Bfast jobs for analyzing AB's SOLiD data vs Illumina data

[genome_anawk1]

Dear Bfast experts,

I am aligning paired end reads to the mouse reference genome and the jobs exceed 40hrs for 5M short reads per parallel job.
Using unpaired option for the analysis each of my parallel jobs take < 5hrs. There are 90 such parallel jobs.
I am using the following command:

#bfast postprocess -n $(pcores -N) -U -t -f $tref -A 1 -i $baf > $sam #unpaired
bfast postprocess -n $(pcores -N) -a 3 -t -f $tref -A 1 -i $baf > $sam # paired end

Q1- Am I doing something incorrect?
Q2- I am seeking suggestions to improve the speed but not compromise on accuracy please.
Q3- Do Illumina and SOLiD paired end analyses differ - so is there a different command to be used.

Earlier when I was analyzing the mate pair library the jobs took a very long time. From the job logs the actual processing time was really lesser than that, but it seems that the jobs were stuck in a non cpu intensive process.

Thanks in advance,
Best,
bfast analyzer.

[nilshomer]

The "-U" option really helps, at very little sensitivity/specificity cost.