Functions
magma_int_t	magma_scustomspmv (magma_int_t m, magma_int_t n, float alpha, float beta, float x, float y, magma_queue_t queue)
	This is an interface to any custom sparse matrix vector product.

magma_int_t	magma_s_spmv (float alpha, magma_s_matrix A, magma_s_matrix x, float beta, magma_s_matrix y, magma_queue_t queue)
	For a given input matrix A and vectors x, y and scalars alpha, beta the wrapper determines the suitable SpMV computing y = alpha * A * x + beta * y.

magma_int_t	magma_s_spmv_shift (float alpha, magma_s_matrix A, float lambda, magma_s_matrix x, float beta, magma_int_t offset, magma_int_t blocksize, magma_index_t *add_rows, magma_s_matrix y, magma_queue_t queue)
	For a given input matrix A and vectors x, y and scalars alpha, beta the wrapper determines the suitable SpMV computing y = alpha * ( A - lambda I ) * x + beta * y.

magma_int_t	magma_s_spmm (float alpha, magma_s_matrix A, magma_s_matrix B, magma_s_matrix *C, magma_queue_t queue)
	For a given input matrix A and B and scalar alpha, the wrapper determines the suitable SpMV computing C = alpha * A * B.

magma_int_t	magma_scuspaxpy (float alpha, magma_s_matrix A, float beta, magma_s_matrix B, magma_s_matrix *AB, magma_queue_t queue)
	This is an interface to the cuSPARSE routine csrgeam computing the sum of two sparse matrices stored in csr format:

magma_int_t	magma_scuspmm (magma_s_matrix A, magma_s_matrix B, magma_s_matrix *AB, magma_queue_t queue)
	This is an interface to the cuSPARSE routine csrmm computing the product of two sparse matrices stored in csr format.

magma_int_t	magma_sge3pt (magma_int_t m, magma_int_t n, float alpha, float beta, magmaFloat_ptr dx, magmaFloat_ptr dy, magma_queue_t queue)
	This routine is a 3-pt-stencil operator derived from a FD-scheme in 2D with Dirichlet boundary.

magma_int_t	magma_sgeaxpy (float alpha, magma_s_matrix X, float beta, magma_s_matrix *Y, magma_queue_t queue)
	This routine computes Y = alpha * X + beta * Y on the GPU.

magma_int_t	magma_sgecsr5mv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t p, float alpha, magma_int_t sigma, magma_int_t bit_y_offset, magma_int_t bit_scansum_offset, magma_int_t num_packet, magmaUIndex_ptr dtile_ptr, magmaUIndex_ptr dtile_desc, magmaIndex_ptr dtile_desc_offset_ptr, magmaIndex_ptr dtile_desc_offset, magmaFloat_ptr dcalibrator, magma_int_t tail_tile_start, magmaFloat_ptr dval, magmaIndex_ptr drowptr, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * A * x + beta * y on the GPU.

magma_int_t	magma_sgecsrmv (magma_trans_t transA, magma_int_t m, magma_int_t n, float alpha, magmaFloat_ptr dval, magmaIndex_ptr drowptr, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * A * x + beta * y on the GPU.

magma_int_t	magma_sgecsrmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, float alpha, float lambda, magmaFloat_ptr dval, magmaIndex_ptr drowptr, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magma_int_t offset, magma_int_t blocksize, magma_index_t *addrows, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * ( A -lambda I ) * x + beta * y on the GPU.

magma_int_t	magma_sgecsrreimsplit (magma_s_matrix A, magma_s_matrix ReA, magma_s_matrix ImA, magma_queue_t queue)
	This routine takes an input matrix A in CSR format and located on the GPU and splits it into two matrixes ReA and ImA containing the real and the imaginary contributions of A.

magma_int_t	magma_sgedensereimsplit (magma_s_matrix A, magma_s_matrix ReA, magma_s_matrix ImA, magma_queue_t queue)
	This routine takes an input matrix A in DENSE format and located on the GPU and splits it into two matrixes ReA and ImA containing the real and the imaginary contributions of A.

