setmatrix_bcyclic: CPU => multi-GPU

Functions
void	magma_csetmatrix_1D_col_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const magmaFloatComplex *hA, magma_int_t lda, magmaFloatComplex_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.
void	magma_csetmatrix_1D_row_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const magmaFloatComplex *hA, magma_int_t lda, magmaFloatComplex_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.
void	magma_dsetmatrix_1D_col_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const double *hA, magma_int_t lda, magmaDouble_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.
void	magma_dsetmatrix_1D_row_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const double *hA, magma_int_t lda, magmaDouble_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.
void	magma_ssetmatrix_1D_col_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const float *hA, magma_int_t lda, magmaFloat_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.
void	magma_ssetmatrix_1D_row_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const float *hA, magma_int_t lda, magmaFloat_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.
void	magma_zsetmatrix_1D_col_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const magmaDoubleComplex *hA, magma_int_t lda, magmaDoubleComplex_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.
void	magma_zsetmatrix_1D_row_bcyclic (magma_int_t ngpu, magma_int_t m, magma_int_t n, magma_int_t nb, const magmaDoubleComplex *hA, magma_int_t lda, magmaDoubleComplex_ptr *dA, magma_int_t ldda, magma_queue_t queues[])
	Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.

Detailed Description

Function Documentation

magma_csetmatrix_1D_col_bcyclic()

void magma_csetmatrix_1D_col_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const magmaFloatComplex *	hA,
		magma_int_t	lda,
		magmaFloatComplex_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,nlocal), where nlocal is the columns assigned to each GPU.
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= m.
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_csetmatrix_1D_row_bcyclic()

void magma_csetmatrix_1D_row_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const magmaFloatComplex *	hA,
		magma_int_t	lda,
		magmaFloatComplex_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,n).
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= (1 + m/(nb*ngpu))*nb
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_dsetmatrix_1D_col_bcyclic()

void magma_dsetmatrix_1D_col_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const double *	hA,
		magma_int_t	lda,
		magmaDouble_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,nlocal), where nlocal is the columns assigned to each GPU.
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= m.
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_dsetmatrix_1D_row_bcyclic()

void magma_dsetmatrix_1D_row_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const double *	hA,
		magma_int_t	lda,
		magmaDouble_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,n).
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= (1 + m/(nb*ngpu))*nb
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_ssetmatrix_1D_col_bcyclic()

void magma_ssetmatrix_1D_col_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const float *	hA,
		magma_int_t	lda,
		magmaFloat_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,nlocal), where nlocal is the columns assigned to each GPU.
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= m.
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_ssetmatrix_1D_row_bcyclic()

void magma_ssetmatrix_1D_row_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const float *	hA,
		magma_int_t	lda,
		magmaFloat_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,n).
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= (1 + m/(nb*ngpu))*nb
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_zsetmatrix_1D_col_bcyclic()

void magma_zsetmatrix_1D_col_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const magmaDoubleComplex *	hA,
		magma_int_t	lda,
		magmaDoubleComplex_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D column block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,nlocal), where nlocal is the columns assigned to each GPU.
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= m.
[in]	queues	Array of dimension (ngpu), with one queue per GPU.

magma_zsetmatrix_1D_row_bcyclic()

void magma_zsetmatrix_1D_row_bcyclic	(	magma_int_t	ngpu,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nb,
		const magmaDoubleComplex *	hA,
		magma_int_t	lda,
		magmaDoubleComplex_ptr *	dA,
		magma_int_t	ldda,
		magma_queue_t	queues[] )

Copy matrix hA on CPU host to dA, which is distributed 1D row block cyclic over multiple GPUs.

Parameters

[in]	ngpu	Number of GPUs over which dAT is distributed.
[in]	m	Number of rows of matrix hA. m >= 0.
[in]	n	Number of columns of matrix hA. n >= 0.
[in]	nb	Block size. nb > 0.
[in]	hA	The m-by-n matrix A on the CPU, of dimension (lda,n).
[in]	lda	Leading dimension of matrix hA. lda >= m.
[out]	dA	Array of ngpu pointers, one per GPU, that store the disributed m-by-n matrix A on the GPUs, each of dimension (ldda,n).
[in]	ldda	Leading dimension of each matrix dAT on each GPU. ldda >= (1 + m/(nb*ngpu))*nb
[in]	queues	Array of dimension (ngpu), with one queue per GPU.