Invalid data type. Input matrices must be double-precision. opts.CHOLB specified, but B is not upper triangular. Input matrices must be the same size for generalized problems. opts.cholB must be a logical scalar. opts.isreal must be a logical scalar. opts.issym must be a logical scalar. Number of basis vectors opts.p must be a positive integer <= n. opts.permB must be a permutation of 1:n. Convergence tolerance opts.tol must be a strictly positive real scalar. Sigma can be a scalar double or one of these options: ''LM'', ''SM'', ''LA'', ''SA'', ''BE'',''LR'', ''SR'', ''LI'', or ''SI''. First argument must be a double matrix or a function. Diagnostic level opts.disp must be 0, 1, or 2. Number of eigenvalues requested, k, must be a positive integer not greater than n = {0,number,integer}. Number of eigenvalues requested, k, must be a positive integer not greater than n = {0,number,integer}. Instead, it was {1}. Number of eigenvalues requested, k, must be a positive integer not greater than n = {0,number,integer}. Instead, it was a struct. Size of problem, ''n'', must be a positive integer. First argument must be a square matrix or a function. Start vector opts.v0 must be real for real problems. Options argument must be a structure. Maximum number of iterations opts.maxit must be a positive integer. Too many input arguments. Start vector opts.v0 must be a double column vector of length n. Ignoring opts.permB since B is not its sparse Cholesky factor. Invalid sigma value. For real symmetric problems, sigma must be ''LM'', ''SM'', ''LA'', ''SA'', or ''BE'', or a scalar double. Invalid sigma value. For nonsymmetric and complex problems, sigma must be ''LM'', ''SM'', ''LR'', ''SR'', ''LI'', or ''SI'', or a scalar double. For real symmetric problems, must have number of basis vectors opts.p > k. For nonsymmetric and complex problems, must have number of basis vectors opts.p > k+1. Complex eigenvalue shift sigma on a Hermitian problem (all real eigenvalues). For real symmetric problems, sigma value ''LR'' (Largest Real) is now ''LA'' (Largest Algebraic). For real symmetric problems, sigma value ''SR'' (Smallest Real) is now ''SA'' (Smallest Algebraic). Sigma value ''BE'' only available for real symmetric problems. Computing ''LM'' eigenvalues instead. The shifted operator is singular. The shift is an eigenvalue. Consider using a different value for sigma. Function handle AFUN is complex. Set opts.isreal = false. First input matrix is singular. The shift 0 is an eigenvalue. Consider using a nonzero numeric value for sigma. Singular second input matrix is only supported when sigma is ''SM'' or a scalar double . First input matrix is close to singular or badly scaled. RCOND = {0}. Results may be inaccurate. Second input matrix B is close to singular or badly scaled. RCOND = {0}. Results may be inaccurate. The shifted operator is close to singular or badly scaled. RCOND = {0}. Consider using a different value for sigma. Error with ARPACK routine {0}: info = {1,number,integer} ARPACK call resulted in an internal error: Invalid ipntr. ARPACK call resulted in an internal error: Unknown Ido. Only {0,number,integer} of the {1,number,integer} requested eigenvalues converged. Iteration {0,number,integer}: a few Ritz values of the {1,number,integer}-by-{2,number,integer} matrix: Compute A*X: {0}\n Solve (A-SIGMA*I)*X=Y for X: {0}\n Solve (A-SIGMA*B)*X=B*Y for X: {0}\n Pre-processing, including lu(A-sigma*I): {0}\n Pre-processing, including lu(A-sigma*B): {0}\n Pre-processing: {0}\n ***********CPU Timing Results in seconds*********** Post-processing: {0}\n Total: {0}\n