Improve filtering of unpaired reads

PairedFastqComparator should be improved.
Currently it uses lists for storing unpaired reads and FASTQIterator with FastqFormat for reading and writing (by line).
The lists should be replaced with hash tables and reading and writing should be performed by chunks.
Improve any other part of task if possible.

Currently the task is running for more then 20 hours for the files of 6.5Gb.