Cutoff
ParallelTemperingMonteCarlo.MachineLearningPotential.Cutoff.angular_measure — Method
angular_measure(a,b,c,r2ij,r2ik)
angular_measure(a,b,c)Accepts three vectors a,b,c and can either accept or calculate the squared distances between them r2_ab,r2_bc, centred on vector a. Returns cos(θ) labelled as θ: the angular measure.
ParallelTemperingMonteCarlo.MachineLearningPotential.Cutoff.cutoff_function — Method
cutoff_function(r_scaled)
cutoff_function(r_ij,r_cut)
cutoff_function(dist_vec::T,r_cut) where {T<:Array}Implementation of the type 2 cutoff function. Either accepts scaled radius r_scaled or the interatomiic distance r_ij and the cutoff radius r_cut. Calculation is described in the RuNNer documentation, given as 1/2 (cos(πx) + 1) where x is (r_ij - r_i,c)/(rc - r_i,c). As an inner cutoff is not used by the potentials we are interested in, we have not included a method. A third method is included for creating a matrix or vector to match the distances provided.
ParallelTemperingMonteCarlo.MachineLearningPotential.Cutoff.dist2 — Method
dist2(a,b)Squared distance of two vectors a,b
ParallelTemperingMonteCarlo.MachineLearningPotential.Cutoff.find_distance2_mat — Method
find_distance2_mat(pos)Given a vector called pos comprised of (ideally) static vectors we return a lengthXlength symmetric matrix of the squared distance
ParallelTemperingMonteCarlo.MachineLearningPotential.Cutoff.thetacalc — Method
thetacalc(xy,xz,disxy,disxz)Calculates cosine theta of two vectors xy,xz with their sqared-distances disxy,disxz.