single precision

Functions
magma_int_t	magma_s_precond (magma_s_matrix A, magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	For a given input matrix A and vectors x, y and the preconditioner parameters, the respective preconditioner is chosen.
magma_int_t	magma_s_precondsetup (magma_s_matrix A, magma_s_matrix b, magma_s_solver_par *solver, magma_s_preconditioner *precond, magma_queue_t queue)
	For a given input matrix M and vectors x, y and the preconditioner parameters, the respective preconditioner is preprocessed.
magma_int_t	magma_s_applyprecond (magma_s_matrix A, magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	For a given input matrix A and vectors x, y and the preconditioner parameters, the respective preconditioner is applied.
magma_int_t	magma_s_applyprecond_left (magma_trans_t trans, magma_s_matrix A, magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	For a given input matrix A and vectors x, y and the preconditioner parameters, the respective left preconditioner is applied.
magma_int_t	magma_s_applyprecond_right (magma_trans_t trans, magma_s_matrix A, magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	For a given input matrix A and vectors x, y and the preconditioner parameters, the respective right-preconditioner is applied.
magma_int_t	magma_s_solver (magma_s_matrix A, magma_s_matrix b, magma_s_matrix *x, magma_sopts *zopts, magma_queue_t queue)
	This is an interface that allows to use any iterative solver on the linear system Ax = b.
magma_int_t	magma_sapplycustomprecond_l (magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	This is an interface to the left solve for any custom preconditioner.
magma_int_t	magma_sapplycustomprecond_r (magma_s_matrix b, magma_s_matrix *x, magma_s_preconditioner *precond, magma_queue_t queue)
	This is an interface to the right solve for any custom preconditioner.
magma_int_t	magma_sresidual (magma_s_matrix A, magma_s_matrix b, magma_s_matrix x, float *res, magma_queue_t queue)
	Computes the residual ||b-Ax|| for a solution approximation x.
magma_int_t	magma_sresidual_slice (magma_int_t start, magma_int_t end, magma_s_matrix A, magma_s_matrix b, magma_s_matrix x, float *res, magma_queue_t queue)
	Computes the residual r=||b-Ax|| for the slice r(start:end) for a solution approximation x.
magma_int_t	magma_sresidualvec (magma_s_matrix A, magma_s_matrix b, magma_s_matrix x, magma_s_matrix *r, float *res, magma_queue_t queue)
	Computes the residual r = b-Ax for a solution approximation x.
magma_int_t	magma_scsrsplit (magma_int_t offset, magma_int_t bsize, magma_s_matrix A, magma_s_matrix *D, magma_s_matrix *R, magma_queue_t queue)
	Splits a CSR matrix into two matrices, one containing the diagonal blocks with the diagonal element stored first, one containing the rest of the original matrix.
magma_int_t	magma_smfree (magma_s_matrix *A, magma_queue_t queue)
	Free the memory of a magma_s_matrix.
magma_int_t	magma_sprecondfree (magma_s_preconditioner *precond_par, magma_queue_t queue)
	Free a preconditioner.
magma_int_t	magma_sfrobenius (magma_s_matrix A, magma_s_matrix B, real_Double_t *res, magma_queue_t queue)
	Computes the Frobenius norm of the difference between the CSR matrices A and B.
magma_int_t	magma_smatrix_addrowindex (magma_s_matrix *A, magma_queue_t queue)
	Adds to a CSR matrix an array containing the rowindexes.
magma_int_t	magma_scsrcoo_transpose (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Transposes a matrix that already contains rowidx.
magma_int_t	magma_smatrix_createrowptr (magma_int_t n, magma_index_t *row, magma_queue_t queue)
	This function generates a rowpointer out of a row-wise element count in parallel.
magma_int_t	magma_smatrix_swap (magma_s_matrix *A, magma_s_matrix *B, magma_queue_t queue)
	Swaps two matrices.
magma_int_t	magma_smatrix_tril (magma_s_matrix A, magma_s_matrix *L, magma_queue_t queue)
	Extracts the lower triangular of a matrix: L = tril(A).
magma_int_t	magma_smatrix_triu (magma_s_matrix A, magma_s_matrix *U, magma_queue_t queue)
	Extracts the lower triangular of a matrix: U = triu(A).
magma_int_t	magma_scsr_sort (magma_s_matrix *A, magma_queue_t queue)
	SOrts the elements in a CSR matrix for increasing column index.
magma_int_t	magma_srowentries (magma_s_matrix *A, magma_queue_t queue)
	Checks the maximal number of nonzeros in a row of matrix A.
magma_int_t	magma_s_csr_compressor (float **val, magma_index_t **row, magma_index_t **col, float **valn, magma_index_t **rown, magma_index_t **coln, magma_int_t *n, magma_queue_t queue)
	Helper function to compress CSR containing zero-entries.
magma_int_t	magma_smconvert (magma_s_matrix A, magma_s_matrix *B, magma_storage_t old_format, magma_storage_t new_format, magma_queue_t queue)
	Converter between different sparse storage formats.
magma_int_t	magma_smcsrcompressor (magma_s_matrix *A, magma_queue_t queue)
	Removes zeros in a CSR matrix.
magma_int_t	magma_scsrset (magma_int_t m, magma_int_t n, magma_index_t *row, magma_index_t *col, float *val, magma_s_matrix *A, magma_queue_t queue)
	Passes a CSR matrix to MAGMA.
magma_int_t	magma_scsrget (magma_s_matrix A, magma_int_t *m, magma_int_t *n, magma_index_t **row, magma_index_t **col, float **val, magma_queue_t queue)
	Passes a MAGMA matrix to CSR structure.
magma_int_t	magma_scsrset_gpu (magma_int_t m, magma_int_t n, magmaIndex_ptr row, magmaIndex_ptr col, magmaFloat_ptr val, magma_s_matrix *A, magma_queue_t queue)
	Passes a CSR matrix to MAGMA (located on DEV).
magma_int_t	magma_scsrget_gpu (magma_s_matrix A, magma_int_t *m, magma_int_t *n, magmaIndex_ptr *row, magmaIndex_ptr *col, magmaFloat_ptr *val, magma_queue_t queue)
	Passes a MAGMA matrix to CSR structure (located on DEV).
magma_int_t	magma_smdiagdom (magma_s_matrix M, float *min_dd, float *max_dd, float *avg_dd, magma_queue_t queue)
	This routine takes a CSR matrix and computes the average diagonal dominance.
magma_int_t	magma_smbdiagdom (magma_s_matrix M, magma_s_matrix blocksizes, float *min_dd, float *max_dd, float *avg_dd, magma_queue_t queue)
	This routine takes a CSR matrix and computes the average block-diagonal dominance.
magma_int_t	magma_smdiff (magma_s_matrix A, magma_s_matrix B, real_Double_t *res, magma_queue_t queue)
	Computes the Frobenius norm of the difference between the CSR matrices A and B.
magma_int_t	magma_smfrobenius (magma_s_matrix A, magma_s_matrix B, magma_s_matrix S, float *norm, magma_queue_t queue)
	Computes the Frobenius norm || A - B ||_S on the sparsity pattern of S.
magma_int_t	magma_smgenerator (magma_int_t n, magma_int_t offdiags, magma_index_t *diag_offset, float *diag_vals, magma_s_matrix *A, magma_queue_t queue)
	Generate a symmetric n x n CSR matrix for a stencil.
magma_int_t	magma_sm_27stencil (magma_int_t n, magma_s_matrix *A, magma_queue_t queue)
	Generate a 27-point stencil for a 3D FD discretization.
magma_int_t	magma_sm_5stencil (magma_int_t n, magma_s_matrix *A, magma_queue_t queue)
	Generate a 5-point stencil for a 2D FD discretization.
magma_int_t	magma_ssymbilu (magma_s_matrix *A, magma_int_t levels, magma_s_matrix *L, magma_s_matrix *U, magma_queue_t queue)
	This routine performs a symbolic ILU factorization.
magma_int_t	read_s_csr_from_mtx (magma_storage_t *type, magma_location_t *location, magma_int_t *n_row, magma_int_t *n_col, magma_int_t *nnz, float **val, magma_index_t **row, magma_index_t **col, const char *filename, magma_queue_t queue)
	Reads in a matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.
magma_int_t	magma_swrite_csr_mtx (magma_s_matrix A, magma_order_t MajorType, const char *filename, magma_queue_t queue)
	Writes a CSR matrix to a file using Matrix Market format.
magma_int_t	magma_sprint_csr_mtx (magma_int_t n_row, magma_int_t n_col, magma_int_t nnz, float **val, magma_index_t **row, magma_index_t **col, magma_order_t MajorType, magma_queue_t queue)
	Prints a CSR matrix in Matrix Market format.
magma_int_t	magma_sprint_csr (magma_int_t n_row, magma_int_t n_col, magma_int_t nnz, float **val, magma_index_t **row, magma_index_t **col, magma_queue_t queue)
	Prints a CSR matrix in CSR format.
magma_int_t	magma_sprint_matrix (magma_s_matrix A, magma_queue_t queue)
	Prints a sparse matrix in CSR format.
magma_int_t	magma_s_csr_mtx (magma_s_matrix *A, const char *filename, magma_queue_t queue)
	Reads in a matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.
magma_int_t	magma_s_csr_mtxsymm (magma_s_matrix *A, const char *filename, magma_queue_t queue)
