Hi all, as part of this hackathon we're going to next week, we'll be looking at

mmseqs(-gpu)

with some NVIDIA life sciences people. 1. We'd like to add the

makepaddedseqdb

function to the collection of small function-specific nf-core modules. 2. At the same time we thought we'd expose

--gpu

arguments in the module since it will be a common use of this module. However I wanted to check what the thoughts were on exposing them versus having workflow-contained

ext.args

1.3 of the nf-core modules doc states non-mandatory args must be provided from modules.config. I guess when running in GPU mode these arguments are common, and there are several nextflow pipelines which would also need modified

modules.config

modified to enable GPU mode (e.g. our colleague's nf-binder-design which adopted proteinfold's mmseqs/colab process) Looking for any strong opinions before we initiate work next week.

Hi @Keiran Rowell I think it would be better to post it in #CJRH30T6V to give it more visibility and have opinions of more people

Okay thanks for directing to a more relevant place!