[gmx-users] difference between steep and cg

[Ming Tang]

Dear Gromacs experts,

I am quite confused about the difference between the energy minimization algorithm steep and cg. According to my experience, MPI is not suitable for cg. Sometimes, even if the system has already converged using steep algorithm, it can be further minimized if one changes the algorithm to cg. However, most of the publications choose to use steep descent algorithm. Which one is more accurate? Is cg only suitable for specific analysis like normal mode analysis?

Thanks in advance,