Tailfindr slow
guillaume-vares opened this issue · 1 comments
guillaume-vares commented
In Nanotail module, Tailfindr is extremely slow in our cluster. It takes several days to compute polyA tails for 2 million transcripts. Is there a way to tweak settings to improve speed? (multithreading, parallel commands). We use Singularity containers, and we are not familiar with configuration.
lucacozzuto commented
Tailfindr is below the control of the label big_mem_cpus in the configuration file nextflow.config within the folder. You can add a queue name that is suitable for your cluster and increase the number of cpus. Have a look if you are really using 8 cpus in your node. PS: the program should already be running in parallel on each multi fast5 subset.