magma_int_t	magma_sgeellmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * A * x + beta * y on the GPU.

magma_int_t	magma_sgeellmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, float alpha, float lambda, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magma_int_t offset, magma_int_t blocksize, magmaIndex_ptr addrows, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha ( A - lambda I ) x + beta * y on the GPU.

magma_int_t	magma_sgeellrtmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaIndex_ptr drowlength, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_int_t alignment, magma_int_t blocksize, magma_queue_t queue)
	This routine computes y = alpha * A * x + beta * y on the GPU.

magma_int_t	magma_sgeelltmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, float alpha, float lambda, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magma_int_t offset, magma_int_t blocksize, magmaIndex_ptr addrows, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha ( A - lambda I ) x + beta * y on the GPU.

magma_int_t	magma_smdotc (magma_int_t n, magma_int_t k, magmaFloat_ptr v, magmaFloat_ptr r, magmaFloat_ptr d1, magmaFloat_ptr d2, magmaFloat_ptr skp, magma_queue_t queue)
	Computes the scalar product of a set of vectors v_i such that.

magma_int_t	magma_sgesellpmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t blocksize, magma_int_t slices, magma_int_t alignment, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaIndex_ptr drowptr, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * A^t * x + beta * y on the GPU.

magma_int_t	magma_sgesellcmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t blocksize, magma_int_t slices, magma_int_t alignment, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaIndex_ptr drowptr, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes y = alpha * A^t * x + beta * y on the GPU.

magma_int_t	magma_smdotc_shfl (magma_int_t n, magma_int_t k, magmaFloat_ptr v, magmaFloat_ptr r, magmaFloat_ptr d1, magmaFloat_ptr d2, magmaFloat_ptr skp, magma_queue_t queue)
	Computes the scalar product of a set of vectors v_i such that.

magma_int_t	magma_smdotc4 (magma_int_t n, magmaFloat_ptr v0, magmaFloat_ptr w0, magmaFloat_ptr v1, magmaFloat_ptr w1, magmaFloat_ptr v2, magmaFloat_ptr w2, magmaFloat_ptr v3, magmaFloat_ptr w3, magmaFloat_ptr d1, magmaFloat_ptr d2, magmaFloat_ptr skp, magma_queue_t queue)
	Computes the scalar product of a set of 4 vectors such that.

magma_int_t	magma_smgecsrmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, float alpha, magmaFloat_ptr dval, magmaIndex_ptr drowptr, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

magma_int_t	magma_smgeellmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, magma_int_t nnz_per_row, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

magma_int_t	magma_smgeelltmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, magma_int_t nnz_per_row, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

magma_int_t	magma_smgesellpmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, magma_int_t blocksize, magma_int_t slices, magma_int_t alignment, float alpha, magmaFloat_ptr dval, magmaIndex_ptr dcolind, magmaIndex_ptr drowptr, magmaFloat_ptr dx, float beta, magmaFloat_ptr dy, magma_queue_t queue)
	This routine computes Y = alpha * A^t * X + beta * Y on the GPU.

Detailed Description

Function Documentation

◆ magma_scustomspmv()

magma_int_t magma_scustomspmv	(	magma_int_t	m,
		magma_int_t	n,
		float	alpha,
		float	beta,
		float *	x,
		float *	y,
		magma_queue_t	queue )

This is an interface to any custom sparse matrix vector product.

It should compute y = alpha*FUNCTION(x) + beta*y The vectors are located on the device, the scalars on the CPU.

Parameters

[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	alpha	float scalar alpha
[in]	x	float * input vector x
[in]	beta	float scalar beta
[out]	y	float * output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_s_spmv()

magma_int_t magma_s_spmv	(	float	alpha,
		magma_s_matrix	A,
		magma_s_matrix	x,
		float	beta,
		magma_s_matrix	y,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and scalars alpha, beta the wrapper determines the suitable SpMV computing y = alpha * A * x + beta * y.

