update documentation
This commit is contained in:
@ -67,6 +67,16 @@ resampled from a normal distribution:
|
||||
The algorithm then continues, proposing a new configuration of particles
|
||||
and velocities N integration steps later.
|
||||
|
||||
Typically, HMC is run with a larger timestep than molecular dynamics.
|
||||
The timestep provides control over the acceptance ratio. A larger
|
||||
timestep will lead to larger and more extreme MC moves which are
|
||||
less likely to be accepted.
|
||||
|
||||
This fix is not designed to be used with anything but an NVE simulation.
|
||||
Only atom-specific data are restored on MC move rejection, so anything which
|
||||
adds or remove atoms, changes the box size, or has some external state not
|
||||
dependent on atomic data will have undefined behavior.
|
||||
|
||||
----------
|
||||
|
||||
The keyword/value options are used in the following ways.
|
||||
|
||||
Reference in New Issue
Block a user