I have a system that only have a small part of charged particles, and others are set charge equal to 0. I want to ask how can I optimize my script to minimize compute cost, for example, exclude those neutral particles from coulomb potential calculation?
Thanks a lot for your kindly explanation in advanced!
University of Science & Technology of China
What you can do depends on the specific details of your system and your inputs.
Can you provide some minimal example input deck with just a few 100s of atoms that would demonstrate the situation?
these are just data files without the corresponding input. this is useless.
the situation is not as straightforward as you might think. also, it is pointless to discuss what might be without some proper real performance analysis (LAMMPS provides some useful information by default). if you just speculate you are likely to try optimizing what doesn’t need to be optimized or has very little potential for improvement.
please note that the cost of PPPM long range electrostatics is primarily determined by the grid, i.e. the volume of the simulation cell and related settings. so the cost of that part of the calculation needs to be balanced against the corresponding real-space part. that is where the larger potential for optimization is.