I have some very basic questions for which I cannot find documented answers.
(1) For the Ising models or more generally for classical spin models is it possible to view the final spin configuration of a system after the simulation has run its course?
(2) Is it possible to observe the magnetization and not just the absolute value of the magnetization?
Thanks in advance, -jonas
On Oct 22, 2012, at 7:53 PM, Jonas Anderson jonastyleranderson@gmail.com wrote:
I have some very basic questions for which I cannot find documented answers.
(1) For the Ising models or more generally for classical spin models is it possible to view the final spin configuration of a system after the simulation has run its course?
You can easily add this to the code, but it usually does not make much sense since the correct result is an ensemble of configurations weighted with the respective Boltzmann weight. DO you want to view all configurations or just the final one?
(2) Is it possible to observe the magnetization and not just the absolute value of the magnetization?
Yes, and it is stored into the results, but it should be 0 due to symmetry unless you add a magnetic field.
Best regards
Matthias Troyer
Thanks in advance, -jonas
-- Jonas T. Anderson Post-doc Université de Sherbrooke
On Mon, Oct 22, 2012 at 2:16 PM, Matthias Troyer troyer@phys.ethz.chwrote:
On Oct 22, 2012, at 7:53 PM, Jonas Anderson jonastyleranderson@gmail.com wrote:
I have some very basic questions for which I cannot find documented
answers.
(1) For the Ising models or more generally for classical spin models is
it possible to view the final spin configuration of a system after the simulation has run its course?
You can easily add this to the code, but it usually does not make much sense since the correct result is an ensemble of configurations weighted with the respective Boltzmann weight. DO you want to view all configurations or just the final one?
Sorry, I meant just the final one.
(2) Is it possible to observe the magnetization and not just the
absolute value of the magnetization?
Yes, and it is stored into the results, but it should be 0 due to symmetry unless you add a magnetic field.
Ok, and it is simply the observable 'Magnetization' ? I'm hoping to choose an initial state such as all spin up and then prevent the system from completely thermalizing and observe some residual magnetization. Is it possible to specify the initial state for Ising models?
Thanks again, -jonas
On Oct 22, 2012, at 8:30 PM, Jonas Anderson jonastyleranderson@gmail.com wrote:
On Mon, Oct 22, 2012 at 2:16 PM, Matthias Troyer troyer@phys.ethz.ch wrote:
On Oct 22, 2012, at 7:53 PM, Jonas Anderson jonastyleranderson@gmail.com wrote:
I have some very basic questions for which I cannot find documented answers.
(1) For the Ising models or more generally for classical spin models is it possible to view the final spin configuration of a system after the simulation has run its course?
You can easily add this to the code, but it usually does not make much sense since the correct result is an ensemble of configurations weighted with the respective Boltzmann weight. DO you want to view all configurations or just the final one?
Sorry, I meant just the final one.
Just add a few lines of code to print the final spin value. You can add this for example into the save function which is called when saving the results.
(2) Is it possible to observe the magnetization and not just the absolute value of the magnetization?
Yes, and it is stored into the results, but it should be 0 due to symmetry unless you add a magnetic field.
Ok, and it is simply the observable 'Magnetization' ? I'm hoping to choose an initial state such as all spin up and then prevent the system from completely thermalizing and observe some residual magnetization. Is it possible to specify the initial state for Ising models?
The spins actually start in a polarized configuration. To get others you will have to modify the constructors of the SpinSim class.
If you use cluster updates the magnetization will be zero immediately since cluster updates are very efficient.
Matthias Troyer
comp-phys-alps-users@lists.phys.ethz.ch