Closed coltonbh closed 3 weeks ago
The convergence was determined for the last structure that is written in crestopt.log
and technically there is one more displacement and transformation from internal to Cartesian coordinates before writing to crestopt.xyz
.
Either should be fine and converged, but the printouts correspond to the last in the log file.
Got it. So to confirm, the crest.engrad
corresponds to the last structure in the crestopt.log
file and the crestopt.xyz
structure results from taking that final gradient and applying a step to the last structure in crestopt.log
, correct? There are no energy/gradient calculations on the final structure in crestopt.xyz
, correct?
yes
Describe the issue I'm trying to know which xyz structure matches the gradient values produced by an optimization run in the
crest.engrad
file. The final structure increstopt.log
is different from the structure increstopt.xyz
. Should they be the same? Or does the log omit the final structure? Which structure corresponds to the values increst.engrad
?Thanks so much!
To Reproduce
input.toml
file:structure.xyz
file:Command:
crest input.toml
The last xyz in the
crestopt.log
:The structure in
crestopt.xyz
:Which of these structures corresponds to the gradient in
crest.engrad
? Doescrest.log
omit the final structure? My calculation converges in 206 cycles and there are 206 structures increstopt.log
, so I do not understand how thecrestopt.xyz
is not one of them.Thanks for your help!