Fast and differentiable implementations of matrix exponentials, Krylov exponential matrix-vector multiplications ("expmv"), KIOPS, ExpoKit functions, and more. All your exponential needs in SciML form.
Currently, krylov_phiv_error_estimate.jl only implements the Lanczos algorithm for Hermitian matrix, which gives a tridiagonal H to use with the expT! functiopns's stegr call
Checklist
Additional context
Currently, krylov_phiv_error_estimate.jl only implements the Lanczos algorithm for Hermitian matrix, which gives a tridiagonal H to use with the expT! functiopns's stegr call
https://github.com/SciML/ExponentialUtilities.jl/blob/18e5161ad28e49e2b92c5f4879495dbd55a0b52e/src/krylov_phiv_error_estimate.jl#L29
This PR passes on the
ishermitian
argument and errors if false.Fixes #160
-- whitespace changes is with JuliaFormatter's sciml style.