![]() |
MAGMA
1.5.0
Matrix Algebra for GPU and Multicore Architectures
|
Functions | |
magma_int_t | magma_dgecsrmv (magma_trans_t transA, magma_int_t m, magma_int_t n, double alpha, double *d_val, magma_index_t *d_rowptr, magma_index_t *d_colind, double *d_x, double beta, double *d_y) |
This routine computes y = alpha * A * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgecsrmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, double alpha, double lambda, double *d_val, magma_index_t *d_rowptr, magma_index_t *d_colind, double *d_x, double beta, int offset, int blocksize, magma_index_t *add_rows, double *d_y) |
This routine computes y = alpha * ( A -lambda I ) * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgeellmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, double alpha, double *d_val, magma_index_t *d_colind, double *d_x, double beta, double *d_y) |
This routine computes y = alpha * A * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgeellmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, double alpha, double lambda, double *d_val, magma_index_t *d_colind, double *d_x, double beta, int offset, int blocksize, magma_index_t *add_rows, double *d_y) |
This routine computes y = alpha *( A - lambda I ) * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgeellrtmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, double alpha, double *d_val, magma_index_t *d_colind, magma_index_t *d_rowlength, double *d_x, double beta, double *d_y, magma_int_t alignment, magma_int_t blocksize) |
This routine computes y = alpha * A * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgeelltmv_shift (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t nnz_per_row, double alpha, double lambda, double *d_val, magma_index_t *d_colind, double *d_x, double beta, int offset, int blocksize, magma_index_t *add_rows, double *d_y) |
This routine computes y = alpha *( A - lambda I ) * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgesellpmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t blocksize, magma_int_t slices, magma_int_t alignment, double alpha, double *d_val, magma_index_t *d_colind, magma_index_t *d_rowptr, double *d_x, double beta, double *d_y) |
This routine computes y = alpha * A^t * x + beta * y on the GPU. More... | |
magma_int_t | magma_dgesellcmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t blocksize, magma_int_t slices, magma_int_t alignment, double alpha, double *d_val, magma_index_t *d_colind, magma_index_t *d_rowptr, double *d_x, double beta, double *d_y) |
This routine computes y = alpha * A^t * x + beta * y on the GPU. More... | |
magma_int_t | magma_dmdotc (int n, int k, double *v, double *r, double *d1, double *d2, double *skp) |
Computes the scalar product of a set of vectors v_i such that. More... | |
magma_int_t | magma_dmgecsrmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, double alpha, double *d_val, magma_index_t *d_rowptr, magma_index_t *d_colind, double *d_x, double beta, double *d_y) |
This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU. More... | |
magma_int_t | magma_dmgeellmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, magma_int_t nnz_per_row, double alpha, double *d_val, magma_index_t *d_colind, double *d_x, double beta, double *d_y) |
This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU. More... | |
magma_int_t | magma_dmgeelltmv (magma_trans_t transA, magma_int_t m, magma_int_t n, magma_int_t num_vecs, magma_int_t nnz_per_row, double alpha, double *d_val, magma_index_t *d_colind, double *d_x, double beta, double *d_y) |
This routine computes Y = alpha * A * X + beta * Y for X and Y sets of num_vec vectors on the GPU. More... | |
magma_int_t | magma_dmgesellpmv (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, double alpha, double *d_val, magma_index_t *d_colind, magma_index_t *d_rowptr, double *d_x, double beta, double *d_y) |
This routine computes Y = alpha * A^t * X + beta * Y on the GPU. More... | |
magma_int_t magma_dgecsrmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_rowptr, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
This routine computes y = alpha * A * x + beta * y on the GPU.
The input format is CSR (val, row, col).
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
alpha | double scalar multiplier |
d_val | double* array containing values of A in CSR |
d_rowptr | magma_int_t* rowpointer of A in CSR |
d_colind | magma_int_t* columnindices of A in CSR |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dgecsrmv_shift | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
double | alpha, | ||
double | lambda, | ||
double * | d_val, | ||
magma_index_t * | d_rowptr, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
int | offset, | ||
int | blocksize, | ||
magma_index_t * | add_rows, | ||
double * | d_y | ||
) |
This routine computes y = alpha * ( A -lambda I ) * x + beta * y on the GPU.
It is a shifted version of the CSR-SpMV.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
alpha | double scalar multiplier |
lambda | double scalar multiplier |
d_val | double* array containing values of A in CSR |
d_rowptr | magma_int_t* rowpointer of A in CSR |
d_colind | magma_int_t* columnindices of A in CSR |
d_x | double* input vector x |
beta | double scalar multiplier |
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 |
d_y | double* output vector y |
magma_int_t magma_dgeellmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
This routine computes y = alpha * A * x + beta * y on the GPU.