Parameters

[in]	alpha	float scalar alpha
[in]	A	magma_s_matrix sparse matrix A
[in]	x	magma_s_matrix input vector x
[in]	beta	float scalar beta
[out]	y	magma_s_matrix output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_s_spmv_shift()

magma_int_t magma_s_spmv_shift	(	float	alpha,
		magma_s_matrix	A,
		float	lambda,
		magma_s_matrix	x,
		float	beta,
		magma_int_t	offset,
		magma_int_t	blocksize,
		magma_index_t *	add_rows,
		magma_s_matrix	y,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and scalars alpha, beta the wrapper determines the suitable SpMV computing y = alpha * ( A - lambda I ) * x + beta * y.

Parameters

	alpha	float scalar alpha
	A	magma_s_matrix sparse matrix A
	lambda	float scalar lambda
	x	magma_s_matrix input vector x
	beta	float scalar beta
	offset	magma_int_t in case not the main diagonal is scaled
	blocksize	magma_int_t in case of processing multiple vectors
	add_rows	magma_int_t* in case the matrixpowerskernel is used
	y	magma_s_matrix output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_s_spmm()

magma_int_t magma_s_spmm	(	float	alpha,
		magma_s_matrix	A,
		magma_s_matrix	B,
		magma_s_matrix *	C,
		magma_queue_t	queue )

For a given input matrix A and B and scalar alpha, the wrapper determines the suitable SpMV computing C = alpha * A * B.

Parameters

[in]	alpha	float scalar alpha
[in]	A	magma_s_matrix sparse matrix A
[in]	B	magma_s_matrix sparse matrix C
[out]	C	magma_s_matrix * outpur sparse matrix C
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_scuspaxpy()

magma_int_t magma_scuspaxpy	(	float *	alpha,
		magma_s_matrix	A,
		float *	beta,
		magma_s_matrix	B,
		magma_s_matrix *	AB,
		magma_queue_t	queue )

This is an interface to the cuSPARSE routine csrgeam computing the sum of two sparse matrices stored in csr format:

C = alpha * A + beta * B

Parameters

[in]	alpha	float* scalar
[in]	A	magma_s_matrix input matrix
[in]	beta	float* scalar
[in]	B	magma_s_matrix input matrix
[out]	AB	magma_s_matrix* output matrix AB = alpha * A + beta * B
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_scuspmm()

magma_int_t magma_scuspmm	(	magma_s_matrix	A,
		magma_s_matrix	B,
		magma_s_matrix *	AB,
		magma_queue_t	queue )

This is an interface to the cuSPARSE routine csrmm computing the product of two sparse matrices stored in csr format.

Parameters

[in]	A	magma_s_matrix input matrix
[in]	B	magma_s_matrix input matrix
[out]	AB	magma_s_matrix* output matrix AB = A * B
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sge3pt()

magma_int_t magma_sge3pt	(	magma_int_t	m,
		magma_int_t	n,
		float	alpha,
		float	beta,
		magmaFloat_ptr	dx,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine is a 3-pt-stencil operator derived from a FD-scheme in 2D with Dirichlet boundary.

It computes y_i = -2 x_i + x_{i-1} + x_{i+1}

Parameters

[in]	m	magma_int_t number of rows in x and y
[in]	n	magma_int_t number of columns in x and y
[in]	alpha	float scalar multiplier
[in]	beta	float scalar multiplier
[in]	dx	magmaFloat_ptr input vector x
[out]	dy	magmaFloat_ptr output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgeaxpy()

magma_int_t magma_sgeaxpy	(	float	alpha,
		magma_s_matrix	X,
		float	beta,
		magma_s_matrix *	Y,
		magma_queue_t	queue )

This routine computes Y = alpha * X + beta * Y on the GPU.

The input format is magma_s_matrix. It can handle both, dense matrix (vector block) and CSR matrices. For the latter, it interfaces the cuSPARSE library.

