Sparse Extra Functions#

This module holds all sparse extra routines.

The sparse extra routines describe operations that manipulate sparse matrices.

hipsparseXcsrgeamNnz()#

hipsparseStatus_t hipsparseXcsrgeamNnz(hipsparseHandle_t handle, int m, int n, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, int *csrRowPtrC, int *nnzTotalDevHostPtr)#

Sparse matrix sparse matrix addition using CSR storage format.

hipsparseXcsrgeamNnz computes the total CSR non-zero elements and the CSR row offsets, that point to the start of every row of the sparse CSR matrix, of the resulting matrix C. It is assumed that csr_row_ptr_C has been allocated with size m + 1.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgeam()#

hipsparseStatus_t hipsparseScsrgeam(hipsparseHandle_t handle, int m, int n, const float *alpha, const hipsparseMatDescr_t descrA, int nnzA, const float *csrValA, const int *csrRowPtrA, const int *csrColIndA, const float *beta, const hipsparseMatDescr_t descrB, int nnzB, const float *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, float *csrValC, int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseDcsrgeam(hipsparseHandle_t handle, int m, int n, const double *alpha, const hipsparseMatDescr_t descrA, int nnzA, const double *csrValA, const int *csrRowPtrA, const int *csrColIndA, const double *beta, const hipsparseMatDescr_t descrB, int nnzB, const double *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, double *csrValC, int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseCcsrgeam(hipsparseHandle_t handle, int m, int n, const hipComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, hipComplex *csrValC, int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseZcsrgeam(hipsparseHandle_t handle, int m, int n, const hipDoubleComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipDoubleComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipDoubleComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipDoubleComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, hipDoubleComplex *csrValC, int *csrRowPtrC, int *csrColIndC)#

Sparse matrix sparse matrix addition using CSR storage format.

hipsparseXcsrgeam multiplies the scalar \(\alpha\) with the sparse \(m \times n\) matrix \(A\), defined in CSR storage format, multiplies the scalar \(\beta\) with the sparse \(m \times n\) matrix \(B\), defined in CSR storage format, and adds both resulting matrices to obtain the sparse \(m \times n\) matrix \(C\), defined in CSR storage format, such that

\[ C := \alpha \cdot A + \beta \cdot B. \]

It is assumed that csr_row_ptr_C has already been filled and that csr_val_C and csr_col_ind_C are allocated by the user. csr_row_ptr_C and allocation size of csr_col_ind_C and csr_val_C is defined by the number of non-zero elements of the sparse CSR matrix C. Both can be obtained by hipsparseXcsrgeamNnz().

Note

Both scalars \(\alpha\) and \(beta\) have to be valid.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

hipsparseXcsrgeam2_bufferSizeExt()#

hipsparseStatus_t hipsparseScsrgeam2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, const float *alpha, const hipsparseMatDescr_t descrA, int nnzA, const float *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const float *beta, const hipsparseMatDescr_t descrB, int nnzB, const float *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, const float *csrSortedValC, const int *csrSortedRowPtrC, const int *csrSortedColIndC, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseDcsrgeam2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, const double *alpha, const hipsparseMatDescr_t descrA, int nnzA, const double *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const double *beta, const hipsparseMatDescr_t descrB, int nnzB, const double *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, const double *csrSortedValC, const int *csrSortedRowPtrC, const int *csrSortedColIndC, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseCcsrgeam2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, const hipComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipComplex *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const hipComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipComplex *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, const hipComplex *csrSortedValC, const int *csrSortedRowPtrC, const int *csrSortedColIndC, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseZcsrgeam2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, const hipDoubleComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipDoubleComplex *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const hipDoubleComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipDoubleComplex *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, const hipDoubleComplex *csrSortedValC, const int *csrSortedRowPtrC, const int *csrSortedColIndC, size_t *pBufferSizeInBytes)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgeam2_bufferSizeExt returns the size of the temporary storage buffer that is required by hipsparseXcsrgeam2Nnz() and hipsparseXcsrgeam2(). The temporary storage buffer must be allocated by the user.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgeam2Nnz()#

hipsparseStatus_t hipsparseXcsrgeam2Nnz(hipsparseHandle_t handle, int m, int n, const hipsparseMatDescr_t descrA, int nnzA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, int *csrSortedRowPtrC, int *nnzTotalDevHostPtr, void *workspace)#

