modify line for spin_cubic, spin_none. edit docs a bit.
This commit is contained in:
@ -84,12 +84,12 @@ The choice of a line search algorithm for the {spin_oso_cg} and
|
||||
{spin_oso_lbfgs} styles can be specified via the {line} keyword.
|
||||
The {spin_cubic} and {spin_none} only make sense when one of those
|
||||
two minimization styles is declared.
|
||||
|
||||
The {spin_cubic} keyword activates the line search procedure when
|
||||
the {spin_oso_cg} algorithm is used.
|
||||
|
||||
The {spin_none} keyword deactivates the line search procedure when
|
||||
the {spin_oso_lbfgs} algorithm is used.
|
||||
The {spin_cubic} performs the line search based on a cubic interpolation
|
||||
of the energy along the search direction. The {spin_none} keyword
|
||||
deactivates the line search procedure.
|
||||
The {spin_none} is a default value for {line} keyword apart from the case when
|
||||
single-replica calculations are performed with {spin_oso_lbfgs} that
|
||||
uses {spin_cubic} line search.
|
||||
|
||||
[Restrictions:] The line search procedure of styles
|
||||
{spin_oso_cg} and {spin_oso_lbfgs} cannot be used for magnetic
|
||||
|
||||
@ -18,7 +18,7 @@ min_style spin_oso_lbfgs :pre
|
||||
[Examples:]
|
||||
|
||||
min_style spin_oso_lbfgs
|
||||
min_modify discrete_factor 10.0 line spin_none :pre
|
||||
min_modify line spin_none discrete_factor 10.0 :pre
|
||||
|
||||
[Description:]
|
||||
|
||||
@ -55,12 +55,21 @@ Style {spin_oso_cg} defines an orthogonal spin optimization
|
||||
(OSO) combined to a conjugate gradient (CG) algorithm.
|
||||
The "min_modify"_min_modify.html command can be used to
|
||||
couple the {spin_oso_cg} to a line search procedure, and to modify the
|
||||
discretization factor {discrete_factor}.
|
||||
discretization factor {discrete_factor}.
|
||||
By defualt, the style {spin_oso_cg} does not employ line search procedure and
|
||||
and uses the adaptive time-step technique in the same way as style {spin}.
|
||||
|
||||
Style {spin_oso_lbfgs} defines an orthogonal spin optimization
|
||||
(OSO) combined to a limited-memory Broyden-Fletcher-Goldfarb-Shanno
|
||||
(LBFGS) algorithm.
|
||||
By default, style {spin_oso_lbfgs} uses a line search procedure.
|
||||
(L-BFGS) algorithm.
|
||||
By default, style {spin_oso_lbfgs} uses a line search procedure
|
||||
based on cubic interpolation for
|
||||
a single-replica calculation, and it does not use line search procedure
|
||||
for a multireplica calculation (such as in case of GNEB calculation).
|
||||
If the line search procedure is not used then the discrete factor defines
|
||||
the maximum root mean squared rotation angle of spins by equation {pi/(5*Kappa)}.
|
||||
The default value for Kappa is 10.
|
||||
|
||||
The "min_modify"_min_modify.html command can be used to
|
||||
deactivate the line search procedure, and to modify the
|
||||
discretization factor {discrete_factor}.
|
||||
|
||||
@ -106,10 +106,10 @@ the number of total force evaluations exceeds {maxeval} :ul
|
||||
NOTE: the "minimization style"_min_style.html {spin},
|
||||
{spin_oso_cg}, and {spin_oso_lbfgs} replace
|
||||
the force tolerance {ftol} by a torque tolerance.
|
||||
The minimization procedure stops if the 2-norm (length) of the
|
||||
global torque vector (defined as the cross product between the
|
||||
spins and their precession vectors omega) is less than {ftol},
|
||||
or if any of the other criteria are met.
|
||||
The minimization procedure stops if the 2-norm (length) of the torque vector on atom
|
||||
(defined as the cross product between the
|
||||
atomic spin and its precession vectors omega) is less than {ftol},
|
||||
or if any of the other criteria are met. Torque have the same units as the energy.
|
||||
|
||||
NOTE: You can also use the "fix halt"_fix_halt.html command to specify
|
||||
a general criterion for exiting a minimization, that is a calculation
|
||||
|
||||
@ -59,7 +59,7 @@ performance speed-up you would see with one or more physical
|
||||
processors per replica. See the "Howto replica"_Howto_replica.html
|
||||
doc page for further discussion.
|
||||
|
||||
NOTE: As explained below, a GNEB calculation performs a damped dynamics
|
||||
NOTE: As explained below, a GNEB calculation performs a
|
||||
minimization across all the replicas. One of the "spin"_min_spin.html
|
||||
style minimizers has to be defined in your input script.
|
||||
|
||||
@ -170,8 +170,8 @@ command is issued.
|
||||
:line
|
||||
|
||||
A NEB calculation proceeds in two stages, each of which is a
|
||||
minimization procedure, performed via damped dynamics. To enable
|
||||
this, you must first define a damped spin dynamics
|
||||
minimization procedure. To enable
|
||||
this, you must first define a
|
||||
"min_style"_min_style.html, using either the {spin},
|
||||
{spin_oso_cg}, or {spin_oso_lbfgs} style (see
|
||||
"min_spin"_min_spin.html for more information).
|
||||
|
||||
Reference in New Issue
Block a user