Parameters

[in]	alpha	float scalar multiplier.
[in]	X	magma_s_matrix input/output matrix Y.
[in]	beta	float scalar multiplier.
[in,out]	Y	magma_s_matrix* input matrix X.
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgecsr5mv()

magma_int_t magma_sgecsr5mv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	p,
		float	alpha,
		magma_int_t	sigma,
		magma_int_t	bit_y_offset,
		magma_int_t	bit_scansum_offset,
		magma_int_t	num_packet,
		magmaUIndex_ptr	dtile_ptr,
		magmaUIndex_ptr	dtile_desc,
		magmaIndex_ptr	dtile_desc_offset_ptr,
		magmaIndex_ptr	dtile_desc_offset,
		magmaFloat_ptr	dcalibrator,
		magma_int_t	tail_tile_start,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	drowptr,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * A * x + beta * y on the GPU.

The input format is CSR5 (val (tile-wise column-major), row_pointer, col (tile-wise column-major), tile_pointer, tile_desc).

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	p	magma_int_t number of tiles in A
[in]	alpha	float scalar multiplier
[in]	sigma	magma_int_t sigma in A in CSR5
[in]	bit_y_offset	magma_int_t bit_y_offset in A in CSR5
[in]	bit_scansum_offset	magma_int_t bit_scansum_offset in A in CSR5
[in]	num_packet	magma_int_t num_packet in A in CSR5
[in]	dtile_ptr	magmaUIndex_ptr tilepointer of A in CSR5
[in]	dtile_desc	magmaUIndex_ptr tiledescriptor of A in CSR5
[in]	dtile_desc_offset_ptr	magmaIndex_ptr tiledescriptor_offsetpointer of A in CSR5
[in]	dtile_desc_offset	magmaIndex_ptr tiledescriptor_offsetpointer of A in CSR5
[in]	dcalibrator	magmaFloat_ptr calibrator of A in CSR5
[in]	tail_tile_start	magma_int_t start of the last tile in A
[in]	dval	magmaFloat_ptr array containing values of A in CSR
[in]	dval	magmaFloat_ptr array containing values of A in CSR
[in]	drowptr	magmaIndex_ptr rowpointer of A in CSR
[in]	dcolind	magmaIndex_ptr columnindices of A in CSR
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgecsrmv()

magma_int_t magma_sgecsrmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	drowptr,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * A * x + beta * y on the GPU.

The input format is CSR (val, row, col).

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in CSR
[in]	drowptr	magmaIndex_ptr rowpointer of A in CSR
[in]	dcolind	magmaIndex_ptr columnindices of A in CSR
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgecsrmv_shift()

magma_int_t magma_sgecsrmv_shift	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		float	alpha,
		float	lambda,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	drowptr,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magma_int_t	offset,
		magma_int_t	blocksize,
		magma_index_t *	addrows,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * ( A -lambda I ) * x + beta * y on the GPU.

It is a shifted version of the CSR-SpMV.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	alpha	float scalar multiplier
[in]	lambda	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in CSR
[in]	drowptr	magmaIndex_ptr rowpointer of A in CSR
[in]	dcolind	magmaIndex_ptr columnindices of A in CSR
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[in]	offset	magma_int_t in case not the main diagonal is scaled
[in]	blocksize	magma_int_t in case of processing multiple vectors
[in]	addrows	magmaIndex_ptr in case the matrixpowerskernel is used
[out]	dy	magmaFloat_ptr output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgecsrreimsplit()

magma_int_t magma_sgecsrreimsplit	(	magma_s_matrix	A,
		magma_s_matrix *	ReA,
		magma_s_matrix *	ImA,
		magma_queue_t	queue )

This routine takes an input matrix A in CSR format and located on the GPU and splits it into two matrixes ReA and ImA containing the real and the imaginary contributions of A.

The output matrices are allocated within the routine.

