Where did the

1:6

cpu:mem

ratio in nf-core process defaults come from? I’ve been working a lot on getting a pipeline working in AWS and it’s very noticeable to me that this ratio never allows for full utilisation since the vast majority of instances (including the defaults) are all

1:2

,

1:4

, or

1:8

I'm guessing historical bias from some cluster at NGI? I do agree the defaults in the template could do with some more common ratios. All our clusters are 1:8 for example

It isn’t a huge issue for me really, I just redefine my

process_*

labels for my use-case, but I was curious

This sounds like a historical bias from NGI working on the Uppmax systems.

shall we have a ratio resource params to help simplify this setup?

Not sure if NGI is to blame, but yes they have been around for quite some time. > All our clusters are 1:8 for example If RAM prices continue going through the roof like this, we will soon start seeing new clusters with a 1:1 ratio again.