 |
(git:08825b9)
|
|
(git:08825b9)
|
Public Member Functions | |
| subroutine | quadratic_solve_real (n, shift, weights, nstates, iterative, max_subspace, max_iter, tolerance, apply, solve, result, status) |
| Solve sum_j w_j^2 A_j S^-1 A_j c = lambda S c without squaring sparse operators. | |
| subroutine | quadratic_solve_complex (n, shift, weights, nstates, iterative, max_subspace, max_iter, tolerance, apply, solve, result, status, adjoint, unitary) |
| Solve sum_j w_j^2 (A_j-z_j S)^dagger S^-1 (A_j-z_j S) c = lambda S c. | |
Definition at line 32 of file quadratic_pseudospectrum.F.
| subroutine quadratic_pseudospectrum::quadratic_solve::quadratic_solve_real | ( | integer, intent(in) | n, |
| real(dp), dimension(:), intent(in) | shift, | ||
| real(dp), dimension(:), intent(in) | weights, | ||
| integer, intent(in) | nstates, | ||
| logical, intent(in) | iterative, | ||
| integer, intent(in) | max_subspace, | ||
| integer, intent(in) | max_iter, | ||
| real(dp), intent(in) | tolerance, | ||
| procedure(operator_apply) | apply, | ||
| procedure(metric_solve) | solve, | ||
| type(quadratic_result_type), intent(out) | result, | ||
| integer, intent(out) | status ) |
Solve sum_j w_j^2 A_j S^-1 A_j c = lambda S c without squaring sparse operators.
| n | AO/spinor dimension |
| shift | Query values for the observables |
| weights | Nonnegative observable scales |
| nstates | Number of lowest eigenpairs |
| iterative | Whether to use a bounded restarted block Ritz space |
| max_subspace | Maximum trial-space size |
| max_iter | Maximum Ritz iterations |
| tolerance | Relative S^-1 eigenpair residual threshold |
| apply | Apply covariant observables or metric (operator zero) |
| solve | Apply inverse metric using already validated factors |
| RESULT | Normalized eigenstates and separated physical residuals |
| status | 0 success, 1 unconverged, -1 invalid input, -4 numerical failure |
Definition at line 321 of file quadratic_pseudospectrum.F.
| subroutine quadratic_pseudospectrum::quadratic_solve::quadratic_solve_complex | ( | integer, intent(in) | n, |
| complex(dp), dimension(:), intent(in) | shift, | ||
| real(dp), dimension(:), intent(in) | weights, | ||
| integer, intent(in) | nstates, | ||
| logical, intent(in) | iterative, | ||
| integer, intent(in) | max_subspace, | ||
| integer, intent(in) | max_iter, | ||
| real(dp), intent(in) | tolerance, | ||
| procedure(operator_apply) | apply, | ||
| procedure(metric_solve) | solve, | ||
| type(quadratic_result_type), intent(out) | result, | ||
| integer, intent(out) | status, | ||
| procedure(operator_apply) | adjoint, | ||
| logical, dimension(:), intent(in), optional | unitary ) |
Solve sum_j w_j^2 (A_j-z_j S)^dagger S^-1 (A_j-z_j S) c = lambda S c.
| n | AO/spinor dimension |
| shift | Complex queries for arbitrary, possibly nonnormal operators |
| weights | Nonnegative residual scales |
| nstates | Number of lowest eigenpairs |
| iterative | Whether to use a bounded restarted block Ritz space |
| max_subspace | Maximum trial-space size |
| max_iter | Maximum Ritz iterations |
| tolerance | Relative S^-1 eigenpair residual threshold |
| apply | Apply covariant operators or the positive metric (operator zero) |
| solve | Solve with the already validated positive metric |
| RESULT | Eigenpairs, residuals and complex expectations; means holds their real parts |
| status | 0 success, 1 unconverged, -1 input, -4 numerical failure |
| adjoint | Apply A_j^dagger, not the coefficient-space S-adjoint; positive operator indices |
| unitary | True for compressions of continuum unitary operators; retain full norms and projection leakage |
Definition at line 356 of file quadratic_pseudospectrum.F.
1.15.0