We try to use -er=150 to remove the strange broken in the phase diagram. But it becomes too slow due to the large er=150. Is it to use phb in parallel or separate the calculation to several parts?
Sorry I don’t have a parallel version of phb. However, you could used multiple copies of emc2 to explore where a phase transition occur at various T and used to those input to launch separate phb run with different starting points (specified by -T and -mu obtained from the various emc2 runs).
Now I should mention that the jumps in the phase diagram is unlikely to be due to a too small -er . There could be a genuine phase transition. Or try reducing -dx .