Benchmarked a few solutions:

use Benchmark qw(:all);

my $str = 'iiiiiiiiMMMMMMMMMMMooooooooooooMMMMMMMMMMiiiiiMMMMMMMMoooo'
+;

my $these = {
    'count' => sub {
        my $count = 0;
        $count++ while $str =~ /(M+)/g;
        $count;
    },
    'regex' => sub {
        my $count = () = $str =~ /(M+)/g;
    },
    'split' => sub {
        $#{[split /M+/, $str, -1]};
    },
};

printf "%-10s %d\n", $_, $these->{$_}->() for sort keys %$these;
print "\n";

cmpthese(-1, $these);
[download]

Got these:

count      3
regex      3
split      3

          Rate split regex count
split 335814/s    --   -3%  -30%
regex 345739/s    3%    --  -27%
count 476625/s   42%   38%    --
[download]