running time problem

[unknow1024]

hello,i have some problem that i can not deal it.
i am running the command hhblits_omp -d uniclust30_2017_10/uniclust30_2017_10 -i length200.part_001_fas -o part_part_001.hhr -oa3m part_part_001/part_part_001.a3m -n 3 -cpu 8
the input includes 300k sequence,and its length is less than 200.
But it seems a bit slow?I run it for 11 hour but only get 3800 sequece results.(i need to deal with 10M seqs, it is the training data from a paper)Is it really rational?(i know that 8 threads maybe a bit small)
Another problem is the memory using almost 80G(the command is used for 30k seqs for each runing)when i get 3800 seqs' result.Is it rational?Or maybe is my wrong setting leads to this problem?
best wish.