Problems? Is your data what you think it is? PerlMonks

### Re^4: sort an array with +ve & -ve numbers in it

by grinder (Bishop)
 on Apr 03, 2009 at 21:52 UTC ( #755350=note: print w/ replies, xml ) Need Help??

Easier and simpler, but less efficient

Woah! I would benchmark that before making such a statement. I expect you'll need a seriously huge list before the O(n log n) starts to lose to your straight O(n).

There's no debate that it will, but the problem is the 0(n) algorithm is using a comparatively large number of slow ops, where as the sort compiles down into a single op, and there you're running at C speed (simple {\$a <=> \$b} blocks are recognised and special-cased during the parse). This will drown out the extra cost for a long, long time (that is: for a long list of values). On my machine, the cross-over occurs between 100 000 and 1 000 000 elements (and I had to run the million element benchmark for 15 seconds in order to give it enough time to settle down)

```               Rate with_scan_1 with_sort_1
with_scan_1 35662/s          --        -53%
with_sort_1 76332/s        114%          --
Rate with_scan_2 with_sort_2
with_scan_2  6838/s          --        -40%
with_sort_2 11437/s         67%          --
Rate with_scan_3 with_sort_3
with_scan_3 759/s          --        -11%
with_sort_3 853/s         12%          --
Rate with_sort_4 with_scan_4
with_sort_4 62.6/s          --        -18%
with_scan_4 76.2/s         22%          --
Rate with_sort_5 with_scan_5
with_sort_5 3.38/s          --        -47%
with_scan_5 6.38/s         89%          --
s/iter with_sort_6 with_scan_6
with_sort_6   4.88          --        -66%
with_scan_6   1.65        196%          --

And since either choice is crazy fast enough for me, I'd throw my lot in with the more succinct version -- less chance of introducing semantic mistakes and off-by-one errors). For instance, I had to think for a little while about how you initialised \$min and \$max...

```#! /usr/local/bin/perl

use strict;
use warnings;

use Benchmark 'cmpthese';

my @s1 = (999, -999, map {rand(900)-900} 0..10);
my @s2 = (999, -999, map {rand(900)-900} 0..100);
my @s3 = (999, -999, map {rand(900)-900} 0..1000);
my @s4 = (999, -999, map {rand(900)-900} 0..10000);
my @s5 = (999, -999, map {rand(900)-900} 0..100000);
my @s6 = (999, -999, map {rand(900)-900} 0..1000000);

sub with_scan {
my ( \$min, \$max ) = @_;
for my \$element ( @_ ) {
\$min = \$element if \$min > \$element;
\$max = \$element if \$max < \$element;
}
return (\$min, \$max);
}

sub with_sort {
return (sort {\$a <=> \$b} @_)[0,-1];
}

print join(',', with_scan(@s1)), \$/;
print join(',', with_sort(@s1)), \$/;

cmpthese(
-3, {
'with_sort_1' => sub {with_sort(@s1)},
'with_scan_1' => sub {with_scan(@s1)},
}
);

cmpthese(
-3, {
'with_sort_2' => sub {with_sort(@s2)},
'with_scan_2' => sub {with_scan(@s2)},
}
);

cmpthese(
-3, {
'with_sort_3' => sub {with_sort(@s3)},
'with_scan_3' => sub {with_scan(@s3)},
}
);

cmpthese(
-3, {
'with_sort_4' => sub {with_sort(@s4)},
'with_scan_4' => sub {with_scan(@s4)},
}
);

cmpthese(
-3, {
'with_sort_5' => sub {with_sort(@s5)},
'with_scan_5' => sub {with_scan(@s5)},
}
);

cmpthese(
-15, {
'with_sort_6' => sub {with_sort(@s6)},
'with_scan_6' => sub {with_scan(@s6)},
}
);

• another intruder with the mooring in the heart of the Perl

Comment on Re^4: sort an array with +ve & -ve numbers in it

Create A New User
Node Status?
node history
Node Type: note [id://755350]
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?