	Reads in a SYMMETRIC matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.
magma_int_t	magma_smlumerge (magma_s_matrix L, magma_s_matrix U, magma_s_matrix *A, magma_queue_t queue)
	Takes an strictly lower triangular matrix L and an upper triangular matrix U and merges them into a matrix A containing the upper and lower triangular parts.
magma_int_t	magma_smscale (magma_s_matrix *A, magma_scale_t scaling, magma_queue_t queue)
	Scales a matrix.
magma_int_t	magma_smscale_matrix_rhs (magma_s_matrix *A, magma_s_matrix *b, magma_s_matrix *scaling_factors, magma_scale_t scaling, magma_queue_t queue)
	Scales a matrix and a right hand side vector of a Ax = b system.
magma_int_t	magma_smdiagadd (magma_s_matrix *A, float add, magma_queue_t queue)
	Adds a multiple of the Identity matrix to a matrix: A = A+add * I.
magma_int_t	magma_smscale_generate (magma_int_t n, magma_scale_t *scaling, magma_side_t *side, magma_s_matrix *A, magma_s_matrix *scaling_factors, magma_queue_t queue)
	Generates n vectors of scaling factors from the A matrix and stores them in the factors matrix as column vectors in column major ordering.
magma_int_t	magma_smscale_apply (magma_int_t n, magma_side_t *side, magma_s_matrix *scaling_factors, magma_s_matrix *A, magma_queue_t queue)
	Applies n diagonal scaling matrices to a matrix A; n=[1,2], factor[i] is applied to side[i] of the matrix.
magma_int_t	magma_sdimv (magma_s_matrix *vecA, magma_s_matrix *vecB, magma_queue_t queue)
	Multiplies a diagonal matrix (vecA) and a vector (vecB).
magma_int_t	magma_smshrink (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Shrinks a non-square matrix (m < n) to the smaller dimension.
magma_int_t	magma_smslice (magma_int_t num_slices, magma_int_t slice, magma_s_matrix A, magma_s_matrix *B, magma_s_matrix *ALOC, magma_s_matrix *ANLOC, magma_index_t *comm_i, float *comm_v, magma_int_t *start, magma_int_t *end, magma_queue_t queue)
	Takes a matrix and extracts a slice for solving the system in parallel:
magma_int_t	magma_smtransfer (magma_s_matrix A, magma_s_matrix *B, magma_location_t src, magma_location_t dst, magma_queue_t queue)
	Copies a matrix from memory location src to memory location dst.
magma_int_t	s_transpose_csr (magma_int_t n_rows, magma_int_t n_cols, magma_int_t nnz, float *values, magma_index_t *rowptr, magma_index_t *colind, magma_int_t *new_n_rows, magma_int_t *new_n_cols, magma_int_t *new_nnz, float **new_values, magma_index_t **new_rowptr, magma_index_t **new_colind, magma_queue_t queue)
	Transposes a matrix stored in CSR format on the CPU host.
magma_int_t	magma_smtranspose (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Interface to cuSPARSE transpose.
magma_int_t	magma_s_cucsrtranspose (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Helper function to transpose CSR matrix.
magma_int_t	magma_smtransposeconjugate (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	This function forms the transpose conjugate of a matrix.
magma_int_t	magma_smtranspose_cpu (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Generates a transpose of A on the CPU.
magma_int_t	magma_smtransposeconj_cpu (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Generates a transpose conjugate of A on the CPU.
magma_int_t	magma_smtransposestruct_cpu (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Generates a transpose of the nonzero pattern of A on the CPU.
magma_int_t	magma_smtransposeabs_cpu (magma_s_matrix A, magma_s_matrix *B, magma_queue_t queue)
	Generates a transpose with absolute values of A on the CPU.
magma_int_t	magma_ssolverinfo (magma_s_solver_par *solver_par, magma_s_preconditioner *precond_par, magma_queue_t queue)
	Prints information about a previously called solver.
magma_int_t	magma_ssolverinfo_free (magma_s_solver_par *solver_par, magma_s_preconditioner *precond_par, magma_queue_t queue)
	Frees any memory assocoiated with the verbose mode of solver_par.
magma_int_t	magma_ssolverinfo_init (magma_s_solver_par *solver_par, magma_s_preconditioner *precond_par, magma_queue_t queue)
	Initializes all solver and preconditioner parameters.
magma_int_t	magma_seigensolverinfo_init (magma_s_solver_par *solver_par, magma_queue_t queue)
	Initializes space for eigensolvers.
magma_int_t	magma_ssort (float *x, magma_int_t first, magma_int_t last, magma_queue_t queue)
	Sorts an array of values in increasing order.
magma_int_t	magma_smsort (float *x, magma_index_t *col, magma_index_t *row, magma_int_t first, magma_int_t last, magma_queue_t queue)
	Sorts an array of values in increasing order.
magma_int_t	magma_sindexsort (magma_index_t *x, magma_int_t first, magma_int_t last, magma_queue_t queue)
	Sorts an array of integers in increasing order.
magma_int_t	magma_sindexsortval (magma_index_t *x, float *y, magma_int_t first, magma_int_t last, magma_queue_t queue)
	Sorts an array of integers, updates a respective array of values.
magma_int_t	magma_sorderstatistics (float *val, magma_int_t length, magma_int_t k, magma_int_t r, float *element, magma_queue_t queue)
	Identifies the kth smallest/largest element in an array.
magma_int_t	magma_sbitonic_sort (magma_int_t start, magma_int_t length, float *seq, magma_int_t flag, magma_queue_t queue)
	Approximates the k-th smallest element in an array by using order-statistics with step-size inc.
magma_int_t	magma_sparse_opts (int argc, char **argv, magma_sopts *opts, int *matrices, magma_queue_t queue)
	Parses input options for a solver.
magma_int_t	magma_svinit (magma_s_matrix *x, magma_location_t mem_loc, magma_int_t num_rows, magma_int_t num_cols, float values, magma_queue_t queue)
	Allocates memory for magma_s_matrix and initializes it with the passed value.
magma_int_t	magma_svinit_rand (magma_s_matrix *x, magma_location_t mem_loc, magma_int_t num_rows, magma_int_t num_cols, magma_queue_t queue)
	Allocates memory for magma_s_matrix and initializes it with random values.
magma_int_t	magma_sprint_vector (magma_s_matrix x, magma_int_t offset, magma_int_t visulen, magma_queue_t queue)
	Visualizes part of a vector of type magma_s_matrix.
magma_int_t	magma_svread (magma_s_matrix *x, magma_int_t length, char *filename, magma_queue_t queue)
	Reads in a float vector of length "length".
magma_int_t	magma_swrite_vector (magma_s_matrix A, const char *filename, magma_queue_t queue)
	Writes a vector to a file.
magma_int_t	magma_svset (magma_int_t m, magma_int_t n, float *val, magma_s_matrix *v, magma_queue_t queue)
	Passes a vector to MAGMA.
magma_int_t	magma_svcopy (magma_s_matrix v, magma_int_t *m, magma_int_t *n, float *val, magma_queue_t queue)
	Passes a MAGMA vector back.
magma_int_t	magma_svset_dev (magma_int_t m, magma_int_t n, magmaFloat_ptr val, magma_s_matrix *v, magma_queue_t queue)
	Passes a vector to MAGMA (located on DEV).
magma_int_t	magma_svget (magma_s_matrix v, magma_int_t *m, magma_int_t *n, float **val, magma_queue_t queue)
	Passes a MAGMA vector back.
magma_int_t	magma_svget_dev (magma_s_matrix v, magma_int_t *m, magma_int_t *n, magmaFloat_ptr *val, magma_queue_t queue)
	Passes a MAGMA vector back (located on DEV).
magma_int_t	magma_svcopy_dev (magma_s_matrix v, magma_int_t *m, magma_int_t *n, magmaFloat_ptr val, magma_queue_t queue)
	Passes a MAGMA vector back (located on DEV).
magma_int_t	magma_svtranspose (magma_s_matrix x, magma_s_matrix *y, magma_queue_t queue)
	Transposes a vector from col to row major and vice versa.
magma_int_t	magma_sdiagcheck (magma_s_matrix dA, magma_queue_t queue)
	This routine checks for a CSR matrix whether there exists a zero on the diagonal.
magma_int_t	magma_scsr_sort_gpu (magma_s_matrix *A, magma_queue_t queue)
	Generates a matrix \(U = A \cup B\).
magma_int_t	magma_smconjugate (magma_s_matrix *A, magma_queue_t queue)
	This function conjugates a matrix.
magma_int_t	magma_smcsrcompressor_gpu (magma_s_matrix *A, magma_queue_t queue)
	Removes zeros in a CSR matrix.
magma_int_t	magma_slobpcg_res (magma_int_t num_rows, magma_int_t num_vecs, magmaFloat_ptr evalues, magmaFloat_ptr X, magmaFloat_ptr R, magmaFloat_ptr res, magma_queue_t queue)
	This routine computes for Block-LOBPCG, the set of residuals.
magma_int_t	magma_slobpcg_shift (magma_int_t num_rows, magma_int_t num_vecs, magma_int_t shift, magmaFloat_ptr x, magma_queue_t queue)
	For a Block-LOBPCG, the set of residuals (entries consecutive in memory) shrinks and the vectors are shifted in case shift residuals drop below threshold.

