Closed wwood closed 9 years ago
Just changed to to not use a tempdir for bamm make, because this may have resulted in /tmp filling up in some cases. Better for it to just work than to run faster but fail in some circumstances.
I'll merge in the next day or so unless there is comment.
No comment! On 03/07/2015 9:33 am, "Ben J Woodcroft" notifications@github.com wrote:
Just changed to to not use a tempdir for bamm make, because this may have resulted in /tmp filling up in some cases. Better for it to just work than to run faster but fail in some circumstances.
I'll merge in the next day or so unless there is comment.
— Reply to this email directly or view it on GitHub https://github.com/Ecogenomics/BamM/pull/26#issuecomment-118193158.
Turns out sorting 250GB BAM files can be very slow, and you really do need lots of CPUs. Before, samtools sort was not multithreaded.
The most recent patch assumes that
bwa
andsamtools sort
don't often work at the same time, so it is safe to reuse the CPUs between these two phases when making (and at worst causes a little kernel scheduling time loss).Also here merged is the next branch, which brings a new
variance
coverage mode, and some better documentation, among other things.