Warn users about resources when using the plugin with large systems

[yiquintero]

Description

From the LAMMPS plugin review:

Overall, this would be a contribution that needs to explicitly warn about the growing memory use and performance impact of applying it for larger systems, since it will cancel parallelization and data distribution.

Solution

1. Obtain a set of LAMMPS input files to benchmark the plugin
   Define a system that can be easily modified to vary the amount of time steps and total number of particles. The plugin should be used to calculate the 4 quantities and, ideally, the calls to OCTP should be easy to comment-off from the LAMMPS input file.

2. Measure memory consumption and wall-time introduced by the plugin

   1. Run the simulation of each input file and measure the wall-time and memory consumption.
   2. Remove the calls to OCTP from the input files and run the simulations again, also taking note of the wall-time and memory consumption.
   3. Make the following plots:
      a. number of particles vs execution time
      b. number of particles vs memory consumption
      c. time steps vs execution time
      d. time steps vs memory consumption
   4. Based on the information in the plots, determine the number of particles and time steps that should warrant a warning to the user.

3. Determine the best way to warn users about increasing memory resources and execution times.
   A few ideas: Include plots and message in README file, write a message to the LAMMPS simulation log file.