performance problem in python 2.2

[Mark Day]

In article <ahsh4t$i13$1 at peabody.colorado.edu>, Fernando Perez
<fperez528 at yahoo.com> wrote:

> As you see, perl/python are in the same league, even though perl is a factor 
> of 2 faster.

Maybe you missed Tim's comment about putting the loop in a function to
make it faster.  Using that trick made the Python version as fast as
Perl on my machine.

-Mark

% time ./p1.py 1000000
2.71050724088e-08
./p1.py 1000000  2.76s user 0.06s system 83% cpu 3.375 total

% time ./p2.py 1000000
2.71050724088e-08
./p2.py 1000000  1.88s user 0.07s system 82% cpu 2.369 total

% time ./p3.pl 1000000
2.71050724087729e-08
./p3.pl 1000000  1.90s user 0.01s system 85% cpu 2.246 total

% cat p1.py
#!/usr/bin/env python

import sys

n = 1.0
p = 2.0**64 # making p a float for fariness of comparison, not
correctness
c = int(sys.argv[1])

for i in xrange(1,c+1):
    n = (n * (p-i)) / p
print 1-n


% cat p2.py 
#!/usr/bin/env python

import sys

def foo(arg):
        n = 1.0
        p = 2.0**64 # making p a float for fariness of comparison, not
correctness
        c = int(arg)

        for i in xrange(1,c+1):
                n = (n * (p-i)) / p
        print 1-n

foo(sys.argv[1])

% cat p3.pl 
#!/usr/bin/perl

$p = 2**64;
$c = $ARGV[0];


$n = 1;
foreach $i (1..$c) {
        $n = ($n * ($p-$i)) / $p
}
print 1-$n, "\n";