Parameters

[in]	A	magma_s_matrix input matrix A.
[out]	ReA	magma_s_matrix* output matrix contaning real contributions.
[out]	ImA	magma_s_matrix* output matrix contaning real contributions.
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgedensereimsplit()

magma_int_t magma_sgedensereimsplit	(	magma_s_matrix	A,
		magma_s_matrix *	ReA,
		magma_s_matrix *	ImA,
		magma_queue_t	queue )

This routine takes an input matrix A in DENSE format and located on the GPU and splits it into two matrixes ReA and ImA containing the real and the imaginary contributions of A.

The output matrices are allocated within the routine.

Parameters

[in]	A	magma_s_matrix input matrix A.
[out]	ReA	magma_s_matrix* output matrix contaning real contributions.
[out]	ImA	magma_s_matrix* output matrix contaning real contributions.
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgeellmv()

magma_int_t magma_sgeellmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nnz_per_row,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * A * x + beta * y on the GPU.

Input format is ELLPACK.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	nnz_per_row	magma_int_t number of elements in the longest row
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in ELLPACK
[in]	dcolind	magmaIndex_ptr columnindices of A in ELLPACK
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgeellmv_shift()

magma_int_t magma_sgeellmv_shift	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nnz_per_row,
		float	alpha,
		float	lambda,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magma_int_t	offset,
		magma_int_t	blocksize,
		magmaIndex_ptr	addrows,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha *( A - lambda I ) * x + beta * y on the GPU.

Input format is ELLPACK. It is the shifted version of the ELLPACK SpMV.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	nnz_per_row	magma_int_t number of elements in the longest row
[in]	alpha	float scalar multiplier
[in]	lambda	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in ELLPACK
[in]	dcolind	magmaIndex_ptr columnindices of A in ELLPACK
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[in]	offset	magma_int_t in case not the main diagonal is scaled
[in]	blocksize	magma_int_t in case of processing multiple vectors
[in]	addrows	magmaIndex_ptr in case the matrixpowerskernel is used
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgeellrtmv()

magma_int_t magma_sgeellrtmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nnz_per_row,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaIndex_ptr	drowlength,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_int_t	alignment,
		magma_int_t	blocksize,
		magma_queue_t	queue )

This routine computes y = alpha * A * x + beta * y on the GPU.

Input format is ELLRT. The ideas are taken from "Improving the performance of the sparse matrix vector product with GPUs", (CIT 2010), and modified to provide correct values.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	nnz_per_row	magma_int_t max number of nonzeros in a row
[in]	alpha	float scalar alpha
[in]	dval	magmaFloat_ptr val array
[in]	dcolind	magmaIndex_ptr col indices
[in]	drowlength	magmaIndex_ptr number of elements in each row
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar beta
[out]	dy	magmaFloat_ptr output vector y
[in]	blocksize	magma_int_t threads per block
[in]	alignment	magma_int_t threads assigned to each row
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgeelltmv_shift()

magma_int_t magma_sgeelltmv_shift	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	nnz_per_row,
		float	alpha,
		float	lambda,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magma_int_t	offset,
		magma_int_t	blocksize,
		magmaIndex_ptr	addrows,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha *( A - lambda I ) * x + beta * y on the GPU.

Input format is ELL.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	nnz_per_row	magma_int_t number of elements in the longest row
[in]	alpha	float scalar multiplier
[in]	lambda	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in ELL
[in]	dcolind	magmaIndex_ptr columnindices of A in ELL
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[in]	offset	magma_int_t in case not the main diagonal is scaled
[in]	blocksize	magma_int_t in case of processing multiple vectors
[in]	addrows	magmaIndex_ptr in case the matrixpowerskernel is used
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smdotc()

magma_int_t magma_smdotc	(	magma_int_t	n,
		magma_int_t	k,
		magmaFloat_ptr	v,
		magmaFloat_ptr	r,
		magmaFloat_ptr	d1,
		magmaFloat_ptr	d2,
		magmaFloat_ptr	skp,
		magma_queue_t	queue )

Computes the scalar product of a set of vectors v_i such that.

skp = ( <v_0,r>, <v_1,r>, .. )

Returns the vector skp.

