>>185
追試。環境は ActivePerl Build810 (5.8.4)

#!/usr/local/bin/perl
use Benchmark;
our @list = qw(hoge fuga piyo foo bar baz);

sub t_l { /hoge/ && return $_ for @_ }

timethese(-5, {
　　g_b => sub { my $a = (grep {/hoge/} @list)[0] },
　　g_e => sub { my $a = (grep /hoge/, @list)[0] },
　　g_l => sub { my($a) = grep /hoge/, @list },
　　t_l => sub { my $a = t_l(@list) },
});
__END__

Benchmark: running g_b, g_e, g_l, t_l for at least 5 CPU seconds...
g_b: 6 wallclock secs ( 5.16 usr + 0.00 sys = 5.16 CPU) @ 49723.26/s (n=256572)
g_e: 5 wallclock secs ( 5.32 usr + 0.00 sys = 5.32 CPU) @ 193839.10/s (n=1031224)
g_l: 5 wallclock secs ( 5.11 usr + 0.00 sys = 5.11 CPU) @ 199724.07/s (n=1020590)
t_l: 5 wallclock secs ( 5.33 usr + 0.00 sys = 5.33 CPU) @ 193475.42/s (n=1031224)