[SOLVED] PHP vs Python

[The Little Guy]

Is there any speed difference between python and php?

[The Little Guy]

OK, I ran 2 Scripts.

 

PHP:

<?php
$p = 10000000;
$t = 0;

$s = microtime(true);
for($n=0;$n < $p;$n++){
$t = md5($n + $n);
$t = 0;
}
$e = microtime(true);
echo $e-$s;
?>

 

Python:

#!/usr/bin/python 

import md5
from time import clock

p = 10000000
n = 0
t = 0
start = clock()
while n < p:
t = n + n
md5.new(str(n)).digest()
n = n + 1
t = ''
stop = clock()
print "Content-type: text/html\n\n"
print round(stop-start,4)[/coe]

 

The final outputs (rounded to the nearest 100^th):

PHP:

- 30.84

- 25.66

- 21.90

- 24.75

- 19.25

 

Python:

- 23.89

- 23.25

- 23.46

- 23.44

- 23.45

 

Averages:

PHP: 24.48

Python: 23.49

 

WOW, I really thought Python was going to win by a lot, since PHP's First round was almost 31. Guess I was wrong.

 

Python does look like a more constant language though seeing that all the results stayed at 23, otherwise PHP just caches the values for faster speeds when ran multiple times.

 

Any thoughts on this test?

[Mchl]

My results:

 

#PHPPython

143.6428.60

244.0827.66

343.7228.34

444.4527.81

544.0728.49

643.5127.97

743.5628.00

844.6028.52

944.6627.86

1044.2827.80

Avg:44.0628.11

 

Note no reduction in time for consecutive PHP executions.

 

It's run on Vista Business, Intel CoreDuo T9500

 

[The Little Guy]

I assume you copied and pasted my scripts correct?

 

If so, did you have a procedure of how you ran it?

 

Such as...

 

run PHP 10 times, then python 10 times

or

run php 1 time then python 1 time, 10 times

 

 

I did the second way.

[Mchl]

It was 10 times PHP then 10 times Python

 

[Daniel0]

FYI, the second time you run the python script it should be faster because it caches the byte code so it doesn't have to be interpreted again unless the source file has been changed. Those are the .pyc files you'll see. You can achieve the same effect with an opcode cache for PHP such as APC as well of course.

[Mchl]

Not much source here to interpret though.

[Daniel0]

No, but on a large scale project with many files, having an opcode cache can improve the performance considerably. Besides, if we say that we run a small script like the above ones a thousand times then you should see a noticeable difference by saving 999 interpretations. A site like The Pirate Bay gets many thousands of hits each minute so caching is crucial for system performance regardless of the script's complexity.

[Mchl]

In those cases yes.

 

For the purpose of this benchmark, bytecode caching has probably a negligible influence (unless we indeed try to run it several hundreds of times).

[Daniel0]

The issue in question was whether Python was faster than PHP. I suppose this is for real-world usage instead of tiny benchmarks only covering a confined amount of structures and features of each language in a sample script that wouldn't be used otherwise. The only thing tested in the above benchmark was loops, integer addition and the MD5 algorithm.

[Mchl]

Only thing it proves is that Python loops faster than PHP on my PC 

[corbin]

Python is almost 3 times slower on my box...  Like already mentioned, this is a horrible comparison.  For example, what if the md5 function is just super slow on my computer under python?

[The Little Guy]

what would be a good comparison? I was running my scripts on a shared web server.

[corbin]

I don't think there is a fair comparison that isn't long and drawn out.  Basically you would have to create the same application in both languages, making sure everything is handled the same.  Even then though, it could be inaccurate, although it would give a real world reading. 

[448191]

Those benchmarks are useless. All you are testing is the speed of creating MD5 hashes.

[corbin]

Python is almost 3 times slower on my box...  Like already mentioned, this is a horrible comparison.  For example, what if the md5 function is just super slow on my computer under python?

 

 

I just realized what I typed earlier, and I want to reword that to say:

 

Python is almost 3 times slower on my box with this script...  Like already mentioned, this is a horrible comparison.  For example, what if the md5 function is just super slow on my computer under python?

[The Little Guy]

I just wrote a mysql query, I did the exact same thing in Python as PHP, and the query took this long:

 

Python - 0.01

PHP - 0.075

 

Using SQL_CALC_FOUND_ROWS, it would have returned: 3,649 rows, otherwise with the limit it returned 5 rows.

[corbin]

Once again, all that proves is that feature a is faster in language b than in language c.  That doesn't make language b better than c or the other way around x.x.

 

 

This thread is seeming pointless to me ;p.

[The Little Guy]

I never said PHP was faster than Python.

 

First I was asking, then testing. I also got some good replies, such as yours:

 

all that proves is that feature a is faster in language b than in language c

 

Ill mark this Solved. If anyone wants to continue, be my guest.

[Mchl]

If you like doing such things, then I'd love to see some comparison between libmysql (default library that connects PHP to MySQL) and mysqlnd (new native driver)

[corbin]

http://blog.felho.hu/what-is-new-in-php-53-part-3-mysqlnd.html

 

No idea if that's accurate/true or not, but interesting read (found through Google).

 

 

It links to this as a benchmark:

 

http://blog.ulf-wendel.de/?p=138

 

 

Once again, I can't/won't vouch for it being accurate/true.

[Mchl]

Once again, I can't/won't vouch for it being accurate/true.

 

Since mysqlnd is not a finished product yet (or so I understand) then of course all benchmarks have to be taken with a grain of salt.

Nevertheless it's nice to see some. Thanks for the links.