## This file provides the time taken to read a FASTA file in parallel.

   The system is a Core i7 Quad-Core (Haswell) at 2.6 GHz running OS X v10.9.5.
   Testing occured from a RAMdisk volume to factor out disk IO.

   fasta_rdr1.pl: serial iterator, parallel consumers
   fasta_rdr2.pl: parallel iterator mode
   fasta_rdr3.pl: parallel callback mode

## Synopsis

   fasta_rdr [ /path/to/fastafile.fa [ trim_seq ] ]

   NPROCS=4 perl fasta_rdr hg19.fa   0   run with 4 workers, no trimming
   NPROCS=8 perl fasta_rdr uniref.fa 1   run with 8 workers, trim white space

## Results for 2 million sequences extracted from UniRef100.fasta.gz.

   $ gunzip -c uniref100.fasta.gz | head -15687827 > uniref.fa
   1.0 GiB, 15.7 million lines

   -- extract id, description, and sequence (no trimming)

   NPROCS=4 perl fasta_rdr1.pl uniref.fa 0           | wc -l    29.041s

   NPROCS=1 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 0 | wc -l     6.976s
   NPROCS=2 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 0 | wc -l     3.506s
   NPROCS=4 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 0 | wc -l     2.100s

   NPROCS=1 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 0 | wc -l     3.341s
   NPROCS=2 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 0 | wc -l     1.637s
   NPROCS=4 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 0 | wc -l     0.993s

   -- trim white space

   NPROCS=4 perl fasta_rdr1.pl uniref.fa 1           | wc -l    31.748s

   NPROCS=1 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 1 | wc -l     8.947s
   NPROCS=2 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 1 | wc -l     4.564s
   NPROCS=4 CKSIZE=4m perl fasta_rdr2.pl uniref.fa 1 | wc -l     2.682s

   NPROCS=1 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 1 | wc -l     5.472s
   NPROCS=2 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 1 | wc -l     2.735s
   NPROCS=4 CKSIZE=4m perl fasta_rdr3.pl uniref.fa 1 | wc -l     1.577s