PetscDAEnsembleTFactor#

Compute and store factorization of T matrix

Synopsis#

Collective

Input Parameters#

  • da - the PetscDA context

  • S - normalized innovation matrix (obs_size x m)

Notes#

This function computes \(T = (1/\rho) I + S^T * S\) (where \(\rho\) is the inflation factor set via PetscDAEnsembleSetInflation()) and stores its symmetric eigendecomposition, i.e. eigenvectors \(V\) and eigenvalues \(D\) such that \(T = V * D * V^T\).

The implementation uses matrix reuse (MAT_REUSE_MATRIX) to minimize memory allocation overhead when the ensemble size remains constant across analysis cycles.

See Also#

PetscDA: Data Assimilation, PetscDA, PETSCDALETKF, PetscDAEnsembleApplyTInverse(), PetscDAEnsembleApplySqrtTInverse()

Level#

advanced

Location#

src/ml/da/impls/ensemble/daensemble.c

Implementations#

PetscDAEnsembleTFactor_Eigen() in src/ml/da/impls/ensemble/daensemble.c


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