Detailed Description

Function Documentation

magma_s_precond()

magma_int_t magma_s_precond	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and the preconditioner parameters, the respective preconditioner is chosen.

It approximates x for A x = y.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[in]	b	magma_s_matrix input vector b
[in]	x	magma_s_matrix* output vector x
[in,out]	precond	magma_s_preconditioner preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_s_precondsetup()

magma_int_t magma_s_precondsetup	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_solver_par *	solver,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

For a given input matrix M and vectors x, y and the preconditioner parameters, the respective preconditioner is preprocessed.

E.g. for Jacobi: the scaling-vetor, for ILU the factorization.

Parameters

[in]	A	magma_s_matrix sparse matrix M
[in]	b	magma_s_matrix input vector y
[in]	solver	magma_s_solver_par solver structure using the preconditioner
[in,out]	precond	magma_s_preconditioner preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_s_applyprecond()

magma_int_t magma_s_applyprecond	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and the preconditioner parameters, the respective preconditioner is applied.

E.g. for Jacobi: the scaling-vetor, for ILU the triangular solves.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[in]	b	magma_s_matrix input vector b
[in,out]	x	magma_s_matrix* output vector x
[in]	precond	magma_s_preconditioner preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_s_applyprecond_left()

magma_int_t magma_s_applyprecond_left	(	magma_trans_t	trans,
		magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and the preconditioner parameters, the respective left preconditioner is applied.

E.g. for Jacobi: the scaling-vetor, for ILU the left triangular solve.

Parameters

[in]	trans	magma_trans_t mode of the preconditioner: MagmaTrans or MagmaNoTrans
[in]	A	magma_s_matrix sparse matrix A
[in]	b	magma_s_matrix input vector b
[in,out]	x	magma_s_matrix* output vector x
[in]	precond	magma_s_preconditioner preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_s_applyprecond_right()

magma_int_t magma_s_applyprecond_right	(	magma_trans_t	trans,
		magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

For a given input matrix A and vectors x, y and the preconditioner parameters, the respective right-preconditioner is applied.

E.g. for Jacobi: the scaling-vetor, for ILU the right triangular solve.

Parameters

[in]	trans	magma_trans_t mode of the preconditioner: MagmaTrans or MagmaNoTrans
[in]	A	magma_s_matrix sparse matrix A
[in]	b	magma_s_matrix input vector b
[in,out]	x	magma_s_matrix* output vector x
[in]	precond	magma_s_preconditioner preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_s_solver()

magma_int_t magma_s_solver	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_sopts *	zopts,
		magma_queue_t	queue )

This is an interface that allows to use any iterative solver on the linear system Ax = b.

All linear algebra objects are expected to be on the device, the linear algebra objects are MAGMA-sparse specific structures (dense matrix b, dense matrix x, sparse/dense matrix A). The additional parameter zopts contains information about the solver and the preconditioner. the type of solver the relative / absolute stopping criterion the maximum number of iterations the preconditioner type ... Please see magmasparse_types.h for details about the fields and magma_sutil_sparse.cpp for the possible options.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[in]	b	magma_s_matrix input vector b
[in]	x	magma_s_matrix* output vector x
[in]	zopts	magma_sopts options for solver and preconditioner
[in]	queue	magma_queue_t Queue to execute in.

magma_sapplycustomprecond_l()

magma_int_t magma_sapplycustomprecond_l	(	magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

This is an interface to the left solve for any custom preconditioner.

It should compute x = FUNCTION(b) The vectors are located on the device.

Parameters

[in]	b	magma_s_matrix RHS
[in,out]	x	magma_s_matrix* vector to precondition
[in,out]	precond	magma_s_preconditioner* preconditioner parameters
[in]	queue	magma_queue_t Queue to execute in.

magma_sapplycustomprecond_r()

magma_int_t magma_sapplycustomprecond_r	(	magma_s_matrix	b,
		magma_s_matrix *	x,
		magma_s_preconditioner *	precond,
		magma_queue_t	queue )

This is an interface to the right solve for any custom preconditioner.

It should compute x = FUNCTION(b) The vectors are located on the device.

Parameters

[in]	b	magma_s_matrix RHS
[in,out]	x	magma_s_matrix* vector to precondition
[in,out]	precond	magma_s_preconditioner* preconditioner parameters
[in]	queue	magma_queue_t Queue to execute in.

magma_sresidual()

magma_int_t magma_sresidual	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix	x,
		float *	res,
		magma_queue_t	queue )

Computes the residual ||b-Ax|| for a solution approximation x.

Parameters

[in]	A	magma_s_matrix input matrix A
[in]	b	magma_s_matrix RHS b
[in]	x	magma_s_matrix solution approximation
[out]	res	float* return residual
[in]	queue	magma_queue_t Queue to execute in.

magma_sresidual_slice()

magma_int_t magma_sresidual_slice	(	magma_int_t	start,
		magma_int_t	end,
		magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix	x,
		float *	res,
		magma_queue_t	queue )

Computes the residual r=||b-Ax|| for the slice r(start:end) for a solution approximation x.

Parameters

[in]	start	magma_int_t start of slice (row-index)
[in]	end	magma_int_t end of slice (row-index)
[in]	A	magma_s_matrix input matrix A
[in]	b	magma_s_matrix RHS b
[in]	x	magma_s_matrix solution approximation
[out]	res	float* return residual
[in]	queue	magma_queue_t Queue to execute in.

magma_sresidualvec()

magma_int_t magma_sresidualvec	(	magma_s_matrix	A,
		magma_s_matrix	b,
		magma_s_matrix	x,
		magma_s_matrix *	r,
		float *	res,
		magma_queue_t	queue )

Computes the residual r = b-Ax for a solution approximation x.

