How to estimate the duration of building a Centrifuge index, with a given hardware and input-sequences size?

Question

[patrickdm]

Hello.

I just started my first attempt in building a Centrifuge index for archaea+bacteria with an input-sequences.fna of ~69GB (after downloading and joining the refseqs).

centrifuge-build is running on a linux machine with an AMD Ryzen 7 2700X(16) @3.700GHz processor and 32GB of memory, using 8 threads.

I'm wondering how long could it take, assuming it won't crash somewhere along the process for lack of resources, or killed prematurely by human impatience -of mine- ;)

Does anyone have some references/experiences for the process duration with any hw/input-size? Thank you.

Edit. I realize I should have started with a much smaller dataset, and benchmark that :-| I guess it is what I'm going to do.