juditzador
commented
1 year ago It is true that at the L1 and L2 levels there are repetitions. It depends how you view it. It is true that in many cases it is a duplication of the same calculations, however, we are somewhat reluctant to simply get rid of this, because we found that this provides robustness for difficult-to-find saddle points. Nevertheless, we are working on a careful solution that will decrease the cost and keep the robustness of the code. However, reagrding the L3 calculations, I think you are not using the correct approach. The necessary set of L3 calculations are collected in a /molpro directory above the kinbot directories, and this only contains the molpro files that need to be run. Nothing needs to be selected manually. It also comes with a batch submission script that should help you submit all jobs just with a single command.
During the PES calculation, many species will appear in different wells, and repeatedly optimized several times, also, generated the L3 submit files several times, which needs to further selected manually. Is it possible to create a global database during PES calculation, and skip the species already calculated to reduce the cost.