Open wasade opened 9 months ago
Was running out the door before. We've been doing processing at 5gb chunks, and requesting 120GB to ensure we have sufficient shared memory, until barnacle supports shared memory as a resource. We have been using 16 cores but our efficiency great, so recommend reducing to 10 or 12
...I think this covers the differences.
The changed demux method is here (
/home/mcdonadt/2023.08.09-pangenome-human-depletion/human-depletion-se-pe/demux-inline-all.py
). The splitter is (/home/mcdonadt/2023.08.09-pangenome-human-depletion/human-depletion-se-pe/splitter/target/release/splitter
) with the source in the corresponding parent directory. This code depends as well onfastq_pair
which is installable from bioconda