Releases

version:0.54_4

• bug corrections for mi versions
• doc implementation

version:0.53_32

• mpi version with halo & bounds particles in network

version:0.52_0

(20/03/25)

• mpi memory optimzed (S0,H,dS with slice particles number)
• stochastic J function name added
• mpi tests all passed.

version:0.50_5

(25/02/25)

• adding mpi/omp/serial packages for network
• modifing structures & network builders
• adding a randomJ simulations

version:0.40_0

(08/11/24)

• adding macro cell dipolar & macro cell external demagnetized operators primary tests
• adding options :
  • use-macro-cell-dipolar : true to use macro cell dipolar operator instead of deipolar operator
  • demagnetized-updating-steps-number : steps number to update demagnetized field on macro cells (0:macro cell dipolar operator >0:acro cell external demagnetized operator)
  • macro-cell-size : macro cell size per direction in Angstrom
  • macro-cell-mergin : macro cell margin to build the macro cells
• bug correction
  • in SM_LandauLifschitzSystem::computeNetTorque() for parallel version
  • in SM_HeissenbergOperatror::computeMagneticField() for not alone operator

version:0.35_9

(22/10/24)

• add an option to save S or not at end of simulation –is-S-saved
• begin and start index of simulations introduced
• adding duration for simulation of curie temperature
• modiffing the time step cycle for using log of time steps
• adding profiling routines in LLSystem class
• adding DipolarMap function for computing Hdip over a network whose S is initiliated to S0 options
• adding a macro cell dipolar field
• changing the MMC algorithm:
  • a MMC-alorithm consists in computing P MMC-spin where P is the number of particles
  • one simulation consists in computing steppers-number MMC-algorithms
• changing the adimensioning varaible with constants values
• changing the computing of esperance of mean S
• bug corrections:
  • multicores CT & MMC temperature range NAN detected & bag global esperance
  • error in time step cycle and in stochastic output list for MPI version

version:0.34_8

(30/09/24)

• Error of factor 0.5 for computing anisotropy energy & field
• removing computeMagneticFieldAndEnergy replaced by computeMagneticField
• energy is no longer computed at any step
• adding a temperature range & time step range
• hysteresis stochastuc output repalced by Hysteresis Cycle & empty strochastic output
• adding a steppers-number option which replaces time-steppers-number used also for MMC method
• adding cycles for monocore/mulitcores
• bugs corrections when testing

version:0.33_2

(24/09/10)

• adding the monte carlo algorithm
• modifying the stochatic functions seed
• re-organisation of the packages
• degugging parllele error
• Parallel MC is not valid

version:0.32

(24/09/04)

• adding a time stepper analysis for time curie temperature method
• adding a Monte-Carlo skeleton classes

version:0.31

(24/08/27)

• adding BCC/FCC structure in parallell
• adding periodic structure for network
• Wdt=sqrt(dt) and the relation eps(T) is adapted
• adding hysteresis package
• adding Boltzmann package

version:0.30

(24/07/10)

• adding material / bcc network / Heun system
• removing the adaptative time step
• creating modules
• testing validation

version:0.20

(24/05/17)

• bug correction on jump for curie temperatire
• adding an options to make adaptative time step

version:0.19

(24/03/21)

• bug coorection on printing options with help
• adding an options to print log only for core 0 –log-per-core={true,false}

version:0.18

(24/03/20)

• modiffing math/random functions base on std, boost & external trng library
• modifiing the option random into random-distribution with values in ("boost-normal","std-normal","trng-normal")
• adding the option random-generator with values in ("mt19937","lcg64","mrg2","mrg3","mrg3s","mrg4","mrg4s","mrg5","yarn2")
• adding a jump random number generator capability for MPI simulations

ersion:0.17

(24/03/12)

• adding MPI for simulations loop for Curie Temperature and for noise rates loop for time curie simlations
• the seed used for time curie temperature simulation is constant for all noise rate
• the seed used for curie temperature simulation depends on the cores
• adding mpi module

version:0.16

(24/03/08)

• changing the name stochMicrom->stochMagnet
• updating the number of epsilons for curie stochastic data
• adding openMP in operator loops

version:0.15

(24/02/16)

• bug correction in H contribution : only H anisotropy is computed whe use-anisotropy is true
• adding a network with the options
  • grid3D : a particle is linked with its 26 neighbor particles
  • cruxGrid3D : a particle is linked with only its 6 neighbor particles in the canonical directions

version:0.14

(24/01/25) modification of the options: particles -> stochastic-particles simulations -> stochastic-simulations adding uniaxial/planar anisotropy with options

• K
• Ku
• use-uniaxial-anisotropy
• use-planar-anisotropy adding generateMu runner

version 0.13

(24/01/17) Curie temperature : Esperance is modified & J computing depends on 1/r^3 adding a curie stochastic data adding a time curie stochastic data adding a math/random package

version:0.11

(23/05/24) Curie temperature tests

version:0.10

(23/05/23)

  Version of the software with implementation of :
          - stratonovich model
          - normalized stratonivich model
          - ito model

  The magnetic moment is computed on a network of points thanks to the Landau Lifchtitz equation integrated by a stochastic euler numerical method.

  The stochastic noise implemented are:
      - no noise for deterministic simulation
      - constant noise
      - inverse noise
  The evolution of energy for any time is computed
  The studied case tests are
      - 1 particle on a line network
      - 2 particles on a line network
      - 20 particles on a line network
      - 25 particles on a grid 5x5x1 network

 The magnetic field and energy has been tested by verified that the energy verifies dE/dmu = - H and by taylor expansion of order 2 of E

 The html doc is updated.