During the torsion angle dynamics in the flexref module, the time steps are increased by a factor defined by tadfactor for the rigid body stages of the protocol. This in combination with fully flexible segment definitions might lead to failures (the MD becomes unstable and crashes). In haddock2.X those cases were handled by the python code and the model was rerun with a smaller factor.
The flexref CNS code should automatically detected the presence of fully flexible segments and reduce the tadfactor from the default value of 8 to e.g. 4.
Motivation
This will cause less model failures and more stable runs.
Desired feature/enhancement
During the torsion angle dynamics in the
flexref
module, the time steps are increased by a factor defined bytadfactor
for the rigid body stages of the protocol. This in combination with fully flexible segment definitions might lead to failures (the MD becomes unstable and crashes). In haddock2.X those cases were handled by the python code and the model was rerun with a smaller factor.The
flexref
CNS code should automatically detected the presence of fully flexible segments and reduce thetadfactor
from the default value of 8 to e.g. 4.Motivation
This will cause less model failures and more stable runs.