in reply to use constant and Compiler Optimizations

Stuff that gets optimized away does not necessarily take no time at all. They just should take less time than the unoptimized version in the general case (and on my system not optimizing is actually faster than optimizing for this code, which frankly surprised me). See the code below.

Your example uses extremely fast constructs; the difference between doing a single multiplication and one "no-op" vs doing two multiplications (while still doing a relatively expensive function call each iteration for both constructs) is not very informative or useful. 

use strict;
use warnings;
use Benchmark qw( cmpthese );
use constant DEBUG => 0;


my $debug = 0;
my $b = 100;
cmpthese -10, {
    constant   => sub { my $a = 10 * $b; $a = 10 * $b if DEBUG; },
    noconstant => sub { my $a = 10 * $b; },
    var => sub {  my $a = 10 * $b; $a = 10 * $b if $debug; },
};

[download]
Update: as an aside, a really hard-core optimizer would eliminate ALL of the code in these examples.

"What should it profit a man, if he should win a flame war, yet lose his cool?"

Replies are listed 'Best First'.
Re^2: use constant and Compiler Optimizations
by ikegami (Patriarch) on Jul 18, 2008 at 20:16 UTC

    Stuff that gets optimized away does not necessarily take no time at all.

    If you look at the tree the compiler generates, you'll see the call to bar isn't even present in it. The only thing left is the nextstate which is an incredibly fast operation. 

    >perl -MO=Concise,-exec -e"foo(); bar() if 0; baz();"
1  <0> enter
2  <;> nextstate(main 1 -e:1) v
3  <0> pushmark s
4  <#> gv[*foo] s/EARLYCV
5  <1> entersub[t2] vKS/TARG,1
6  <;> nextstate(main 1 -e:1) v     <---
7  <;> nextstate(main 1 -e:1) v
8  <0> pushmark s
9  <#> gv[*baz] s/EARLYCV
a  <1> entersub[t4] vKS/TARG,1
b  <@> leave[1 ref] vKP/REFC
-e syntax OK

    [download]
[reply]
[d/l]
[select]
      Yes, but funnily enough, doing an actual boolean test on an integer appears to sometimes be faster on my machine than doing the optimization. I'm not sure why, and the results vary. The output of my test script above with -10 and -2 as the first argument to timethis() cmpthese() respectively is: 

                      Rate        var   constant noconstant
var        5014619/s         --       -10%       -40%
constant   5551194/s        11%         --       -33%
noconstant 8292829/s        65%        49%         --

      [download]
                      Rate   constant        var noconstant
constant   5505020/s         --       -17%       -23%
var        6604512/s        20%         --        -7%
noconstant 7120433/s        29%         8%         --

      [download]
      "What should it profit a man, if he should win a flame war, yet lose his cool?"
[reply]
[d/l]
[select]

        You called it right earlier. Two layers of subroutine call (your's and the one Benchmark adds) wrapped around such trivial amounts of code serves to swamp the code you're trying to test.

        Get rid of one layer of subroutine call and add an internal multiplier and your results will be both more consistant and a truer reflection of the actual cost of what you are testing. At least they are on my system: 

        use strict;
use warnings;
use Benchmark qw( cmpthese );
use constant DEBUG => 0;

our $N ||= -3;
our $debug = 0;
our $b = 100;

cmpthese $N, {
    constant   => q{
        for( 1 .. 1000 ){ my $a = 10 * $b; $a = 10 * $b if DEBUG; }
    },
    noconstant => q{ for( 1 .. 1000 ){ my $a = 10 * $b; } },
    var => q{
        our $debug;
        for( 1.. 1000 ){  my $a = 10 * $b; $a = 10 * $b if $debug; }
    },
};

__END__
C:\test>junk6 -N=-1
             Rate        var   constant noconstant
var        2222/s         --       -19%       -19%
constant   2731/s        23%         --        -1%
noconstant 2745/s        24%         1%         --

C:\test>junk6 -N=-3
             Rate        var   constant noconstant
var        2133/s         --       -20%       -25%
constant   2662/s        25%         --        -6%
noconstant 2831/s        33%         6%         --

C:\test>junk6 -N=-10
             Rate        var   constant noconstant
var        2172/s         --       -18%       -21%
constant   2659/s        22%         --        -3%
noconstant 2746/s        26%         3%         --

C:\test>junk6 -N=-20
             Rate        var   constant noconstant
var        2171/s         --       -18%       -21%
constant   2632/s        21%         --        -5%
noconstant 2759/s        27%         5%         --

        [download]
        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.
[reply]
[d/l]

In Section Seekers of Perl Wisdom