This thread might have some relevance Perl's poor disk IO performance. Also, wild as it may seem, launching a fast grepper might speed things by just feeding in lines with "^area". Lots of overhead with that approach, but with huge files, might be worth it? I don't know. Benchmarking needed. If the problem is slow Perl line by line I/O, threads won't help.