`scipy.linalg._flapack:sstemr`

Summary

Compute selected eigenvalues and optionally eigenvectors of a symmetric tridiagonal matrix by the MRRR algorithm (LAPACK sstemr).

d : ndarray: Diagonal elements of the symmetric tridiagonal matrix, length n.
e : ndarray: Off-diagonal elements in the leading n - 1 entries, length n. The final entry need not be set; LAPACK uses it as workspace.
range : int: Which eigenvalues to compute: 0 for all, 1 for those in (vl, vu], 2 for those with indices il through iu.
vl : float: Lower bound of the half-open interval; used only when range is 1.
vu : float: Upper bound of the half-open interval; used only when range is 1.
il : int: Index of the smallest eigenvalue to return, 1-based; used only when range is 2.
iu : int: Index of the largest eigenvalue to return, 1-based; used only when range is 2.
compute_v : int, optional: If nonzero, eigenvectors are computed. Default is 1.
lwork : int, optional: Size of the workspace. Default is 18 * n when compute_v is nonzero and 12 * n otherwise. Use stemr_lwork for the optimal value.
liwork : int, optional: Size of the integer workspace. Default is 10 * n when compute_v is nonzero and 8 * n otherwise. Use stemr_lwork for the optimal value.
overwrite_d : int, optional: If nonzero, d may be overwritten in place. Default is 0.
overwrite_e : int, optional: If nonzero, e may be overwritten in place. Default is 0.

m : int: Number of eigenvalues found; only w[:m] and z[:, :m] are meaningful.
w : ndarray: The computed eigenvalues in ascending order, length n.
z : ndarray: Orthonormal eigenvectors as columns, shape (n, n). Not referenced when compute_v is 0.
info : int: 0 on success; if negative, the -info-th argument had an illegal value; if positive, the algorithm failed to converge.