It returns both, the actual residual and the residual vector

Parameters

[in]	A	magma_s_matrix input matrix A
[in]	b	magma_s_matrix RHS b
[in]	x	magma_s_matrix solution approximation
[in,out]	r	magma_s_matrix* residual vector
[out]	res	float* return residual
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrsplit()

magma_int_t magma_scsrsplit	(	magma_int_t	offset,
		magma_int_t	bsize,
		magma_s_matrix	A,
		magma_s_matrix *	D,
		magma_s_matrix *	R,
		magma_queue_t	queue )

Splits a CSR matrix into two matrices, one containing the diagonal blocks with the diagonal element stored first, one containing the rest of the original matrix.

Parameters

[in]	offset	magma_int_t size of the first block
[in]	bsize	magma_int_t size of the diagonal blocks
[in]	A	magma_s_matrix CSR input matrix
[out]	D	magma_s_matrix* CSR matrix containing diagonal blocks
[out]	R	magma_s_matrix* CSR matrix containing rest
[in]	queue	magma_queue_t Queue to execute in.

magma_smfree()

magma_int_t magma_smfree	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Free the memory of a magma_s_matrix.

Note, this routine performs a magma_queue_sync on the queue passed to it prior to freeing any memory.

Parameters

[in,out]	A	magma_s_matrix* matrix to free
[in]	queue	magma_queue_t Queue to execute in.

magma_sprecondfree()

magma_int_t magma_sprecondfree	(	magma_s_preconditioner *	precond_par,
		magma_queue_t	queue )

Free a preconditioner.

Note, this routine performs a magma_queue_sync on the queue passed to it prior to freeing any memory.

Parameters

[in,out]	precond_par	magma_s_preconditioner* structure containing all preconditioner information
[in]	queue	magma_queue_t Queue to execute in.

magma_sfrobenius()

magma_int_t magma_sfrobenius	(	magma_s_matrix	A,
		magma_s_matrix	B,
		real_Double_t *	res,
		magma_queue_t	queue )

Computes the Frobenius norm of the difference between the CSR matrices A and B.

They need to share the same sparsity pattern!

Parameters

[in]	A	magma_s_matrix sparse matrix in CSR
[in]	B	magma_s_matrix sparse matrix in CSR
[out]	res	real_Double_t* Frobenius norm of difference
[in]	queue	magma_queue_t Queue to execute in.

magma_smatrix_addrowindex()

magma_int_t magma_smatrix_addrowindex	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Adds to a CSR matrix an array containing the rowindexes.

Parameters

[in,out]	A	magma_s_matrix* Matrix where rowindexes should be added.
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrcoo_transpose()

magma_int_t magma_scsrcoo_transpose	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Transposes a matrix that already contains rowidx.

The idea is to use a linked list.

Parameters

[in]	A	magma_s_matrix Matrix to transpose.
[out]	B	magma_s_matrix* Transposed matrix.
[in]	queue	magma_queue_t Queue to execute in.

magma_smatrix_createrowptr()

magma_int_t magma_smatrix_createrowptr	(	magma_int_t	n,
		magma_index_t *	row,
		magma_queue_t	queue )

This function generates a rowpointer out of a row-wise element count in parallel.

Parameters

[in]	n	magma_indnt_t row-count.
[in,out]	row	magma_index_t* Input: Vector of size n+1 containing the row-counts (offset by one). Output: Rowpointer.
[in]	queue	magma_queue_t Queue to execute in.

magma_smatrix_swap()

magma_int_t magma_smatrix_swap	(	magma_s_matrix *	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Swaps two matrices.

Useful if a loop modifies the name of a matrix.

Parameters

[in,out]	A	magma_s_matrix* Matrix to be swapped with B.
[in,out]	B	magma_s_matrix* Matrix to be swapped with A.
[in]	queue	magma_queue_t Queue to execute in.

magma_smatrix_tril()

magma_int_t magma_smatrix_tril	(	magma_s_matrix	A,
		magma_s_matrix *	L,
		magma_queue_t	queue )

Extracts the lower triangular of a matrix: L = tril(A).

The values of A are preserved.

Parameters

[in]	A	magma_s_matrix Element part of this.
[out]	L	magma_s_matrix* Lower triangular part of A.
[in]	queue	magma_queue_t Queue to execute in.

magma_smatrix_triu()

magma_int_t magma_smatrix_triu	(	magma_s_matrix	A,
		magma_s_matrix *	U,
		magma_queue_t	queue )

Extracts the lower triangular of a matrix: U = triu(A).

The values of A are preserved.

Parameters

[in]	A	magma_s_matrix Element part of this.
[out]	U	magma_s_matrix* Lower triangular part of A.
[in]	queue	magma_queue_t Queue to execute in.

magma_scsr_sort()

magma_int_t magma_scsr_sort	(	magma_s_matrix *	A,
		magma_queue_t	queue )

SOrts the elements in a CSR matrix for increasing column index.

Parameters

[in,out]	A	magma_s_matrix* CSR matrix, sorted on output.
[in]	queue	magma_queue_t Queue to execute in.

magma_srowentries()

magma_int_t magma_srowentries	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Checks the maximal number of nonzeros in a row of matrix A.

Inserts the data into max_nnz_row.

Parameters

[in,out]	A	magma_s_matrix* sparse matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_s_csr_compressor()

magma_int_t magma_s_csr_compressor	(	float **	val,
		magma_index_t **	row,
		magma_index_t **	col,
		float **	valn,
		magma_index_t **	rown,
		magma_index_t **	coln,
		magma_int_t *	n,
		magma_queue_t	queue )

Helper function to compress CSR containing zero-entries.

Parameters

[in]	val	float** input val pointer to compress
[in]	row	magma_int_t** input row pointer to modify
[in]	col	magma_int_t** input col pointer to compress
[in]	valn	float** output val pointer
[out]	rown	magma_int_t** output row pointer
[out]	coln	magma_int_t** output col pointer
[out]	n	magma_int_t* number of rows in matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smconvert()

magma_int_t magma_smconvert	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_storage_t	old_format,
		magma_storage_t	new_format,
		magma_queue_t	queue )

Converter between different sparse storage formats.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[out]	B	magma_s_matrix* copy of A in new format
[in]	old_format	magma_storage_t original storage format
[in]	new_format	magma_storage_t new storage format
[in]	queue	magma_queue_t Queue to execute in.

magma_smcsrcompressor()

magma_int_t magma_smcsrcompressor	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Removes zeros in a CSR matrix.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrset()

magma_int_t magma_scsrset	(	magma_int_t	m,
		magma_int_t	n,
		magma_index_t *	row,
		magma_index_t *	col,
		float *	val,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Passes a CSR matrix to MAGMA.

Parameters

[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	row	magma_index_t* row pointer
[in]	col	magma_index_t* column indices
[in]	val	float* array containing matrix entries
[out]	A	magma_s_matrix* matrix in magma sparse matrix format
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrget()

magma_int_t magma_scsrget	(	magma_s_matrix	A,
		magma_int_t *	m,
		magma_int_t *	n,
		magma_index_t **	row,
		magma_index_t **	col,
		float **	val,
		magma_queue_t	queue )

Passes a MAGMA matrix to CSR structure.

Parameters

[in]	A	magma_s_matrix magma sparse matrix in CSR format
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	row	magma_index_t* row pointer
[out]	col	magma_index_t* column indices
[out]	val	float* array containing matrix entries
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrset_gpu()

magma_int_t magma_scsrset_gpu	(	magma_int_t	m,
		magma_int_t	n,
		magmaIndex_ptr	row,
		magmaIndex_ptr	col,
		magmaFloat_ptr	val,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Passes a CSR matrix to MAGMA (located on DEV).

Parameters

[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	row	magmaIndex_ptr row pointer
[in]	col	magmaIndex_ptr column indices
[in]	val	magmaFloat_ptr array containing matrix entries
[out]	A	magma_s_matrix* matrix in magma sparse matrix format
[in]	queue	magma_queue_t Queue to execute in.

magma_scsrget_gpu()

magma_int_t magma_scsrget_gpu	(	magma_s_matrix	A,
		magma_int_t *	m,
		magma_int_t *	n,
		magmaIndex_ptr *	row,
		magmaIndex_ptr *	col,
		magmaFloat_ptr *	val,
		magma_queue_t	queue )

Passes a MAGMA matrix to CSR structure (located on DEV).

