Adjust the temperature with a canonical sampling thermostat that uses global velocity rescaling with Hamiltonian dynamics (temp/csvr) -(Bussi1), or Langevin dynamics (temp/csld) (Bussi2). -In the case of temp/csvr the thermostat is similar to the empirical -Berendsen thermostat in temp/berendsen, but -chooses the actual scaling factor from a suitably chosen (gaussian) -distribution rather than having it determined from the time constant -directly. In the second case the velocities are updated to a linear -combination of the current velocities with a gaussian distribution -of velocities at the desired temperature. Both termostats are applied -in every timestep. +(Bussi1), or Langevin dynamics (temp/csld) +(Bussi2). In the case of temp/csvr the thermostat is +similar to the empirical Berendsen thermostat in +temp/berendsen, but chooses the actual +scaling factor from a suitably chosen (gaussian) distribution rather +than having it determined from the time constant directly. In the +second case the velocities are updated to a linear combination of the +current velocities with a gaussian distribution of velocities at the +desired temperature. Both termostats are applied in every timestep.
The thermostat is applied to only the translational degrees of freedom for the particles, which is an important consideration for finite-size particles which have rotational degrees of freedom are being -thermostatted with these fixes. The translational degrees of freedom can -also have a bias velocity removed from them before thermostatting +thermostatted with these fixes. The translational degrees of freedom +can also have a bias velocity removed from them before thermostatting takes place; see the description below.
The desired temperature at each timestep is a ramped value during the @@ -78,10 +78,10 @@ time. Thus it is easy to specify a time-dependent temperature.
IMPORTANT NOTE: Unlike the fix nvt command which performs Nose/Hoover thermostatting AND time integration, these fixes -do NOT perform time integration. They only modify velocities to -effect thermostatting. Thus you must use a separate time integration -fix, like fix nve to actually update the positions of -atoms using the modified velocities. Likewise, these fixes should not +do NOT perform time integration. They only modify velocities to effect +thermostatting. Thus you must use a separate time integration fix, +like fix nve to actually update the positions of atoms +using the modified velocities. Likewise, these fixes should not normally be used on atoms that also have their temperature controlled by another fix - e.g. by fix nvt or fix langevin commands. diff --git a/doc/fix_temp_csvr.txt b/doc/fix_temp_csvr.txt index 024998e0e6..1d78a8625f 100644 --- a/doc/fix_temp_csvr.txt +++ b/doc/fix_temp_csvr.txt @@ -30,21 +30,21 @@ fix 1 all temp/csld 100.0 300.0 10.0 123321 :pre Adjust the temperature with a canonical sampling thermostat that uses global velocity rescaling with Hamiltonian dynamics ({temp/csvr}) -"(Bussi1)"_#Bussi1, or Langevin dynamics ({temp/csld}) "(Bussi2)"_#Bussi2. -In the case of {temp/csvr} the thermostat is similar to the empirical -Berendsen thermostat in "temp/berendsen"_fix_temp_berendsen.html, but -chooses the actual scaling factor from a suitably chosen (gaussian) -distribution rather than having it determined from the time constant -directly. In the second case the velocities are updated to a linear -combination of the current velocities with a gaussian distribution -of velocities at the desired temperature. Both termostats are applied -in every timestep. +"(Bussi1)"_#Bussi1, or Langevin dynamics ({temp/csld}) +"(Bussi2)"_#Bussi2. In the case of {temp/csvr} the thermostat is +similar to the empirical Berendsen thermostat in +"temp/berendsen"_fix_temp_berendsen.html, but chooses the actual +scaling factor from a suitably chosen (gaussian) distribution rather +than having it determined from the time constant directly. In the +second case the velocities are updated to a linear combination of the +current velocities with a gaussian distribution of velocities at the +desired temperature. Both termostats are applied in every timestep. The thermostat is applied to only the translational degrees of freedom for the particles, which is an important consideration for finite-size particles which have rotational degrees of freedom are being -thermostatted with these fixes. The translational degrees of freedom can -also have a bias velocity removed from them before thermostatting +thermostatted with these fixes. The translational degrees of freedom +can also have a bias velocity removed from them before thermostatting takes place; see the description below. The desired temperature at each timestep is a ramped value during the @@ -67,10 +67,10 @@ time. Thus it is easy to specify a time-dependent temperature. IMPORTANT NOTE: Unlike the "fix nvt"_fix_nh.html command which performs Nose/Hoover thermostatting AND time integration, these fixes -do NOT perform time integration. They only modify velocities to -effect thermostatting. Thus you must use a separate time integration -fix, like "fix nve"_fix_nve.html to actually update the positions of -atoms using the modified velocities. Likewise, these fixes should not +do NOT perform time integration. They only modify velocities to effect +thermostatting. Thus you must use a separate time integration fix, +like "fix nve"_fix_nve.html to actually update the positions of atoms +using the modified velocities. Likewise, these fixes should not normally be used on atoms that also have their temperature controlled by another fix - e.g. by "fix nvt"_fix_nh.html or "fix langevin"_fix_langevin.html commands.