TaoGetConvergenceHistory#

Gets the arrays used that hold the convergence history.

Synopsis#

#include "petsctao.h" 
PetscErrorCode TaoGetConvergenceHistory(Tao tao, PetscReal **obj, PetscReal **resid, PetscReal **cnorm, PetscInt **lits, PetscInt *nhist)

Collective

Input Parameter#

tao - the Tao context

Output Parameters#

obj - array used to hold objective value history
resid - array used to hold residual history
cnorm - array used to hold constraint violation history
lits - integer array used to hold linear solver iteration count
nhist - size of obj, resid, cnorm, and lits

Notes#

This routine must be preceded by calls to TaoSetConvergenceHistory() and TaoSolve(), otherwise it returns useless information.

This routine is useful, e.g., when running a code for purposes of accurate performance monitoring, when no I/O should be done during the section of code that is being timed.

Fortran Notes#

The calling sequence is

   call TaoGetConvergenceHistory(Tao tao, PetscInt nhist, PetscErrorCode ierr)

In other words this gets the current number of entries in the history. Access the history through the array you passed to TaoSetConvergenceHistory()