Parameters

[in]	A	magma_s_matrix magma sparse matrix in CSR format
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	row	magmaIndex_ptr row pointer
[out]	col	magmaIndex_ptr column indices
[out]	val	magmaFloat_ptr array containing matrix entries
[in]	queue	magma_queue_t Queue to execute in.

magma_smdiagdom()

magma_int_t magma_smdiagdom	(	magma_s_matrix	M,
		float *	min_dd,
		float *	max_dd,
		float *	avg_dd,
		magma_queue_t	queue )

This routine takes a CSR matrix and computes the average diagonal dominance.

For each row i, it computes the abs(d_ii)/sum_j(abs(a_ij)). It returns max, min, and average.

Parameters

[in]	M	magma_s_matrix System matrix.
[out]	min_dd	float Smallest diagonal dominance.
[out]	max_dd	float Largest diagonal dominance.
[out]	avg_dd	float Average diagonal dominance.
[in]	queue	magma_queue_t Queue to execute in.

magma_smbdiagdom()

magma_int_t magma_smbdiagdom	(	magma_s_matrix	M,
		magma_s_matrix	blocksizes,
		float *	min_dd,
		float *	max_dd,
		float *	avg_dd,
		magma_queue_t	queue )

This routine takes a CSR matrix and computes the average block-diagonal dominance.

For each row i, it computes the abs( D_(i,:) ) / abs( A(i,:) \ D_(i,:) ). It returns max, min, and average. The input vector bsz contains the blocksizes.

Parameters

[in]	M	magma_s_matrix System matrix.
[in]	blocksizes	magma_s_matrix Vector containing blocksizes (as DoubleComplex).
[out]	min_dd	float Smallest diagonal dominance.
[out]	max_dd	float Largest diagonal dominance.
[out]	avg_dd	float Average diagonal dominance.
[in]	queue	magma_queue_t Queue to execute in.

magma_smdiff()

magma_int_t magma_smdiff	(	magma_s_matrix	A,
		magma_s_matrix	B,
		real_Double_t *	res,
		magma_queue_t	queue )

Computes the Frobenius norm of the difference between the CSR matrices A and B.

They do not need to share the same sparsity pattern!

    res = ||A-B||_F = sqrt( sum_ij (A_ij-B_ij)^2 )

Parameters

[in]	A	magma_s_matrix sparse matrix in CSR
[in]	B	magma_s_matrix sparse matrix in CSR
[out]	res	real_Double_t* residual
[in]	queue	magma_queue_t Queue to execute in.

magma_smfrobenius()

magma_int_t magma_smfrobenius	(	magma_s_matrix	A,
		magma_s_matrix	B,
		magma_s_matrix	S,
		float *	norm,
		magma_queue_t	queue )

Computes the Frobenius norm || A - B ||_S on the sparsity pattern of S.

Parameters

[in]	A	magma_s_matrix input sparse matrix in CSR
[in]	B	magma_s_matrix input sparse matrix in CSR
[in]	S	magma_s_matrix input sparsity pattern in CSR
[out]	norm	float* Frobenius norm of difference on sparsity pattern S
[in]	queue	magma_queue_t Queue to execute in.

magma_smgenerator()

magma_int_t magma_smgenerator	(	magma_int_t	n,
		magma_int_t	offdiags,
		magma_index_t *	diag_offset,
		float *	diag_vals,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Generate a symmetric n x n CSR matrix for a stencil.

Parameters

[in]	n	magma_int_t number of rows
[in]	offdiags	magma_int_t number of offdiagonals
[in]	diag_offset	magma_int_t* array containing the offsets (length offsets+1)
[in]	diag_vals	float* array containing the values (length offsets+1)
[out]	A	magma_s_matrix* matrix to generate
[in]	queue	magma_queue_t Queue to execute in.

magma_sm_27stencil()

magma_int_t magma_sm_27stencil	(	magma_int_t	n,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Generate a 27-point stencil for a 3D FD discretization.

Parameters

[in]	n	magma_int_t number of rows
[out]	A	magma_s_matrix* matrix to generate
[in]	queue	magma_queue_t Queue to execute in.

magma_sm_5stencil()

magma_int_t magma_sm_5stencil	(	magma_int_t	n,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Generate a 5-point stencil for a 2D FD discretization.

Parameters

[in]	n	magma_int_t number of rows
[out]	A	magma_s_matrix* matrix to generate
[in]	queue	magma_queue_t Queue to execute in.

magma_ssymbilu()

magma_int_t magma_ssymbilu	(	magma_s_matrix *	A,
		magma_int_t	levels,
		magma_s_matrix *	L,
		magma_s_matrix *	U,
		magma_queue_t	queue )

This routine performs a symbolic ILU factorization.

The algorithm is taken from an implementation written by Edmond Chow.

Parameters

[in,out]	A	magma_s_matrix* matrix in magma sparse matrix format containing the original matrix on input, and L,U on output
[in]	levels	magma_magma_int_t_t fill in level
[out]	L	magma_s_matrix* output lower triangular matrix in magma sparse matrix format empty on function call
[out]	U	magma_s_matrix* output upper triangular matrix in magma sparse matrix format empty on function call
[in]	queue	magma_queue_t Queue to execute in.

read_s_csr_from_mtx()

magma_int_t read_s_csr_from_mtx	(	magma_storage_t *	type,
		magma_location_t *	location,
		magma_int_t *	n_row,
		magma_int_t *	n_col,
		magma_int_t *	nnz,
		float **	val,
		magma_index_t **	row,
		magma_index_t **	col,
		const char *	filename,
		magma_queue_t	queue )

Reads in a matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.

It duplicates the off-diagonal entries in the symmetric case.

Parameters

[out]	type	magma_storage_t* storage type of matrix
[out]	location	magma_location_t* location of matrix
[out]	n_row	magma_int_t* number of rows in matrix
[out]	n_col	magma_int_t* number of columns in matrix
[out]	nnz	magma_int_t* number of nonzeros in matrix
[out]	val	float** value array of CSR output
[out]	row	magma_index_t** row pointer of CSR output
[out]	col	magma_index_t** column indices of CSR output
[in]	filename	const char* filname of the mtx matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_swrite_csr_mtx()

magma_int_t magma_swrite_csr_mtx	(	magma_s_matrix	A,
		magma_order_t	MajorType,
		const char *	filename,
		magma_queue_t	queue )

Writes a CSR matrix to a file using Matrix Market format.

Parameters

[in]	A	magma_s_matrix matrix to write out
[in]	MajorType	magma_index_t Row or Column sort default: 0 = RowMajor, 1 = ColMajor TODO: use named constants (e.g., MagmaRowMajor), not numbers.
[in]	filename	const char* output-filname of the mtx matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_sprint_csr_mtx()

magma_int_t magma_sprint_csr_mtx	(	magma_int_t	n_row,
		magma_int_t	n_col,
		magma_int_t	nnz,
		float **	val,
		magma_index_t **	row,
		magma_index_t **	col,
		magma_order_t	MajorType,
		magma_queue_t	queue )

Prints a CSR matrix in Matrix Market format.

Parameters

[in]	n_row	magma_int_t* number of rows in matrix
[in]	n_col	magma_int_t* number of columns in matrix
[in]	nnz	magma_int_t* number of nonzeros in matrix
[in]	val	float** value array of CSR
[in]	row	magma_index_t** row pointer of CSR
[in]	col	magma_index_t** column indices of CSR
[in]	MajorType	magma_index_t Row or Column sort default: 0 = RowMajor, 1 = ColMajor
[in]	queue	magma_queue_t Queue to execute in.

magma_sprint_csr()

magma_int_t magma_sprint_csr	(	magma_int_t	n_row,
		magma_int_t	n_col,
		magma_int_t	nnz,
		float **	val,
		magma_index_t **	row,
		magma_index_t **	col,
		magma_queue_t	queue )

Prints a CSR matrix in CSR format.

