I just found issue where QM energy was unreasonable high in torsion fitting. This has been corrected by adding an error causing program to crash if too high, otherwise boltsman fitting will take care of it and you wouldnt know something is wrong. Due to recent changes in initial QM OPT/ torsion OPT criteria such as now doing MP2, now doing 5 cycles intead of 3 and tolerances for comparing input structure to "quickly terminated QM OPT" structure, this caused OPT to change, however SP had already been successfully completed on a previous run with a "bad opt structure" that had hydrogens flying away... anyway, not an issue unless restarting. You can manually check your restarted files via command "find . -name "use_weightspng"" (without outer quotes), this will tell you if any fragment use boltsman weights. Sometimes this is good like in partial double bonds (we dont spin amide/acid by default), however there are other cases like trigonal center on b-c in a-b-c-d that cause high energy torsion barrier, so can manually check each case before submitting or, can just rerun poltype and it will crash if one of your fragments has high energy (> 50 kcal/mol). Anyway the fix for me is to delete the "bad" SP log files. If you are deleting any SP log files, also delete all the tinker files too since fitting is after (.key. .xyz..alz,.out)
I just found issue where QM energy was unreasonable high in torsion fitting. This has been corrected by adding an error causing program to crash if too high, otherwise boltsman fitting will take care of it and you wouldnt know something is wrong. Due to recent changes in initial QM OPT/ torsion OPT criteria such as now doing MP2, now doing 5 cycles intead of 3 and tolerances for comparing input structure to "quickly terminated QM OPT" structure, this caused OPT to change, however SP had already been successfully completed on a previous run with a "bad opt structure" that had hydrogens flying away... anyway, not an issue unless restarting. You can manually check your restarted files via command "find . -name "use_weightspng"" (without outer quotes), this will tell you if any fragment use boltsman weights. Sometimes this is good like in partial double bonds (we dont spin amide/acid by default), however there are other cases like trigonal center on b-c in a-b-c-d that cause high energy torsion barrier, so can manually check each case before submitting or, can just rerun poltype and it will crash if one of your fragments has high energy (> 50 kcal/mol). Anyway the fix for me is to delete the "bad" SP log files. If you are deleting any SP log files, also delete all the tinker files too since fitting is after (.key. .xyz..alz,.out)