Sparse matrix sparse matrix addition using CSR storage format.

hipsparseXcsrgeam2Nnz computes the total CSR non-zero elements and the CSR row offsets, that point to the start of every row of the sparse CSR matrix, of the resulting matrix C. It is assumed that csr_row_ptr_C has been allocated with size m + 1.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgeam2()#

hipsparseStatus_t hipsparseScsrgeam2(hipsparseHandle_t handle, int m, int n, const float *alpha, const hipsparseMatDescr_t descrA, int nnzA, const float *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const float *beta, const hipsparseMatDescr_t descrB, int nnzB, const float *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, float *csrSortedValC, int *csrSortedRowPtrC, int *csrSortedColIndC, void *pBuffer)#
hipsparseStatus_t hipsparseDcsrgeam2(hipsparseHandle_t handle, int m, int n, const double *alpha, const hipsparseMatDescr_t descrA, int nnzA, const double *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const double *beta, const hipsparseMatDescr_t descrB, int nnzB, const double *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, double *csrSortedValC, int *csrSortedRowPtrC, int *csrSortedColIndC, void *pBuffer)#
hipsparseStatus_t hipsparseCcsrgeam2(hipsparseHandle_t handle, int m, int n, const hipComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipComplex *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const hipComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipComplex *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, hipComplex *csrSortedValC, int *csrSortedRowPtrC, int *csrSortedColIndC, void *pBuffer)#
hipsparseStatus_t hipsparseZcsrgeam2(hipsparseHandle_t handle, int m, int n, const hipDoubleComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipDoubleComplex *csrSortedValA, const int *csrSortedRowPtrA, const int *csrSortedColIndA, const hipDoubleComplex *beta, const hipsparseMatDescr_t descrB, int nnzB, const hipDoubleComplex *csrSortedValB, const int *csrSortedRowPtrB, const int *csrSortedColIndB, const hipsparseMatDescr_t descrC, hipDoubleComplex *csrSortedValC, int *csrSortedRowPtrC, int *csrSortedColIndC, void *pBuffer)#

Sparse matrix sparse matrix addition using CSR storage format.

hipsparseXcsrgeam2 multiplies the scalar \(\alpha\) with the sparse \(m \times n\) matrix \(A\), defined in CSR storage format, multiplies the scalar \(\beta\) with the sparse \(m \times n\) matrix \(B\), defined in CSR storage format, and adds both resulting matrices to obtain the sparse \(m \times n\) matrix \(C\), defined in CSR storage format, such that

\[ C := \alpha \cdot A + \beta \cdot B. \]

It is assumed that csr_row_ptr_C has already been filled and that csr_val_C and csr_col_ind_C are allocated by the user. csr_row_ptr_C and allocation size of csr_col_ind_C and csr_val_C is defined by the number of non-zero elements of the sparse CSR matrix C. Both can be obtained by hipsparseXcsrgeam2Nnz().

Note

Both scalars \(\alpha\) and \(beta\) have to be valid.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

hipsparseXcsrgemmNnz()#

hipsparseStatus_t hipsparseXcsrgemmNnz(hipsparseHandle_t handle, hipsparseOperation_t transA, hipsparseOperation_t transB, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, int *csrRowPtrC, int *nnzTotalDevHostPtr)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgemmNnz computes the total CSR non-zero elements and the CSR row offsets, that point to the start of every row of the sparse CSR matrix, of the resulting multiplied matrix C. It is assumed that csr_row_ptr_C has been allocated with size m + 1.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Please note, that for matrix products with more than 8192 intermediate products per row, additional temporary storage buffer is allocated by the algorithm.