Parameters

[in]	n_row	magma_int_t* number of rows in matrix
[in]	n_col	magma_int_t* number of columns in matrix
[in]	nnz	magma_int_t* number of nonzeros in matrix
[in]	val	float** value array of CSR
[in]	row	magma_index_t** row pointer of CSR
[in]	col	magma_index_t** column indices of CSR
[in]	queue	magma_queue_t Queue to execute in.

magma_sprint_matrix()

magma_int_t magma_sprint_matrix	(	magma_s_matrix	A,
		magma_queue_t	queue )

Prints a sparse matrix in CSR format.

Parameters

[in]	A	magma_s_matrix sparse matrix in Magma_CSR format
[in]	queue	magma_queue_t Queue to execute in.

magma_s_csr_mtx()

magma_int_t magma_s_csr_mtx	(	magma_s_matrix *	A,
		const char *	filename,
		magma_queue_t	queue )

Reads in a matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.

It duplicates the off-diagonal entries in the symmetric case.

Parameters

[out]	A	magma_s_matrix* matrix in magma sparse matrix format
[in]	filename	const char* filname of the mtx matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_s_csr_mtxsymm()

magma_int_t magma_s_csr_mtxsymm	(	magma_s_matrix *	A,
		const char *	filename,
		magma_queue_t	queue )

Reads in a SYMMETRIC matrix stored in coo format from a Matrix Market (.mtx) file and converts it into CSR format.

It does not duplicate the off-diagonal entries!

Parameters

[out]	A	magma_s_matrix* matrix in magma sparse matrix format
[in]	filename	const char* filname of the mtx matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smlumerge()

magma_int_t magma_smlumerge	(	magma_s_matrix	L,
		magma_s_matrix	U,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Takes an strictly lower triangular matrix L and an upper triangular matrix U and merges them into a matrix A containing the upper and lower triangular parts.

Parameters

[in]	L	magma_s_matrix input strictly lower triangular matrix L
[in]	U	magma_s_matrix input upper triangular matrix U
[out]	A	magma_s_matrix* output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smscale()

magma_int_t magma_smscale	(	magma_s_matrix *	A,
		magma_scale_t	scaling,
		magma_queue_t	queue )

Scales a matrix.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in]	scaling	magma_scale_t scaling type (unit rownorm / unit diagonal)
[in]	queue	magma_queue_t Queue to execute in.

magma_smscale_matrix_rhs()

magma_int_t magma_smscale_matrix_rhs	(	magma_s_matrix *	A,
		magma_s_matrix *	b,
		magma_s_matrix *	scaling_factors,
		magma_scale_t	scaling,
		magma_queue_t	queue )

Scales a matrix and a right hand side vector of a Ax = b system.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in,out]	b	magma_s_matrix* input/output right hand side vector
[out]	scaling_factors	magma_s_matrix* output scaling factors vector
[in]	scaling	magma_scale_t scaling type (unit rownorm / unit diagonal)
[in]	queue	magma_queue_t Queue to execute in.

magma_smdiagadd()

magma_int_t magma_smdiagadd	(	magma_s_matrix *	A,
		float	add,
		magma_queue_t	queue )

Adds a multiple of the Identity matrix to a matrix: A = A+add * I.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in]	add	float scaling for the identity matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smscale_generate()

magma_int_t magma_smscale_generate	(	magma_int_t	n,
		magma_scale_t *	scaling,
		magma_side_t *	side,
		magma_s_matrix *	A,
		magma_s_matrix *	scaling_factors,
		magma_queue_t	queue )

Generates n vectors of scaling factors from the A matrix and stores them in the factors matrix as column vectors in column major ordering.

Parameters

[in]	n	magma_int_t number of diagonal scaling matrices
[in]	scaling	magma_scale_t* array of scaling specifiers
[in]	side	magma_side_t* array of side specifiers
[in]	A	magma_s_matrix* input matrix
[out]	scaling_factors	magma_s_matrix* array of diagonal matrices
[in]	queue	magma_queue_t Queue to execute in.

magma_smscale_apply()

magma_int_t magma_smscale_apply	(	magma_int_t	n,
		magma_side_t *	side,
		magma_s_matrix *	scaling_factors,
		magma_s_matrix *	A,
		magma_queue_t	queue )

Applies n diagonal scaling matrices to a matrix A; n=[1,2], factor[i] is applied to side[i] of the matrix.

Parameters

[in]	n	magma_int_t number of diagonal scaling matrices
[in]	side	magma_side_t* array of side specifiers
[in]	scaling_factors	magma_s_matrix* array of diagonal matrices
[in,out]	A	magma_s_matrix* input/output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_sdimv()

magma_int_t magma_sdimv	(	magma_s_matrix *	vecA,
		magma_s_matrix *	vecB,
		magma_queue_t	queue )

Multiplies a diagonal matrix (vecA) and a vector (vecB).

Parameters

[in]	vecA	magma_s_matrix* input matrix
[in,out]	vecB	magma_s_matrix* input/output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smshrink()

magma_int_t magma_smshrink	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Shrinks a non-square matrix (m < n) to the smaller dimension.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[out]	B	magma_s_matrix* sparse matrix A
[in]	queue	magma_queue_t Queue to execute in.

magma_smslice()

magma_int_t magma_smslice	(	magma_int_t	num_slices,
		magma_int_t	slice,
		magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_s_matrix *	ALOC,
		magma_s_matrix *	ANLOC,
		magma_index_t *	comm_i,
		float *	comm_v,
		magma_int_t *	start,
		magma_int_t *	end,
		magma_queue_t	queue )

Takes a matrix and extracts a slice for solving the system in parallel:

B = A( i:i+n, : ) and ALOC = A(i:i+n,i:i+n) and ANLOCA(0:start - end:n,:)

B is of size n x n, ALOC of size end-start x end-start, ANLOC of size end-start x n

The last slice might be smaller. For the non-local parts, B is the identity. comm contains 1ess in the locations that are non-local but needed to solve local system.

Parameters

[in]	num_slices	magma_int_t number of slices
[in]	slice	magma_int_t slice id (0.. num_slices-1)
[in]	A	magma_s_matrix sparse matrix in CSR
[out]	B	magma_s_matrix* sparse matrix in CSR
[out]	ALOC	magma_s_matrix* sparse matrix in CSR
[out]	ANLOC	magma_s_matrix* sparse matrix in CSR
[in,out]	comm_i	magma_int_t* communication plan
[in,out]	comm_v	float* communication plan
[in]	queue	magma_queue_t Queue to execute in.
[out]	start	magma_int_t* start of slice (row-index)
[out]	end	magma_int_t* end of slice (row-index)

magma_smtransfer()

magma_int_t magma_smtransfer	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_location_t	src,
		magma_location_t	dst,
		magma_queue_t	queue )

Copies a matrix from memory location src to memory location dst.

Parameters

[in]	A	magma_s_matrix sparse matrix A
[out]	B	magma_s_matrix* copy of A
[in]	src	magma_location_t original location A
[in]	dst	magma_location_t location of the copy of A
[in]	queue	magma_queue_t Queue to execute in.

s_transpose_csr()

magma_int_t s_transpose_csr	(	magma_int_t	n_rows,
		magma_int_t	n_cols,
		magma_int_t	nnz,
		float *	values,
		magma_index_t *	rowptr,
		magma_index_t *	colind,
		magma_int_t *	new_n_rows,
		magma_int_t *	new_n_cols,
		magma_int_t *	new_nnz,
		float **	new_values,
		magma_index_t **	new_rowptr,
		magma_index_t **	new_colind,
		magma_queue_t	queue )

Transposes a matrix stored in CSR format on the CPU host.

Parameters

[in]	n_rows	magma_int_t number of rows in input matrix
[in]	n_cols	magma_int_t number of columns in input matrix
[in]	nnz	magma_int_t number of nonzeros in input matrix
[in]	values	float* value array of input matrix
[in]	rowptr	magma_index_t* row pointer of input matrix
[in]	colind	magma_index_t* column indices of input matrix
[in]	new_n_rows	magma_index_t* number of rows in transposed matrix
[in]	new_n_cols	magma_index_t* number of columns in transposed matrix
[in]	new_nnz	magma_index_t* number of nonzeros in transposed matrix
[in]	new_values	float** value array of transposed matrix
[in]	new_rowptr	magma_index_t** row pointer of transposed matrix
[in]	new_colind	magma_index_t** column indices of transposed matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smtranspose()

magma_int_t magma_smtranspose	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Interface to cuSPARSE transpose.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_s_cucsrtranspose()

magma_int_t magma_s_cucsrtranspose	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Helper function to transpose CSR matrix.