Input format is ELLPACK.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
nnz_per_row | magma_int_t number of elements in the longest row |
alpha | double scalar multiplier |
d_val | double* array containing values of A in ELLPACK |
d_colind | magma_int_t* columnindices of A in ELLPACK |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dgeellmv_shift | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double | lambda, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
int | offset, | ||
int | blocksize, | ||
magma_index_t * | add_rows, | ||
double * | d_y | ||
) |
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.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
nnz_per_row | magma_int_t number of elements in the longest row |
alpha | double scalar multiplier |
lambda | double scalar multiplier |
d_val | double* array containing values of A in ELLPACK |
d_colind | magma_int_t* columnindices of A in ELLPACK |
d_x | double* input vector x |
beta | double scalar multiplier |
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 |
d_y | double* input/output vector y |
magma_int_t magma_dgeellrtmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
magma_index_t * | d_rowlength, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y, | ||
magma_int_t | alignment, | ||
magma_int_t | blocksize | ||
) |
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.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows |
n | magma_int_t number of columns |
nnz_per_row | magma_int_t max number of nonzeros in a row |
alpha | double scalar alpha |
d_val | double* val array |
d_colind | magma_int_t* col indices |
d_rowlength | magma_int_t* number of elements in each row |
d_x | double* input vector x |
beta | double scalar beta |
d_y | double* output vector y |
blocksize | magma_int_t threads per block |
alignment | magma_int_t threads assigned to each row |
magma_int_t magma_dgeelltmv_shift | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double | lambda, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
int | offset, | ||
int | blocksize, | ||
magma_index_t * | add_rows, | ||
double * | d_y | ||
) |
This routine computes y = alpha *( A - lambda I ) * x + beta * y on the GPU.
Input format is ELL.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
nnz_per_row | magma_int_t number of elements in the longest row |
alpha | double scalar multiplier |
lambda | double scalar multiplier |
d_val | double* array containing values of A in ELL |
d_colind | magma_int_t* columnindices of A in ELL |
d_x | double* input vector x |
beta | double scalar multiplier |
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 |
d_y | double* input/output vector y |
magma_int_t magma_dgesellcmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | blocksize, | ||
magma_int_t | slices, | ||
magma_int_t | alignment, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
magma_index_t * | d_rowptr, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
This routine computes y = alpha * A^t * x + beta * y on the GPU.
Input format is SELLC/SELLP.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
blocksize | magma_int_t number of rows in one ELL-slice |
slices | magma_int_t number of slices in matrix |
alignment | magma_int_t number of threads assigned to one row (=1) |
alpha | double scalar multiplier |
d_val | double* array containing values of A in SELLC/P |
d_colind | magma_int_t* columnindices of A in SELLC/P |
d_rowptr | magma_int_t* rowpointer of SELLP |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dgesellpmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | blocksize, | ||
magma_int_t | slices, | ||
magma_int_t | alignment, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
magma_index_t * | d_rowptr, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
This routine computes y = alpha * A^t * x + beta * y on the GPU.
Input format is SELLP.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
blocksize | magma_int_t number of rows in one ELL-slice |
slices | magma_int_t number of slices in matrix |
alignment | magma_int_t number of threads assigned to one row |
alpha | double scalar multiplier |
d_val | double* array containing values of A in SELLP |
d_colind | magma_int_t* columnindices of A in SELLP |
d_rowptr | magma_int_t* rowpointer of SELLP |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dmdotc | ( | int | n, |
int | k, | ||
double * | v, | ||
double * | r, | ||
double * | d1, | ||
double * | d2, | ||
double * | skp | ||
) |
Computes the scalar product of a set of vectors v_i such that.
skp = ( <v_0,r>, <v_1,r>, .. )
Returns the vector skp.
n | int length of v_i and r |
k | int vectors v_i |
v | double* v = (v_0 .. v_i.. v_k) |
r | double* r |
d1 | double* workspace |
d2 | double* workspace |
skp | double* vector[k] of scalar products (<v_i,r>...) |
magma_int_t magma_dmgecsrmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | num_vecs, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_rowptr, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
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.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
num_vecs | mama_int_t number of vectors |
alpha | double scalar multiplier |
d_val | double* array containing values of A in CSR |
d_rowptr | magma_int_t* rowpointer of A in CSR |
d_colind | magma_int_t* columnindices of A in CSR |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dmgeellmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | num_vecs, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
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.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
num_vecs | mama_int_t number of vectors |
nnz_per_row | magma_int_t number of elements in the longest row |
alpha | double scalar multiplier |
d_val | double* array containing values of A in ELLPACK |
d_colind | magma_int_t* columnindices of A in ELLPACK |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dmgeelltmv | ( | magma_trans_t | transA, |
magma_int_t | m, | ||
magma_int_t | n, | ||
magma_int_t | num_vecs, | ||
magma_int_t | nnz_per_row, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
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.
transA | magma_trans_t transposition parameter for A |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
num_vecs | mama_int_t number of vectors |
nnz_per_row | magma_int_t number of elements in the longest row |
alpha | double scalar multiplier |
d_val | double* array containing values of A in ELL |
d_colind | magma_int_t* columnindices of A in ELL |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |
magma_int_t magma_dmgesellpmv | ( | 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, | ||
double | alpha, | ||
double * | d_val, | ||
magma_index_t * | d_colind, | ||
magma_index_t * | d_rowptr, | ||
double * | d_x, | ||
double | beta, | ||
double * | d_y | ||
) |
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!
transA | magma_trans_t transpose A? |
m | magma_int_t number of rows in A |
n | magma_int_t number of columns in A |
num_vecs | magma_int_t number of columns in X and Y |
blocksize | magma_int_t number of rows in one ELL-slice |
slices | magma_int_t number of slices in matrix |
alignment | magma_int_t number of threads assigned to one row |
alpha | double scalar multiplier |
d_val | double* array containing values of A in SELLP |
d_colind | magma_int_t* columnindices of A in SELLP |
d_rowptr | magma_int_t* rowpointer of SELLP |
d_x | double* input vector x |
beta | double scalar multiplier |
d_y | double* input/output vector y |