I agree with other Monks that unpack is the way to go but if you are more comfortable with split I would use that in combination with splice and join in stages. Something like (not tested)
my @flds = split m{\s+}, $line;
my $memAddr = splice @flds, 0, 1;
my $ascii = splice @flds, -1, 1;
my $data = join q{}, @flds;