Re^4: quick sort. How do it faster

I'm surprised that introducing that temporary variable for swaps makes such a difference, but it does appear to. Do you have any explanation?

Not really. Just that it seems to be some inefficiency inherent in the way list ops are implemented in perl.

Perhaps even more surprising is that -- for numeric data -- the old saw, triple xor-assignment is more efficient than both the list assignment and the introduced temporary:

cmpthese -1,{
    a=>q[ my( $a, $b    ) = ( 1, 2 ); ( $a, $b ) = ( $b, $a )   for 1.
+.1000; ],
    b=>q[ my( $a, $b,$t ) = ( 1, 2 ); $t = $a, $a = $b, $b = $t for 1.
+.1000; ],
    c=>q[ my( $a, $b    ) = ( 1, 2 ); $a ^= $b ^= $a ^= $b      for 1.
+.1000; ]
};;
    Rate    a    b    c
a 3253/s   -- -34% -42%
b 4913/s  51%   -- -13%
c 5632/s  73%  15%   --
[download]

That really ought not be the case.

With the rise and rise of 'Social' network sites: 'Computers are making people easier to use everyday'

Examine what is said, not who speaks -- Silence betokens consent -- Love the truth but pardon error.

"Science is about questioning the status quo. Questioning authority".

In the absence of evidence, opinion is indistinguishable from prejudice.

Comment on Re^4: quick sort. How do it faster Download Code