Note

Currently, only trans_A == trans_B == HIPSPARSE_OPERATION_NON_TRANSPOSE is supported.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgemm()#

hipsparseStatus_t hipsparseScsrgemm(hipsparseHandle_t handle, hipsparseOperation_t transA, hipsparseOperation_t transB, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const float *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const float *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, float *csrValC, const int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseDcsrgemm(hipsparseHandle_t handle, hipsparseOperation_t transA, hipsparseOperation_t transB, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const double *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const double *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, double *csrValC, const int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseCcsrgemm(hipsparseHandle_t handle, hipsparseOperation_t transA, hipsparseOperation_t transB, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const hipComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const hipComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, hipComplex *csrValC, const int *csrRowPtrC, int *csrColIndC)#
hipsparseStatus_t hipsparseZcsrgemm(hipsparseHandle_t handle, hipsparseOperation_t transA, hipsparseOperation_t transB, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const hipDoubleComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const hipDoubleComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrC, hipDoubleComplex *csrValC, const int *csrRowPtrC, int *csrColIndC)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgemm multiplies the sparse \(m \times k\) matrix \(A\), defined in CSR storage format with the sparse \(k \times n\) matrix \(B\), defined in CSR storage format, and stores the result in the sparse \(m \times n\) matrix \(C\), defined in CSR storage format, such that

\[ C := op(A) \cdot op(B), \]
with
\[\begin{split} op(A) = \left\{ \begin{array}{ll} A, & \text{if trans_A == HIPSPARSE_OPERATION_NON_TRANSPOSE} \\ A^T, & \text{if trans_A == HIPSPARSE_OPERATION_TRANSPOSE} \\ A^H, & \text{if trans_A == HIPSPARSE_OPERATION_CONJUGATE_TRANSPOSE} \end{array} \right. \end{split}\]
and
\[\begin{split} op(B) = \left\{ \begin{array}{ll} B, & \text{if trans_B == HIPSPARSE_OPERATION_NON_TRANSPOSE} \\ B^T, & \text{if trans_B == HIPSPARSE_OPERATION_TRANSPOSE} \\ B^H, & \text{if trans_B == HIPSPARSE_OPERATION_CONJUGATE_TRANSPOSE} \end{array} \right. \end{split}\]

It is assumed that csr_row_ptr_C has already been filled and that csr_val_C and csr_col_ind_C are allocated by the user. csr_row_ptr_C and allocation size of csr_col_ind_C and csr_val_C is defined by the number of non-zero elements of the sparse CSR matrix C. Both can be obtained by hipsparseXcsrgemmNnz().

Note

Currently, only trans_A == HIPSPARSE_OPERATION_NON_TRANSPOSE is supported.

Note

Currently, only trans_B == HIPSPARSE_OPERATION_NON_TRANSPOSE is supported.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Please note, that for matrix products with more than 4096 non-zero entries per row, additional temporary storage buffer is allocated by the algorithm.

hipsparseXcsrgemm2_bufferSizeExt()#

hipsparseStatus_t hipsparseScsrgemm2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, int k, const float *alpha, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const float *beta, const hipsparseMatDescr_t descrD, int nnzD, const int *csrRowPtrD, const int *csrColIndD, csrgemm2Info_t info, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseDcsrgemm2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, int k, const double *alpha, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const double *beta, const hipsparseMatDescr_t descrD, int nnzD, const int *csrRowPtrD, const int *csrColIndD, csrgemm2Info_t info, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseCcsrgemm2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, int k, const hipComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const hipComplex *beta, const hipsparseMatDescr_t descrD, int nnzD, const int *csrRowPtrD, const int *csrColIndD, csrgemm2Info_t info, size_t *pBufferSizeInBytes)#
hipsparseStatus_t hipsparseZcsrgemm2_bufferSizeExt(hipsparseHandle_t handle, int m, int n, int k, const hipDoubleComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const hipDoubleComplex *beta, const hipsparseMatDescr_t descrD, int nnzD, const int *csrRowPtrD, const int *csrColIndD, csrgemm2Info_t info, size_t *pBufferSizeInBytes)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgemm2_bufferSizeExt returns the size of the temporary storage buffer that is required by hipsparseXcsrgemm2Nnz() and hipsparseXcsrgemm2(). The temporary storage buffer must be allocated by the user.

Note

Please note, that for matrix products with more than 4096 non-zero entries per row, additional temporary storage buffer is allocated by the algorithm.

