LinearSolve solvers

In this page, we list several recommended algorithms provided by LinearSolve.jl for solving steadystate and spectrum in hierarchical equations of motion approach.

Remember to import LinearSolve.jl

using LinearSolve

(click here to see the full algorithm list provided by LinearSolve.jl)

A generic GMRES implementation from Krylov (Default algorithm)

KrylovJL_GMRES(rtol=1e-12, atol=1e-14)

UMFPACKFactorization

This algorithm performs better when there is more structure to the sparsity pattern (depends on the complexity of your system and baths).

UMFPACKFactorization()

KLUFactorization

This algorithm performs better when there is less structure to the sparsity pattern (depends on the complexity of your system and baths).

KLUFactorization()

Pardiso

This algorithm is based on Intel openAPI Math Kernel Library (MKL) Pardiso

Note

Using this algorithm requires adding the package Pardiso.jl, i.e. using Pardiso

using Pardiso
using LinearSolve
MKLPardisoFactorize()
MKLPardisoIterate()