Using Checkpoints for hysteresis loop calculations with varying field steps

[Maryam]

Dear VAMPIRE users,

I am working on calculating a hysteresis loop for a spinel system using VAMPIRE 7.0.

I would like to know if the checkpoint system is the recommended approach for generating a hysteresis loop where I need to change the applied magnetic field at each step. Specifically, I intend to use different field step sizes (e.g., 0.05 T near the coercive field and 0.1 T for saturation regions).

If this "chaining" of simulations via checkpoints is the correct workflow, could you please clarify:

1. Is it sufficient to just load the checkpoint for each new field step, or are there any specific parameters in the input or material files that I must modify or remove between runs to avoid conflicts?

2. Does the checkpoint automatically handle the transition, or is there a standard practice to maintain consistency across these sequential field-dependent steps?

Thank you for your guidance.

Best regards, Mary

[Richard Evans]

Hi Mary,

Yes chaining like this should work, if a bit clunky. I would recommend just doing a series of time-series calculations driven with bash/python at constant field (set in the input file) and using the checkpoints for this, which should work fine. For T = 0 the static-hysteresis-loop program allows an exit condition for low torque which significantly improves the speed - you can set a large number of steps and it will exit for fields far from the coercivity. For T > 0 then the solution you propose will work, but I would recommend fine field steps (eg 0.0001 T) and low loop steps - in my experience this improves convergence of the loop significantly.

Best,

Richard

[Maryam]

Dear Richard,

Thank you very much for the clear and helpful guidance. I will proceed with the bash scripting approach and adopt the strategy of using fine field steps (e.g., 0.0001 T) for my T > 0 simulations.

To ensure that I implement your advice correctly, could you please clarify the approximate numerical range that you would recommend for "low loop steps"?

Specifically, when using a very fine field step such as 0.0001 T, what would be a reasonable starting value for sim:equilibration-time-steps and sim:loop-time-steps at each field point to achieve a smooth quasi-static transition without unnecessary computational overhead?

Additionally, given the potentially high computational duration of using such fine field steps at T > 0, would you recommend utilizing checkpoints to manage these runs, or is it generally more practical to execute them in a single continuous run?"

Thank you again for your time and support.

Best regards, Mary

[Pravin Sharma]

Hi all, 

I am trying to get the hysteresis for the FM/AFM bilayer with intermixing. I am constantly getting the issue here. When I perform field cooling, save the checkpoint file, and reload it for the hysteresis loop calculation, I observe the exchange bias in the opposite direction. What mistake am I doing? This is very crucial for my project, please help me out here. 

I am using: 

For field cooling: 
sim:integrator-random-seed = 17
sim:equilibration-time-steps=100000
#sim:time-steps-increment = 1000
sim:total-time-steps = 2500000
sim:time-step = 1.0e-15 !s

#==== Field related====================
sim:applied-field-strength = 1.0 !T
sim:applied-field-unit-vector = 0.0, 0.0, 1.0
sim:save-checkpoint = end               #in the field cooling calculations; 


For hysteresis: 

sim:integrator-random-seed = 17
sim:equilibration-time-steps=100000
sim:loop-time-steps = 1000000
#sim:time-steps-increment = 10
sim:time-step = 1.0e-15 !s

#==== Field related====================
sim:maximum-applied-field-strength = 1.0 !T
sim:applied-field-strength-increment = 0.01 !T
sim:applied-field-unit-vector = 0.0, 0.0, 1.0
#============================================
sim:program = hysteresis-loop
sim:integrator = llg-heun
sim:load-checkpoint = restart 

Thank you in advance

Best regards, 

Pravin Sharma