Using the CUSPARSE CSR2CSC function.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_smtransposeconjugate()

magma_int_t magma_smtransposeconjugate	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

This function forms the transpose conjugate of a matrix.

For a real-value matrix, the output is the transpose.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_smtranspose_cpu()

magma_int_t magma_smtranspose_cpu	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Generates a transpose of A on the CPU.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_smtransposeconj_cpu()

magma_int_t magma_smtransposeconj_cpu	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Generates a transpose conjugate of A on the CPU.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_smtransposestruct_cpu()

magma_int_t magma_smtransposestruct_cpu	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Generates a transpose of the nonzero pattern of A on the CPU.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_smtransposeabs_cpu()

magma_int_t magma_smtransposeabs_cpu	(	magma_s_matrix	A,
		magma_s_matrix *	B,
		magma_queue_t	queue )

Generates a transpose with absolute values of A on the CPU.

Parameters

[in]	A	magma_s_matrix input matrix (CSR)
[out]	B	magma_s_matrix* output matrix (CSR)
[in]	queue	magma_queue_t Queue to execute in.

magma_ssolverinfo()

magma_int_t magma_ssolverinfo	(	magma_s_solver_par *	solver_par,
		magma_s_preconditioner *	precond_par,
		magma_queue_t	queue )

Prints information about a previously called solver.

Parameters

[in]	solver_par	magma_s_solver_par* structure containing all solver information
[in,out]	precond_par	magma_s_preconditioner* structure containing all preconditioner information
[in]	queue	magma_queue_t Queue to execute in.

magma_ssolverinfo_free()

magma_int_t magma_ssolverinfo_free	(	magma_s_solver_par *	solver_par,
		magma_s_preconditioner *	precond_par,
		magma_queue_t	queue )

Frees any memory assocoiated with the verbose mode of solver_par.

The other values are set to default.

Parameters

[in,out]	solver_par	magma_s_solver_par* structure containing all solver information
[in,out]	precond_par	magma_s_preconditioner* structure containing all preconditioner information
[in]	queue	magma_queue_t Queue to execute in.

magma_ssolverinfo_init()

magma_int_t magma_ssolverinfo_init	(	magma_s_solver_par *	solver_par,
		magma_s_preconditioner *	precond_par,
		magma_queue_t	queue )

Initializes all solver and preconditioner parameters.

Parameters

[in,out]	solver_par	magma_s_solver_par* structure containing all solver information
[in,out]	precond_par	magma_s_preconditioner* structure containing all preconditioner information
[in]	queue	magma_queue_t Queue to execute in.

magma_seigensolverinfo_init()

magma_int_t magma_seigensolverinfo_init	(	magma_s_solver_par *	solver_par,
		magma_queue_t	queue )

Initializes space for eigensolvers.

Parameters

[in,out]	solver_par	magma_s_solver_par* structure containing all solver information
[in]	queue	magma_queue_t Queue to execute in.

magma_ssort()

magma_int_t magma_ssort	(	float *	x,
		magma_int_t	first,
		magma_int_t	last,
		magma_queue_t	queue )

Sorts an array of values in increasing order.

Parameters

[in,out]	x	float* array to sort
[in]	first	magma_int_t pointer to first element
[in]	last	magma_int_t pointer to last element
[in]	queue	magma_queue_t Queue to execute in.

magma_smsort()

magma_int_t magma_smsort	(	float *	x,
		magma_index_t *	col,
		magma_index_t *	row,
		magma_int_t	first,
		magma_int_t	last,
		magma_queue_t	queue )

Sorts an array of values in increasing order.

Parameters

[in,out]	x	float* array to sort
[in,out]	col	magma_index_t* Target array, will be modified during operation.
[in,out]	row	magma_index_t* Target array, will be modified during operation.
[in]	first	magma_int_t pointer to first element
[in]	last	magma_int_t pointer to last element
[in]	queue	magma_queue_t Queue to execute in.

magma_sindexsort()

magma_int_t magma_sindexsort	(	magma_index_t *	x,
		magma_int_t	first,
		magma_int_t	last,
		magma_queue_t	queue )

Sorts an array of integers in increasing order.

Parameters

[in,out]	x	magma_index_t* array to sort
[in]	first	magma_int_t pointer to first element
[in]	last	magma_int_t pointer to last element
[in]	queue	magma_queue_t Queue to execute in.

magma_sindexsortval()

magma_int_t magma_sindexsortval	(	magma_index_t *	x,
		float *	y,
		magma_int_t	first,
		magma_int_t	last,
		magma_queue_t	queue )

Sorts an array of integers, updates a respective array of values.

Parameters

[in,out]	x	magma_index_t* array to sort
[in,out]	y	float* array to sort
[in]	first	magma_int_t pointer to first element
[in]	last	magma_int_t pointer to last element
[in]	queue	magma_queue_t Queue to execute in.

magma_sorderstatistics()

magma_int_t magma_sorderstatistics	(	float *	val,
		magma_int_t	length,
		magma_int_t	k,
		magma_int_t	r,
		float *	element,
		magma_queue_t	queue )

Identifies the kth smallest/largest element in an array.

Parameters

[in,out]	val	float* Target array, will be modified during operation.
[in]	length	magma_int_t Length of the target array.
[in]	k	magma_int_t Element to be identified (largest/smallest).
[in]	r	magma_int_t rule how to sort: '1' -> largest, '0' -> smallest
[out]	element	float* location of the respective element
[in]	queue	magma_queue_t Queue to execute in.

magma_sbitonic_sort()

magma_int_t magma_sbitonic_sort	(	magma_int_t	start,
		magma_int_t	length,
		float *	seq,
		magma_int_t	flag,
		magma_queue_t	queue )

Approximates the k-th smallest element in an array by using order-statistics with step-size inc.

Parameters

[in]	start	magma_int_t Start position of the target array.
[in]	length	magma_int_t Length of the target array.
[in,out]	seq	float* Target array, will be modified during operation.
[in]	flag	magma_int_t ???
[in]	queue	magma_queue_t Queue to execute in.

magma_sparse_opts()

magma_int_t magma_sparse_opts	(	int	argc,
		char **	argv,
		magma_sopts *	opts,
		int *	matrices,
		magma_queue_t	queue )

Parses input options for a solver.

Parameters

[in]	argc	int command line input
[in]	argv	char** command line input
[in,out]	opts	magma_sopts * magma solver options
[out]	matrices	int counter how many linear systems to process
[in]	queue	magma_queue_t Queue to execute in.

magma_svinit()

magma_int_t magma_svinit	(	magma_s_matrix *	x,
		magma_location_t	mem_loc,
		magma_int_t	num_rows,
		magma_int_t	num_cols,
		float	values,
		magma_queue_t	queue )

Allocates memory for magma_s_matrix and initializes it with the passed value.

Parameters

[out]	x	magma_s_matrix* vector to initialize
[in]	mem_loc	magma_location_t memory for vector
[in]	num_rows	magma_int_t desired length of vector
[in]	num_cols	magma_int_t desired width of vector-block (columns of dense matrix)
[in]	values	float entries in vector
[in]	queue	magma_queue_t Queue to execute in.

magma_svinit_rand()

magma_int_t magma_svinit_rand	(	magma_s_matrix *	x,
		magma_location_t	mem_loc,
		magma_int_t	num_rows,
		magma_int_t	num_cols,
		magma_queue_t	queue )

Allocates memory for magma_s_matrix and initializes it with random values.

Parameters

[out]	x	magma_s_matrix* vector to initialize
[in]	mem_loc	magma_location_t memory for vector
[in]	num_rows	magma_int_t desired length of vector
[in]	num_cols	magma_int_t desired width of vector-block (columns of dense matrix)
[in]	queue	magma_queue_t Queue to execute in.

magma_sprint_vector()

magma_int_t magma_sprint_vector	(	magma_s_matrix	x,
		magma_int_t	offset,
		magma_int_t	visulen,
		magma_queue_t	queue )

Visualizes part of a vector of type magma_s_matrix.

