TSAlphaSetParams#

sets the algorithmic parameters for TSALPHA

Synopsis#

#include "petscts.h"   
PetscErrorCode TSAlphaSetParams(TS ts, PetscReal alpha_m, PetscReal alpha_f, PetscReal gamma)

Logically Collective

Input Parameters#

  • ts - timestepping context

  • alpha_m - algorithmic parameter

  • alpha_f - algorithmic parameter

  • gamma - algorithmic parameter

Options Database Keys#

  • -ts_alpha_alpha_m <alpha_m> - set alpha_m

  • -ts_alpha_alpha_f <alpha_f> - set alpha_f

  • -ts_alpha_gamma - set gamma

Note#

Second-order accuracy can be obtained so long as: \(\gamma = 0.5 + \alpha_m - \alpha_f\)

Unconditional stability requires: \(\alpha_m >= \alpha_f >= 0.5\)

Backward Euler method is recovered with: \(\alpha_m = \alpha_f = \gamma = 1\)

Use of this function is normally only required to hack TSALPHA to use a modified integration scheme. Users should call TSAlphaSetRadius() to set the desired spectral radius of the methods (i.e. high-frequency damping) in order so select optimal values for these parameters.

See Also#

TS: Scalable ODE and DAE Solvers, TS, TSALPHA, TSAlphaSetRadius(), TSAlphaGetParams()

Level#

advanced

Location#

src/ts/impls/implicit/alpha/alpha1.c

Implementations#

TSAlphaSetParams_Alpha() in src/ts/impls/implicit/alpha/alpha1.c


Index of all TS routines
Table of Contents for all manual pages
Index of all manual pages