Parameters

[in]	n	int length of v_i and r
[in]	k	int

vectors v_i

Parameters

[in]	v	magmaFloat_ptr v = (v_0 .. v_i.. v_k)
[in]	r	magmaFloat_ptr r
[in]	d1	magmaFloat_ptr workspace
[in]	d2	magmaFloat_ptr workspace
[out]	skp	magmaFloat_ptr vector[k] of scalar products (<v_i,r>...)
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgesellpmv()

magma_int_t magma_sgesellpmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	blocksize,
		magma_int_t	slices,
		magma_int_t	alignment,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaIndex_ptr	drowptr,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * A^t * x + beta * y on the GPU.

Input format is SELLP.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	blocksize	magma_int_t number of rows in one ELL-slice
[in]	slices	magma_int_t number of slices in matrix
[in]	alignment	magma_int_t number of threads assigned to one row
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in SELLP
[in]	dcolind	magmaIndex_ptr columnindices of A in SELLP
[in]	drowptr	magmaIndex_ptr rowpointer of SELLP
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_sgesellcmv()

magma_int_t magma_sgesellcmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	blocksize,
		magma_int_t	slices,
		magma_int_t	alignment,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaIndex_ptr	drowptr,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes y = alpha * A^t * x + beta * y on the GPU.

Input format is SELLC/SELLP.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	blocksize	magma_int_t number of rows in one ELL-slice
[in]	slices	magma_int_t number of slices in matrix
[in]	alignment	magma_int_t number of threads assigned to one row (=1)
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in SELLC/P
[in]	dcolind	magmaIndex_ptr columnindices of A in SELLC/P
[in]	drowptr	magmaIndex_ptr rowpointer of SELLP
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smdotc_shfl()

magma_int_t magma_smdotc_shfl	(	magma_int_t	n,
		magma_int_t	k,
		magmaFloat_ptr	v,
		magmaFloat_ptr	r,
		magmaFloat_ptr	d1,
		magmaFloat_ptr	d2,
		magmaFloat_ptr	skp,
		magma_queue_t	queue )

Computes the scalar product of a set of vectors v_i such that.

skp = ( <v_0,r>, <v_1,r>, .. )

Returns the vector skp.

Parameters

[in]	n	int length of v_i and r
[in]	k	int

vectors v_i

Parameters

[in]	v	magmaFloat_ptr v = (v_0 .. v_i.. v_k)
[in]	r	magmaFloat_ptr r
[in]	d1	magmaFloat_ptr workspace
[in]	d2	magmaFloat_ptr workspace
[out]	skp	magmaFloat_ptr vector[k] of scalar products (<v_i,r>...)
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smdotc4()

magma_int_t magma_smdotc4	(	magma_int_t	n,
		magmaFloat_ptr	v0,
		magmaFloat_ptr	w0,
		magmaFloat_ptr	v1,
		magmaFloat_ptr	w1,
		magmaFloat_ptr	v2,
		magmaFloat_ptr	w2,
		magmaFloat_ptr	v3,
		magmaFloat_ptr	w3,
		magmaFloat_ptr	d1,
		magmaFloat_ptr	d2,
		magmaFloat_ptr	skp,
		magma_queue_t	queue )

Computes the scalar product of a set of 4 vectors such that.

skp[0,1,2,3] = [ <v_0,w_0>, <v_1,w_1>, <v_2,w_2>, <v3,w_3> ]

Returns the vector skp. In case there are less dot products required, an easy workaround is given by doubling input.