With input vector x , offset, visulen, the entries offset - (offset + visulen) of x are visualized.

Parameters

[in]	x	magma_s_matrix vector to visualize
[in]	offset	magma_int_t start inex of visualization
[in]	visulen	magma_int_t number of entries to visualize
[in]	queue	magma_queue_t Queue to execute in.

magma_svread()

magma_int_t magma_svread	(	magma_s_matrix *	x,
		magma_int_t	length,
		char *	filename,
		magma_queue_t	queue )

Reads in a float vector of length "length".

Parameters

[out]	x	magma_s_matrix * vector to read in
[in]	length	magma_int_t length of vector
[in]	filename	char* file where vector is stored
[in]	queue	magma_queue_t Queue to execute in.

magma_swrite_vector()

magma_int_t magma_swrite_vector	(	magma_s_matrix	A,
		const char *	filename,
		magma_queue_t	queue )

Writes a vector to a file.

Parameters

[in]	A	magma_s_matrix matrix to write out
[in]	filename	const char* output-filname of the mtx matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_svset()

magma_int_t magma_svset	(	magma_int_t	m,
		magma_int_t	n,
		float *	val,
		magma_s_matrix *	v,
		magma_queue_t	queue )

Passes a vector to MAGMA.

Parameters

[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	val	float* array containing vector entries
[out]	v	magma_s_matrix* magma vector
[in]	queue	magma_queue_t Queue to execute in.

magma_svcopy()

magma_int_t magma_svcopy	(	magma_s_matrix	v,
		magma_int_t *	m,
		magma_int_t *	n,
		float *	val,
		magma_queue_t	queue )

Passes a MAGMA vector back.

This function requires the array val to be already allocated (of size m x n).

Parameters

[in]	v	magma_s_matrix magma vector
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	val	float* array of size m x n the vector entries are copied into
[in]	queue	magma_queue_t Queue to execute in.

magma_svset_dev()

magma_int_t magma_svset_dev	(	magma_int_t	m,
		magma_int_t	n,
		magmaFloat_ptr	val,
		magma_s_matrix *	v,
		magma_queue_t	queue )

Passes a vector to MAGMA (located on DEV).

Parameters

[in]	m	magma_int_t number of rows
[in]	n	magma_int_t number of columns
[in]	val	float* array containing vector entries
[out]	v	magma_s_matrix* magma vector
[in]	queue	magma_queue_t Queue to execute in.

magma_svget()

magma_int_t magma_svget	(	magma_s_matrix	v,
		magma_int_t *	m,
		magma_int_t *	n,
		float **	val,
		magma_queue_t	queue )

Passes a MAGMA vector back.

Parameters

[in]	v	magma_s_matrix magma vector
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	val	float* array containing vector entries
[in]	queue	magma_queue_t Queue to execute in.

magma_svget_dev()

magma_int_t magma_svget_dev	(	magma_s_matrix	v,
		magma_int_t *	m,
		magma_int_t *	n,
		magmaFloat_ptr *	val,
		magma_queue_t	queue )

Passes a MAGMA vector back (located on DEV).

Parameters

[in]	v	magma_s_matrix magma vector
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	val	magmaFloat_ptr array containing vector entries
[in]	queue	magma_queue_t Queue to execute in.

magma_svcopy_dev()

magma_int_t magma_svcopy_dev	(	magma_s_matrix	v,
		magma_int_t *	m,
		magma_int_t *	n,
		magmaFloat_ptr	val,
		magma_queue_t	queue )

Passes a MAGMA vector back (located on DEV).

This function requires the array val to be already allocated (of size m x n).

Parameters

[in]	v	magma_s_matrix magma vector
[out]	m	magma_int_t number of rows
[out]	n	magma_int_t number of columns
[out]	val	float* array of size m x n on the device the vector entries are copied into
[in]	queue	magma_queue_t Queue to execute in.

magma_svtranspose()

magma_int_t magma_svtranspose	(	magma_s_matrix	x,
		magma_s_matrix *	y,
		magma_queue_t	queue )

Transposes a vector from col to row major and vice versa.

Parameters

[in]	x	magma_s_matrix input vector
[out]	y	magma_s_matrix* output vector
[in]	queue	magma_queue_t Queue to execute in.

magma_sdiagcheck()

magma_int_t magma_sdiagcheck	(	magma_s_matrix	dA,
		magma_queue_t	queue )

This routine checks for a CSR matrix whether there exists a zero on the diagonal.

This can be the diagonal entry missing or an explicit zero.

Parameters

[in]	dA	magma_s_matrix matrix in CSR format
[in]	queue	magma_queue_t Queue to execute in.

magma_scsr_sort_gpu()

magma_int_t magma_scsr_sort_gpu	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Generates a matrix \(U = A \cup B\).

If both matrices have a nonzero value in the same location, the value of A is used.

This is the GPU version of the operation.

Parameters

[in]	A	magma_s_matrix Input matrix 1.
[in]	B	magma_s_matrix Input matrix 2.
[out]	U	magma_s_matrix* \(U = A \cup B\). If both matrices have a nonzero value in the same location, the value of A is used.
[in]	queue	magma_queue_t Queue to execute in.

magma_smconjugate()

magma_int_t magma_smconjugate	(	magma_s_matrix *	A,
		magma_queue_t	queue )

This function conjugates a matrix.

For a real matrix, no value is changed.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_smcsrcompressor_gpu()

magma_int_t magma_smcsrcompressor_gpu	(	magma_s_matrix *	A,
		magma_queue_t	queue )

Removes zeros in a CSR matrix.

This is a GPU implementation of the CSR compressor.

Parameters

[in,out]	A	magma_s_matrix* input/output matrix
[in]	queue	magma_queue_t Queue to execute in.

magma_slobpcg_res()

magma_int_t magma_slobpcg_res	(	magma_int_t	num_rows,
		magma_int_t	num_vecs,
		magmaFloat_ptr	evalues,
		magmaFloat_ptr	X,
		magmaFloat_ptr	R,
		magmaFloat_ptr	res,
		magma_queue_t	queue )

This routine computes for Block-LOBPCG, the set of residuals.

R = Ax - x evalues It replaces: for(int i=0; i < n; i++) { magma_saxpy(m, MAGMA_S_MAKE(-evalues[i],0),blockX+i*m,1,blockR+i*m,1); } The memory layout of x is:

/ x1[0] x2[0] x3[0] \
| x1[1] x2[1] x3[1] |

x = | x1[2] x2[2] x3[2] | = x1[0] x1[1] x1[2] x1[3] x1[4] x2[0] x2[1] . | x1[3] x2[3] x3[3] | \ x1[4] x2[4] x3[4] /

Parameters

[in]	num_rows	magma_int_t number of rows
[in]	num_vecs	magma_int_t number of vectors
[in]	evalues	magmaFloat_ptr array of eigenvalues/approximations
[in]	X	magmaFloat_ptr block of eigenvector approximations
[in]	R	magmaFloat_ptr block of residuals
[in]	res	magmaFloat_ptr array of residuals
[in]	queue	magma_queue_t Queue to execute in.

magma_slobpcg_shift()

magma_int_t magma_slobpcg_shift	(	magma_int_t	num_rows,
		magma_int_t	num_vecs,
		magma_int_t	shift,
		magmaFloat_ptr	x,
		magma_queue_t	queue )

For a Block-LOBPCG, the set of residuals (entries consecutive in memory)
shrinks and the vectors are shifted in case shift residuals drop below threshold.

The memory layout of x is:

/ x1[0] x2[0] x3[0] \
| x1[1] x2[1] x3[1] |

x = | x1[2] x2[2] x3[2] | = x1[0] x2[0] x3[0] x1[1] x2[1] x3[1] x1[2] . | x1[3] x2[3] x3[3] | \ x1[4] x2[4] x3[4] /

Parameters

[in]	num_rows	magma_int_t number of rows
[in]	num_vecs	magma_int_t number of vectors
[in]	shift	magma_int_t shift number
[in,out]	x	magmaFloat_ptr input/output vector x
[in]	queue	magma_queue_t Queue to execute in.