If performance is in any way a criteria, don't use File::Slurp. And if the size of the file is any more than a few hundred lines, don't use list assignment either:

#! perl -slw
use strict;
use File::Slurp;
use Benchmark qw[ cmpthese ];

print `wc -l $ARGV[ 0 ]`;

open our $infile, '<', $ARGV[ 0 ] or die $!;
our $I //= -1;

cmpthese $I, {
    a=> q[
        sysseek $infile, 0, 0;
        my @lines = read_file( $infile, chomp => 1 );
        print 'a: ', scalar @lines if $I == 1;
    ],
    b=> q[
        sysseek $infile, 0, 0;
        chomp( my @lines = <$infile> );
        print 'b: ', scalar @lines if $I == 1;
    ],
    c=> q[
        sysseek $infile, 0, 0;
        my @lines; $lines[ @lines ] = <$infile> until eof $infile;
        chomp @lines;
        print 'c: ', scalar @lines if $I == 1;
    ],
};

__END__
C:\test>t-slurp -I=1 small.tsv
1000000 small.tsv

a: 1000000
            (warning: too few iterations for a reliable count)

b: 1000000
            (warning: too few iterations for a reliable count)

c: 1000000
            (warning: too few iterations for a reliable count)

  s/iter     a     b     c
a    257    --  -98%  -98%
b   4.94 5115%    --  -16%
c   4.14 6119%   19%    --

C:\test>t-slurp -I=-1 small.tsv
1000000 small.tsv

            (warning: too few iterations for a reliable count)

            (warning: too few iterations for a reliable count)

     s/iter          b          a          c
b      4.88         --        -4%      -100%
a      4.69         4%         --      -100%
c 1.39e-006 350525381% 337007680%         --
[download]