[lammps-users] Parallel efficiency

Dear all,

I am trying appending an additional degree of freedom to the Atom Class. I found that only one CPU was on 100% of its power when the program was running in parallel. Is this related to the communication process of mpich? Could you please tell me what should be done to speed up the running. Any suggestion is grateful, thanks.

Best regards,

Yonggang Zheng

Dalian University of Technology

When you run a LAMMPS problem does it print out (at the end)
that it ran on multiple processors, in the timing breakdown?
If not, then you really didn't run it in parallel.


2009/3/19 Yonggang <[email protected]>: