Re: [Sheflug] Perl tips

[Dave Mitchell <davem [at] xxxxxxxxxxx>]

On Mon, Nov 25, 2002 at 04:19:34PM +0000, Alex Hudson wrote:
> On Mon, Nov 25, 2002 at 04:05:40PM +0000, Dave Mitchell wrote:
> > Except of course that the function will be called NlogN times, which
> > may then make it rather slow.
> 
> O(NlogN) function versus a similar order data structure - I think I 
> did point out that it would be exchanging speed for space :)

Pointless benchmark follows :-)

on my work desktop system (sparc/solaris),
    using a hash took 6 secs and a 7.2M process size;
    using a function took 34 secs and a 5M process.

But since my desktop system isn't even Linux, perhaps I better shut up for
the time being on this list :-)

Dave.

-- 
That he said that that that that is is is debatable, is debatable.


#!/usr/bin/perl -w

local $/;
@ARGV = ('/usr/dict/words');
@words = split /\n/, <>;


sub munge { $_[0] =~ /^(...)(.+)$/ ? "$2$1" : $_[0] }

sub s1 {
    @words2 = sort { munge($a) cmp munge($b) }  [at] words;
}

sub s2 {
    $words{$_} = munge($_) for  [at] words;
    @words2 = sort { $words{$a} cmp $words{$b} }  [at] words;
}

$t=time;
s1; # or s2
$t=time-$t;
print "took $t secs\n";
system("ps -lfyp $$");
___________________________________________________________________

Sheffield Linux User's Group -
http://www.sheflug.co.uk/mailfaq.html

  GNU the choice of a complete generation.