Leading eigenvalues of large sparse unsymmetric matrix

91 Views Asked by At

I have a matrix $R$ which is sparse and all eigenvalues are -ve with a zero eigenvalue. Size of R is more than $10^6 \times 10^6$. But I need to calculate only few large (by value not by magnitude) eigenvalues and eigenvectors including zero (say $1000$). Which one is the best algorithm with better convergence and less error. Can you please suggest few papers for the same.