Comment by project2501a

What I am working on is my masters thesis in bioinformatics supports/HPC: "Predicting the running time of bioinformatics tools", by running (initially) 5 tools 1000 times each with different parameters and then fitting a curve.

I am pretty sure I can get 70% predation rates +/- 10% . Unfortunately, I'm blocked by the lack of hardware. Kind of not-quite school affiliated (so I cannot really ask for national computing resources), so I am trying to build a single threaripper pro node on my own. Hurts the wallet, but if added to slurm as module, this can have implications.