These might be a little bit unfair at small numbers of trajectories, since the Metropolis stuff we’re comparing to (calculated in previous run) blog is not sorted.
Other associated files:
- mod_frenkel_std.job
- frenkel_s00/modified_frenkel_s00.inp
- modified_frenkel_s00.out
- frenkel_s01/modified_frenkel_s01.inp
- modified_frenkel_s01.out
- frenkel_s02/modified_frenkel_s02.inp
- modified_frenkel_s02.out
- frenkel_s03/modified_frenkel_s03.inp
- modified_frenkel_s03.out
- frenkel_s04/modified_frenkel_s04.inp
- modified_frenkel_s04.out
- frenkel_s05/modified_frenkel_s05.inp
- modified_frenkel_s05.out
- frenkel_s06/modified_frenkel_s06.inp
- modified_frenkel_s06.out
- frenkel_s07/modified_frenkel_s07.inp
- modified_frenkel_s07.out
- frenkel_s08/modified_frenkel_s08.inp
- modified_frenkel_s08.out
- frenkel_s09/modified_frenkel_s09.inp
- modified_frenkel_s09.out
- modified_frenkel.data
- ../../frenkel/metrop_final.data
Tags: fbivr, frenkel_testing, grossman, hess_normal, hk_overlap_coherent, hk_prefactor_maslov, manolopoulos, mc_generate_frenkel, morse
Note: this only measures how much the calculations differ from each other. It still tells you nothing about the actual errors in the calculation (until long time, at least). You need to plot something to see how well converged the results are — at least plot the difference between the avg. |C(t=t’)| for the trajectories and the fully converged value of |C(t=t’)|.
Ideally, there’s probably an even better measure of something like the RMS deviation from that converged result (but that would probably require more work.)