Slightly OT but I noticed that two solutions
print a string concatenation while another supplied a list of strings to
print. I wondered if there was any performance difference between the two. It seems that using a list is consistently about 10% faster. Here's the benchmark code.
use strict;
use warnings;
use Benchmark q{cmpthese};
my $nullFile = q{/dev/null};
open my $nullFH, q{>}, $nullFile
or die qq{open: $nullFile: $!\n};
my $rcConcat = sub
{
print $nullFH q{a} . q{!} x 5 for 1 .. 10000;
};
my $rcList = sub
{
print $nullFH q{a}, q{!} x 5 for 1 .. 10000;
};
cmpthese (-3,
{
Concat => $rcConcat,
List => $rcList,
});
close $nullFH
or die qq{close: $nullFile: $!\n};
and the output from 10 runs.
$ for i in 1 2 3 4 5 6 7 8 9 10; do
> echo Run $i
> ./spw613222
> done
Run 1
Rate Concat List
Concat 17.4/s -- -9%
List 19.2/s 10% --
Run 2
Rate Concat List
Concat 17.4/s -- -10%
List 19.3/s 11% --
Run 3
Rate Concat List
Concat 17.5/s -- -10%
List 19.4/s 11% --
Run 4
Rate Concat List
Concat 17.5/s -- -9%
List 19.1/s 9% --
Run 5
Rate Concat List
Concat 17.4/s -- -10%
List 19.3/s 11% --
Run 6
Rate Concat List
Concat 17.4/s -- -10%
List 19.4/s 11% --
Run 7
Rate Concat List
Concat 17.5/s -- -10%
List 19.4/s 11% --
Run 8
Rate Concat List
Concat 17.5/s -- -10%
List 19.4/s 11% --
Run 9
Rate Concat List
Concat 17.5/s -- -10%
List 19.4/s 11% --
Run 10
Rate Concat List
Concat 17.4/s -- -10%
List 19.3/s 11% --
$
Cheers,
JohnGG