks take an
average of serveral thousand iterations in order to find the
average time it take something to do.
Benchmarking with the output of a browser can lead to different
results on each request for a varity of reasons like:
1) processor load (and capacity) at the time
2) memory allocation for the paticular request
(which will vary uppon each request)
3) something I overlooked.
Another thing to note is that there generally is more overhead on
the first few routines of a benchmark, try swapping your test so
that the 'String Array' is done first.
Another thing to note is your test script has a lot of things that
can factor speed issues. In order to test the difference between to
different methods you want to set something up like:
<?php
// initialize variable before time testing
// so memory allocation issues are avoided.
$i=0;
$var = ' ';
$str = 'a b c d';
//start timer
for($i=0; $i <= 1000; $i++ ) {
$var = $str{0};
}
//end timer
//start timer
for($i=0; $i <= 1000; $i++ ) {
$var = substr($str, 0, 1);
}
//end timer
Then check the results, record them and swap the two benchmarks.
check the results and record.
All in all, you probably will still see that the $str{0} is faster,
and as mentioned, its because its substr() is a function and
$str{0} is handled by the Zend Engine.
A time of substr result = 0.0000370000, is rather quick and is can
be assumed that there is a large amount of error (based on the
level of accuracy we are dealing with).
On my 450hz machine I wont accept a page that takes longer than
0.100 to execute. Which is a very long time compared to the
0.000034 for substr().
HTH,
Curt.
Curt
--
Quoth the Raven, "Nevermore."
--
PHP General Mailing List (http://www.php.net/)
To unsubscribe, visit: http://www.php.net/unsub.php