Parameters

[in]	n	int length of v_i and w_i
[in]	v0	magmaFloat_ptr input vector
[in]	w0	magmaFloat_ptr input vector
[in]	v1	magmaFloat_ptr input vector
[in]	w1	magmaFloat_ptr input vector
[in]	v2	magmaFloat_ptr input vector
[in]	w2	magmaFloat_ptr input vector
[in]	v3	magmaFloat_ptr input vector
[in]	w3	magmaFloat_ptr input vector
[in]	d1	magmaFloat_ptr workspace
[in]	d2	magmaFloat_ptr workspace
[out]	skp	magmaFloat_ptr vector[4] of scalar products [<v_i, w_i>] This vector is located on the host
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smgecsrmv()

magma_int_t magma_smgecsrmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	num_vecs,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	drowptr,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

Input format is CSR.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	num_vecs	mama_int_t number of vectors
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in CSR
[in]	drowptr	magmaIndex_ptr rowpointer of A in CSR
[in]	dcolind	magmaIndex_ptr columnindices of A in CSR
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smgeellmv()

magma_int_t magma_smgeellmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	num_vecs,
		magma_int_t	nnz_per_row,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

Input format is ELLPACK.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	num_vecs	mama_int_t number of vectors
[in]	nnz_per_row	magma_int_t number of elements in the longest row
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in ELLPACK
[in]	dcolind	magmaIndex_ptr columnindices of A in ELLPACK
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smgeelltmv()

magma_int_t magma_smgeelltmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	num_vecs,
		magma_int_t	nnz_per_row,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU.

Input format is ELL.

Parameters

[in]	transA	magma_trans_t transposition parameter for A
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	num_vecs	mama_int_t number of vectors
[in]	nnz_per_row	magma_int_t number of elements in the longest row
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in ELL
[in]	dcolind	magmaIndex_ptr columnindices of A in ELL
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

◆ magma_smgesellpmv()

magma_int_t magma_smgesellpmv	(	magma_trans_t	transA,
		magma_int_t	m,
		magma_int_t	n,
		magma_int_t	num_vecs,
		magma_int_t	blocksize,
		magma_int_t	slices,
		magma_int_t	alignment,
		float	alpha,
		magmaFloat_ptr	dval,
		magmaIndex_ptr	dcolind,
		magmaIndex_ptr	drowptr,
		magmaFloat_ptr	dx,
		float	beta,
		magmaFloat_ptr	dy,
		magma_queue_t	queue )

This routine computes Y = alpha * A^t * X + beta * Y on the GPU.

Input format is SELLP. Note, that the input format for X is row-major while the output format for Y is column major!

Parameters

[in]	transA	magma_trans_t transpose A?
[in]	m	magma_int_t number of rows in A
[in]	n	magma_int_t number of columns in A
[in]	num_vecs	magma_int_t number of columns in X and Y
[in]	blocksize	magma_int_t number of rows in one ELL-slice
[in]	slices	magma_int_t number of slices in matrix
[in]	alignment	magma_int_t number of threads assigned to one row
[in]	alpha	float scalar multiplier
[in]	dval	magmaFloat_ptr array containing values of A in SELLP
[in]	dcolind	magmaIndex_ptr columnindices of A in SELLP
[in]	drowptr	magmaIndex_ptr rowpointer of SELLP
[in]	dx	magmaFloat_ptr input vector x
[in]	beta	float scalar multiplier
[out]	dy	magmaFloat_ptr input/output vector y
[in]	queue	magma_queue_t Queue to execute in.

Functions

Detailed Description

Function Documentation

◆ magma_scustomspmv()

◆ magma_s_spmv()

◆ magma_s_spmv_shift()

◆ magma_s_spmm()

◆ magma_scuspaxpy()

◆ magma_scuspmm()

◆ magma_sge3pt()

◆ magma_sgeaxpy()

◆ magma_sgecsr5mv()

◆ magma_sgecsrmv()

◆ magma_sgecsrmv_shift()

◆ magma_sgecsrreimsplit()

◆ magma_sgedensereimsplit()

◆ magma_sgeellmv()

◆ magma_sgeellmv_shift()

◆ magma_sgeellrtmv()

◆ magma_sgeelltmv_shift()

◆ magma_smdotc()

vectors v_i

◆ magma_sgesellpmv()

◆ magma_sgesellcmv()

◆ magma_smdotc_shfl()

vectors v_i

◆ magma_smdotc4()

◆ magma_smgecsrmv()

◆ magma_smgeellmv()

◆ magma_smgeelltmv()

◆ magma_smgesellpmv()