MAFFT cpu usage - Githubissues

AusARG / pipesnake

ausarg/pipesnake is a bioinformatics best-practice analysis pipeline for phylogenomic reconstruction starting from short-read 'second-generation' sequencing data.

MIT License

7 stars 2 forks source link

MAFFT cpu usage #4

Closed IanGBrennan closed 8 months ago

IanGBrennan commented 9 months ago

Description of feature

MAFFT by default seems to use just a single CPU (though not completely certain of this). So, specifying the number of CPU per task with --thread -1 (following this) might use resources better.

Keeping in mind the hierarchy of the base.config / modules.config / specified parameters.

IanGBrennan commented 9 months ago

This is a more general suggestion than just for MAFFT. We can do the same for IQTREE by specifying -T AUTO. I should check remaining programs to see if it's the same.

IanGBrennan commented 9 months ago

I'm now using this as a list double-checking the programs that can be parallelized.

BBMAP (includes DEDUPE and FILTER):

threads=auto (t) Set to number of threads desired. By default, uses all cores available.

TRIMMOMATIC:

-threads
this is already implemented

PEAR:

-j
this is already implemented

SPADES:

--threads
already implemented as --threads $task.cpus but should it be --threads {task.cpus} ?

GBLOCKS: single-thread

MAFFT: (see above post)

IQTREE: (see above post)

IanGBrennan commented 8 months ago

resolved here: https://github.com/AusARG/pipesnake/commit/2d36aeaccf377d3f47041f85c4032b10b7fb2045