I was wondering does anyone know a trick to reduce memory requirement during a bucket sort without significantly reducing the speed

How? (which module or post code); and why? -- reasoning -- are you using a "bucket sort" from Perl?

Given that:

• Your data is so unsuited to a bucket sort:

• Perl's in-built (merge) sort (since 5.8x???) will sort numerical data quite quickly for ranges that most hardware will handle;

  #! perl -slw
  use strict;
  use Time::HiRes qw[ time ];
  
  my @data = qw[
      2 53 23456 75434532 233 45654 665446 33 234 5 6665 3 2 23 43
      62553 27263 2 33 75434532 987762 34455 22 3345 24 567 54 43 12 23
  ];
  
  @data = ( (@data) x 10 );
  
  my $start = time;
  @data = sort{ $a <=> $b } @data;
  printf "sorting %10u integers took: %.6f\n", scalar( @data ), time() -
  + $start;
  
  @data = ( @data ) x 10;
  $start = time;
  @data = sort{ $a <=> $b } @data;
  printf "sorting %10u integers took: %.6f\n", scalar( @data ), time() -
  + $start;
  
  @data = ( @data ) x 10;
  $start = time;
  @data = sort{ $a <=> $b } @data;
  printf "sorting %10u integers took: %.6f\n", scalar( @data ), time() -
  + $start;
  
  @data = ( @data ) x 10;
  $start = time;
  @data = sort{ $a <=> $b } @data;
  printf "sorting %10u integers took: %.6f\n", scalar( @data ), time() -
  + $start;
  
  @data = ( @data ) x 10;
  $start = time;
  @data = sort{ $a <=> $b } @data;
  printf "sorting %10u integers took: %.6f\n", scalar( @data ), time() -
  + $start;
  
  __END__
  C:\test>1113070
  sorting        300 integers took: 0.000152
  sorting       3000 integers took: 0.000316
  sorting      30000 integers took: 0.001214
  sorting     300000 integers took: 0.018363
  sorting    3000000 integers took: 0.192435
  [download]