Can anyone see how to improve the performance--I'd like an order of magnitude improvement if possible--of the following code?

Essentially, the state of the booleans in @bools will switch from false to true in some undefinable order. Each time one becomes true, it will affect the composite state of one or more sets in @sets. And if all the booleans referenced by any particular set are true, then I need to remove that set from @sets.

This code does a simple brute force search of all the sets and checks all the booleans referenced. I keep looking at this thinking that there should be some combination of hashes or heaps and/or references that would allow me to remove completed sets more efficiently, but I'm not seeing it.

Update:Corrected typos.

#! perl -slw
use strict;
use Data::Dump qw[ pp ];
use List::Util qw[ shuffle ];
use Time::HiRes qw[ time ];

our $N //= 1e3;
our $S //= 1e4;
our $Z //= 5;

my @bools = (0) x $N;

my @sets = map [
    map int( rand $N ), 1 .. 1 + int( rand $Z )
],1 .. $S;

my @someOrder = shuffle 0 .. $N-1;

my $start = time;
for my $next ( @someOrder ) {
    $bools[ $next ] = 1;
    for my $set ( reverse 0 .. $#sets ) {
        if( grep( $bools[ $_ ], @{ $sets[ $set ] } ) == @{ $sets[ $set
+ ] } ) {
            splice @sets, $set, 1;
        }
    }
}
printf "Took %.6f\n", time() - $start;

__END__
C:\test>junk27
Took 11.623000
[download]