For general instructions on preparing the pileup format
http://genomeview.org/content/preparing-pileup
For the tab delimited file, you should include .swig somewhere in the filename.
We typically end up with something like:
bigfile.bam.swig.bgz
In a recent version of GenomeView (1255+) we had to split the functionality of the pileup/SNP track and the short read track which shows the individual reads.
Essentially, the pileup was calculated in real-time, but this proved to be impractical for very large data sets. Therefore we now support the pile-up format from samtools, which has the pile-up and SNP functionality.
The pile-up track is described GUI section in the manual.