matrix_norm

Compute a norm of a matrix. Currently Frobenius or L1 norms are supported

The order parameter can be either NormOrder::L1` or NormOrder::FROB. NormOrder::NONE may also be used as an alias for the Frobenius norm.

template<typename Op>
__MATX_INLINE__ auto matx::matrix_norm(const Op &op, NormOrder order = NormOrder::NONE)

Compute a matrix norm.

Computes various types of matrix and matrix norms based on the order

Template Parameters:

Op – Type of input values to evaluate

Parameters:

• op – Input values to evaluate

• order – Order of norm

Returns:

norm operator

template<typename Op, int D>
__MATX_INLINE__ auto matx::matrix_norm(const Op &op, const int (&dims)[D], NormOrder order = NormOrder::NONE)

Compute a matrix norm.

Computes various types of matrix norms based on the order

Template Parameters:

Op – Type of input values to evaluate

Parameters:

• op – Input values to evaluate

• dims – Dimensions to perform norm over

• order – Order of norm

Returns:

norm operator

Examples

(this->out_m = matrix_norm(this->in_m, NormOrder::L1)).run(this->exec);

(this->out_m = matrix_norm(this->in_m, NormOrder::FROB)).run(this->exec);