Porechop uses as many cpu as possible no matter the parallelization I wrote.

It might detect and use whatever is asked. It might even do that in parallel somehow

It seems a pretty lightweight computation still.