Custom solvers

In this example, we show how to define custom solvers. Our system will again be silicon, because we are not very imaginative

In [1]:
using DFTK, LinearAlgebra

a = 10.26
lattice = a / 2 * [[0 1 1.];
                   [1 0 1.];
                   [1 1 0.]]
Si = ElementPsp(:Si, psp=load_psp("hgh/lda/Si-q4"))
atoms = [Si => [ones(3)/8, -ones(3)/8]]

# We take very (very) crude parameters
model = model_LDA(lattice, atoms)
kgrid = [1, 1, 1]
Ecut = 5
basis = PlaneWaveBasis(model, Ecut; kgrid=kgrid);

We define our custom fix-point solver: simply a damped fixed-point

In [2]:
function my_fp_solver(f, x0, max_iter; tol)
    mixing_factor = .7
    x = x0
    fx = f(x)
    for n = 1:max_iter
        inc = fx - x
        if norm(inc) < tol
            break
        end
        x = x + mixing_factor * inc
        fx = f(x)
    end
    (fixpoint=x, converged=norm(fx-x) < tol)
end;

Our eigenvalue solver just forms the dense matrix and diagonalizes it explicitly (this only works for very small systems)

In [3]:
function my_eig_solver(A, X0; maxiter, tol, kwargs...)
    n = size(X0, 2)
    A = Array(A)
    E = eigen(A)
    λ = E.values[1:n]
    X = E.vectors[:, 1:n]
    (λ=λ, X=X, residual_norms=[], iterations=0, converged=true, n_matvec=0)
end;

Finally we also define our custom mixing scheme. It will be a mixture of simple mixing (for the first 2 steps) and than default to Kerker mixing. In the mixing interface δF is (ρ_\text{out} - ρ_\text{in}), i.e. the difference in density between two subsequent SCF steps and the mix function returns δρ, which is added to ρ_\text{in} to yield ρ_\text{next}, the density for the next SCF step.

In [4]:
struct MyMixing
    α  # Damping parameter
end
MyMixing() = MyMixing(0.7)

function DFTK.mix(mixing::MyMixing, basis, δF::RealFourierArray, δF_spin=nothing; n_iter, kwargs...)
    if n_iter <= 2
        # Just do simple mixing on total density and spin density (if it exists)
        (mixing.α * δF, isnothing(δF_spin) ? nothing : mixing.α * δF_spin)
    else
        # Use the KerkerMixing from DFTK
        DFTK.mix(KerkerMixing(α=mixing.α), basis, δF, δF_spin; kwargs...)
    end
end

That's it! Now we just run the SCF with these solvers

In [5]:
scfres = self_consistent_field(basis;
                               tol=1e-8,
                               solver=my_fp_solver,
                               eigensolver=my_eig_solver,
                               mixing=MyMixing());
n     Energy            Eₙ-Eₙ₋₁     ρout-ρin   Diag
---   ---------------   ---------   --------   ----
  1   -7.234302105583         NaN   3.17e-01    0.0 
  2   -7.246539783547   -1.22e-02   1.44e-01    0.0 
  3   -7.248795495550   -2.26e-03   6.55e-02    0.0 
  4   -7.249077010080   -2.82e-04   3.61e-02    0.0 
  5   -7.249170222206   -9.32e-05   2.00e-02    0.0 
  6   -7.249201046259   -3.08e-05   1.12e-02    0.0 
  7   -7.249211378481   -1.03e-05   6.34e-03    0.0 
  8   -7.249214920041   -3.54e-06   3.64e-03    0.0 
  9   -7.249216167364   -1.25e-06   2.12e-03    0.0 
 10   -7.249216619477   -4.52e-07   1.25e-03    0.0 
 11   -7.249216787981   -1.69e-07   7.50e-04    0.0 
 12   -7.249216852389   -6.44e-08   4.56e-04    0.0 
 13   -7.249216877549   -2.52e-08   2.81e-04    0.0 
 14   -7.249216887557   -1.00e-08   1.75e-04    0.0 
 15   -7.249216891595   -4.04e-09   1.10e-04    0.0 

Note that the default convergence criterion is on the difference of energy from one step to the other; when this gets below tol, the "driver" self_consistent_field artificially makes the fixpoint solver think it's converged by forcing f(x) = x. You can customize this with the is_converged keyword argument to self_consistent_field.