Is there already a way to get eigenvalues and eigenvectors with BigInt/BigFloat accuracy? I know about the GenericLinearAlgebra.jl implementation for eigvals, but I need the high accuracy in the eigenvectors. Anybody knows about an implementation already out there? If not, what could be the easiest way to get a working code (I don’t care too much about speed as long as its faster than symbolic math tools like SymPy or Mathematica) ?
Do you need all of the eigenvalues and eigenvectors, or just particular ones? Is your matrix arbitrary, or special in some way (real-symmetric, tridiagonal, etcetera)?
I have arbitrary real matrices. Actually I need to solve a generalized problem, but in my case the matrix is invertable to make it a standard problem. I’d be happy about both a targeted and a dense solver, no real preference.
Take a look at GitHub - RalphAS/GenericSchur.jl: Julia package for Schur decomposition of matrices with generic element types. To get the eigenvectors, you’d have to convert the input matrix to complex before computing the Schur factorization.
1 Like
Thanks. That’s basically what i needed.