In this chapter we record the software for the non-Hermitian eigenvalue problem (NHEP), A x = lambda x, where the square matrix A \= A^*. x \= 0 is called a right eigenvector. A vector y \= 0 satisfying y^* A = lambda y^* is called a left eigenvector of A.

Section	Package Name	Language	Comments
7.3	LAPACK	Fortran 77, C++ wrapper	direct methods
7.4	EB12	Fortran 77	multivector iteration, eigenpairs - leftmost, rightmost, largest modulus eigvalue
7.4	SRRIT	Fortran 77	multivector iteration, calculates dominant invariant subspace
7.4	LOPSI	Fortran 77	multivector iteration, finds dominant eigenpairs
7.6	ARPACK	Fortran 77, C++ wrapper	implicitly restarted Arnoldi method
7.5	EB13	Fortran 77	Arnoldi method, eigenvalues - rightmost, leftmost, max imaginary part
7.8	QMRPACK	Fortran 77	quasi-minimal residual (QMR) based on look-ahead Lanczos
7.8, 7.9	ABLEPACK	MATLAB	adaptive block Lanczos
7.9	LASO2	Fortran IV	block Lanczos
7.9	BLZPACK	Fortran 77	block Lanczos
7.11	QMRPACK	Fortran 77	complex symmetric Lanczos
7.12	JDQR	MATLAB	Jacobi-Davidson