Hi Joel,
Thanks for writing to us. So, the slowness could be due to the diamond version. Version 2 is actually much slower than version 1 because a bug was fixed, you can read more about it here: Announcing HUMAnN 3.6 (Critical Update).
From the log it seems that one genome was mapped in ~20m, right? But the error after that is due to you killing PhyloPhlAn because too slow?
Also, you’re specifying 4 cores, which means 4 input genomes will be mapped at the same time, to speed up the process you can specify more (if you can) so that more genomes will be mapped in the same amount of time.
Thanks a lot,
Francesco