Notes: This is rarely used directly Many br
Many br
Notes: Use preconditioned conjugate gradient to compute Many bran approximate minimizer of the quadratic function Many br
q(s) = g^T * s + .5 * s^T * H * s Many br
subject to the Euclidean norm trust region constraint Many br
|| D * s || <= delta, Many br
where Many br
delta is the trust region radius, Many brg is the gradient vector, and Many brH is Hessian matrix, Many brD is a scaling matrix. Many br
KSPConvergedReason may be Many br
KSP_CONVERGED_CG_NEG_CURVE if convergence is reached along a negative curvature direction,
KSP_CONVERGED_CG_CONSTRAINED if convergence is reached along a constrained step,
other KSP converged/diverged reasons
Level:developer
Location:src/ksp/ksp/impls/qcg/qcg.c
Index of all KSP routines
Table of Contents for all manual pages
Index of all manual pages