The  /o modifier means "compile once" for a regex it modifies. Consider these examples:

>perl -wMstrict -le
"my $s = '1a2b3c';
 ;;
 print qq{no /o};
 for my $i (qw(3 2 1)) {
   print qq{matched '$1'} if $s =~ m{ ($i.) }xms;
   }
 ;;
 print qq{with /o};
 for my $i (qw(3 2 1)) {
   print qq{matched '$1'} if $s =~ m{ ($i.) }xmso;
   }
"
no /o
matched '3c'
matched '2b'
matched '1a'
with /o
matched '3c'
matched '3c'
matched '3c'
[download]

The function of the  /o modifier has been generally replaced by the qr// regex object builder (see in perlop).

I was a bit surprised not to see anything about  /o in perlre, but it is (briefly and obliquely) discussed in qr/STRING/msixpodual (5.14), and the following remains in perlretut (at least through 5.12):

Optimizing pattern evaluation

We pointed out earlier that variables in regexps are substituted before the regexp is evaluated:

$pattern = 'Seuss';
     while (<>) {
         print if /$pattern/;
     }
[download]

This will print any lines containing the word "Seuss". It is not as efficient as it could be, however, because Perl has to re-evaluate (or compile) $pattern each time through the loop. If $pattern won't be changing over the lifetime of the script, we can add the "//o" modifier, which directs Perl to only perform variable substitutions once:

#!/usr/bin/perl
     #    Improved simple_grep
     $regexp = shift;
     while (<>) {
         print if /$regexp/o;  # a good deal faster
     }
[download]