The calculation speed of DPGEN calling VASP is slow

[tc-q]

When i was using dpgen and "nohup dpgen init_bulk param.json machine.json 1>log 2>err&" to submit a task, I find that the fp.log shows as blow

it seems that it only runs on 1 cores, although when i "top", there are 24 cores are used by vasp.
the machine.json shows as blow

why does the question occur? how can i do to solve this problem to make the vasp task run on multiple cores.

[njzjz]

It's likely that your vasp was not compiled with your MPI that is currently used.