NormType#

determines what type of norm to compute with VecNorm(), VecNormBegin()/VecNormEnd() and MatNorm().

Synopsis#

typedef enum {
  NORM_1         = 0,
  NORM_2         = 1,
  NORM_FROBENIUS = 2,
  NORM_INFINITY  = 3,
  NORM_1_AND_2   = 4
} NormType;

NORM_1 - the one norm, \(\|v\| = \sum_i | v_i |\). \(\|A\| = \max_j \| A_{*,j} \|\), maximum column sum
NORM_2 - the two norm, \(\|v\| = \sqrt{\sum_i |v_i|^2}\) (vectors only)
NORM_FROBENIUS - \(\|A\| = \sqrt{\sum_{i,j} |A_{i,j}|^2}\), same as NORM_2 for vectors
NORM_INFINITY - \(\|v\| = \max_i |v_i|\). \(\|A\| = \max_i \| A_{i,*} \|_1\), maximum row sum
NORM_1_AND_2 - computes both the 1 and 2 norm of a vector. The values are stored in two adjacent PetscReal memory locations