Rough simulations that explore how the number of salt pairs varies with the chemical potential

[gregoryross]

Aims

1. See if saltswap simulations produced reasonable results when salt molecules were titrated in by varying the chemical potential.
2. To make a preliminary test of using SAMS to calculate relative hydration free energies of salt and water.

Results in brief

• The most stable simulations were produced using the CUDA GPU kernal. More work is required to verify this.
• Using the CUDA, SAMS (with the binuary update scheme) was used to predict the relative hydration free energies.
  • With 3 repeats SAMS predicted $\Delta F = 317.18 +/- 0.32$ (in thermal units at 300K)
  • Using BAR on the NCMC work values from the SAMS simulations predicted $\Delta F = -317.28 +/- 0.21$ (in thermal units at 300K)
  • Calculations on from the calc_gradients branch (with TI, MBAR to follow) predicted $\Delta F = -316.91 +/- 0.19$ (in thermal units at 300K)
  • (Errors are 95% confidence intervals).