jessieren / DeepVirFinder

Identifying viruses from metagenomic data by deep learning
Other
116 stars 32 forks source link

Server's running time #24

Open paulaistvan opened 3 years ago

paulaistvan commented 3 years ago

Hello,

I'm using DeepVirFinder for some big fasta files like 200 to 300+ mb in size, but its taking around 6 hours to run even after supplying 100 CPUs using -c 100 tag.

I have been reading documentation and other issues listed, and its mentioned that its better to supply GPUs but on the server they are not available.

Can you please answer below questions?

  1. Right now i'm using minimum sequence length threshold of 300, increasing this can help? and what would be the impact of that, is it good to increase it?
  2. Also, increasing CPUs can help with performance, if 100 CPUs take roughly 6 hours for 250 mb fasta file, can i increase them as GPUs are not available, also with GPU how better is performance?