\(C = \alpha A A^H + \beta C\) where \(C\) is Hermitian More...

Functions
template<typename TA , typename TC >
void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, real_type< TA, TC > alpha, TA const A, int64_t lda, real_type< TA, TC > beta, TC C, int64_t ldc)
	Hermitian rank-k update:

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< float > const &alpha, std::vector< float * > const &Aarray, std::vector< int64_t > const &lda, std::vector< float > const &beta, std::vector< float * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info)
	CPU, variable-size batched, float version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< double > const &alpha, std::vector< double * > const &Aarray, std::vector< int64_t > const &lda, std::vector< double > const &beta, std::vector< double * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info)
	CPU, variable-size batched, double version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< float > const &alpha, std::vector< std::complex< float > * > const &Aarray, std::vector< int64_t > const &lda, std::vector< float > const &beta, std::vector< std::complex< float > * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info)
	CPU, variable-size batched, complex<float> version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< double > const &alpha, std::vector< std::complex< double > * > const &Aarray, std::vector< int64_t > const &lda, std::vector< double > const &beta, std::vector< std::complex< double > * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info)
	CPU, variable-size batched, complex<double> version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< float > const &alpha, std::vector< float * > const &Aarray, std::vector< int64_t > const &lda, std::vector< float > const &beta, std::vector< float * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info, blas::Queue &queue)
	GPU device, variable-size batched, float version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< double > const &alpha, std::vector< double * > const &Aarray, std::vector< int64_t > const &lda, std::vector< double > const &beta, std::vector< double * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info, blas::Queue &queue)
	GPU device, variable-size batched, double version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< float > const &alpha, std::vector< std::complex< float > * > const &Aarray, std::vector< int64_t > const &lda, std::vector< float > const &beta, std::vector< std::complex< float > * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info, blas::Queue &queue)
	GPU device, variable-size batched, complex<float> version.

void	blas::batch::herk (blas::Layout layout, std::vector< blas::Uplo > const &uplo, std::vector< blas::Op > const &trans, std::vector< int64_t > const &n, std::vector< int64_t > const &k, std::vector< double > const &alpha, std::vector< std::complex< double > * > const &Aarray, std::vector< int64_t > const &lda, std::vector< double > const &beta, std::vector< std::complex< double > * > const &Carray, std::vector< int64_t > const &ldc, size_t batch_size, std::vector< int64_t > &info, blas::Queue &queue)
	GPU device, variable-size batched, complex<double> version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, float alpha, float const A, int64_t lda, float beta, float C, int64_t ldc, blas::Queue &queue)
	GPU device, float version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, double alpha, double const A, int64_t lda, double beta, double C, int64_t ldc, blas::Queue &queue)
	GPU device, double version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, float alpha, std::complex< float > const A, int64_t lda, float beta, std::complex< float > C, int64_t ldc, blas::Queue &queue)
	GPU device, complex<float> version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, double alpha, std::complex< double > const A, int64_t lda, double beta, std::complex< double > C, int64_t ldc, blas::Queue &queue)
	GPU device, complex<double> version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, float alpha, float const A, int64_t lda, float beta, float C, int64_t ldc)
	CPU, float version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, double alpha, double const A, int64_t lda, double beta, double C, int64_t ldc)
	CPU, double version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, float alpha, std::complex< float > const A, int64_t lda, float beta, std::complex< float > C, int64_t ldc)
	CPU, complex<float> version.

void	blas::herk (blas::Layout layout, blas::Uplo uplo, blas::Op trans, int64_t n, int64_t k, double alpha, std::complex< double > const A, int64_t lda, double beta, std::complex< double > C, int64_t ldc)
	CPU, complex<double> version.

Detailed Description

\(C = \alpha A A^H + \beta C\) where \(C\) is Hermitian

Function Documentation

◆ herk()

template<typename TA , typename TC >

void blas::herk	(	blas::Layout	layout,
		blas::Uplo	uplo,
		blas::Op	trans,
		int64_t	n,
		int64_t	k,
		real_type< TA, TC >	alpha,
		TA const *	A,
		int64_t	lda,
		real_type< TA, TC >	beta,
		TC *	C,
		int64_t	ldc
	)

Hermitian rank-k update:

\[ C = \alpha A A^H + \beta C, \]

or

\[ C = \alpha A^H A + \beta C, \]

where alpha and beta are real scalars, C is an n-by-n Hermitian matrix, and A is an n-by-k or k-by-n matrix.

Generic implementation for arbitrary data types.

Parameters

[in]	layout	Matrix storage, Layout::ColMajor or Layout::RowMajor.
[in]	uplo	What part of the matrix C is referenced, the opposite triangle being assumed from symmetry: Uplo::Lower: only the lower triangular part of C is referenced. Uplo::Upper: only the upper triangular part of C is referenced.
[in]	trans	The operation to be performed: Op::NoTrans: \(C = \alpha A A^H + \beta C\). Op::ConjTrans: \(C = \alpha A^H A + \beta C\). In the real case, Op::Trans is interpreted as Op::ConjTrans. In the complex case, Op::Trans is illegal (see syrk: Symmetric rank k update instead).
[in]	n	Number of rows and columns of the matrix C. n >= 0.
[in]	k	If trans = NoTrans: number of columns of the matrix A. k >= 0. Otherwise: number of rows of the matrix A. k >= 0.
[in]	alpha	Scalar alpha. If alpha is zero, A is not accessed.
[in]	A	If trans = NoTrans: the n-by-k matrix A, stored in an lda-by-k array [RowMajor: n-by-lda]. Otherwise: the k-by-n matrix A, stored in an lda-by-n array [RowMajor: k-by-lda].
[in]	lda	Leading dimension of A. If trans = NoTrans: lda >= max(1, n) [RowMajor: lda >= max(1, k)], If Otherwise: lda >= max(1, k) [RowMajor: lda >= max(1, n)].
[in]	beta	Scalar beta. If beta is zero, C need not be set on input.
[in]	C	The n-by-n Hermitian matrix C, stored in an lda-by-n array [RowMajor: n-by-lda].
[in]	ldc	Leading dimension of C. ldc >= max(1, n).

Functions

Detailed Description

Function Documentation

◆ herk()