Note

Please note, that for matrix products with more than 8192 intermediate products per row, additional temporary storage buffer is allocated by the algorithm.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgemm2Nnz()#

hipsparseStatus_t hipsparseXcsrgemm2Nnz(hipsparseHandle_t handle, int m, int n, int k, const hipsparseMatDescr_t descrA, int nnzA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const int *csrRowPtrB, const int *csrColIndB, const hipsparseMatDescr_t descrD, int nnzD, const int *csrRowPtrD, const int *csrColIndD, const hipsparseMatDescr_t descrC, int *csrRowPtrC, int *nnzTotalDevHostPtr, const csrgemm2Info_t info, void *pBuffer)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgemm2Nnz computes the total CSR non-zero elements and the CSR row offsets, that point to the start of every row of the sparse CSR matrix, of the resulting multiplied matrix C. It is assumed that csr_row_ptr_C has been allocated with size m + 1. The required buffer size can be obtained by hipsparseXcsrgemm2_bufferSizeExt().

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Please note, that for matrix products with more than 8192 intermediate products per row, additional temporary storage buffer is allocated by the algorithm.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

hipsparseXcsrgemm2()#

hipsparseStatus_t hipsparseScsrgemm2(hipsparseHandle_t handle, int m, int n, int k, const float *alpha, const hipsparseMatDescr_t descrA, int nnzA, const float *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const float *csrValB, const int *csrRowPtrB, const int *csrColIndB, const float *beta, const hipsparseMatDescr_t descrD, int nnzD, const float *csrValD, const int *csrRowPtrD, const int *csrColIndD, const hipsparseMatDescr_t descrC, float *csrValC, const int *csrRowPtrC, int *csrColIndC, const csrgemm2Info_t info, void *pBuffer)#
hipsparseStatus_t hipsparseDcsrgemm2(hipsparseHandle_t handle, int m, int n, int k, const double *alpha, const hipsparseMatDescr_t descrA, int nnzA, const double *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const double *csrValB, const int *csrRowPtrB, const int *csrColIndB, const double *beta, const hipsparseMatDescr_t descrD, int nnzD, const double *csrValD, const int *csrRowPtrD, const int *csrColIndD, const hipsparseMatDescr_t descrC, double *csrValC, const int *csrRowPtrC, int *csrColIndC, const csrgemm2Info_t info, void *pBuffer)#
hipsparseStatus_t hipsparseCcsrgemm2(hipsparseHandle_t handle, int m, int n, int k, const hipComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const hipComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipComplex *beta, const hipsparseMatDescr_t descrD, int nnzD, const hipComplex *csrValD, const int *csrRowPtrD, const int *csrColIndD, const hipsparseMatDescr_t descrC, hipComplex *csrValC, const int *csrRowPtrC, int *csrColIndC, const csrgemm2Info_t info, void *pBuffer)#
hipsparseStatus_t hipsparseZcsrgemm2(hipsparseHandle_t handle, int m, int n, int k, const hipDoubleComplex *alpha, const hipsparseMatDescr_t descrA, int nnzA, const hipDoubleComplex *csrValA, const int *csrRowPtrA, const int *csrColIndA, const hipsparseMatDescr_t descrB, int nnzB, const hipDoubleComplex *csrValB, const int *csrRowPtrB, const int *csrColIndB, const hipDoubleComplex *beta, const hipsparseMatDescr_t descrD, int nnzD, const hipDoubleComplex *csrValD, const int *csrRowPtrD, const int *csrColIndD, const hipsparseMatDescr_t descrC, hipDoubleComplex *csrValC, const int *csrRowPtrC, int *csrColIndC, const csrgemm2Info_t info, void *pBuffer)#

Sparse matrix sparse matrix multiplication using CSR storage format.

hipsparseXcsrgemm2 multiplies the scalar \(\alpha\) with the sparse \(m \times k\) matrix \(A\), defined in CSR storage format, and the sparse \(k \times n\) matrix \(B\), defined in CSR storage format, and adds the result to the sparse \(m \times n\) matrix \(D\) that is multiplied by \(\beta\). The final result is stored in the sparse \(m \times n\) matrix \(C\), defined in CSR storage format, such that

\[ C := \alpha \cdot A \cdot B + \beta \cdot D \]

It is assumed that csr_row_ptr_C has already been filled and that csr_val_C and csr_col_ind_C are allocated by the user. csr_row_ptr_C and allocation size of csr_col_ind_C and csr_val_C is defined by the number of non-zero elements of the sparse CSR matrix C. Both can be obtained by hipsparseXcsrgemm2Nnz(). The required buffer size for the computation can be obtained by hipsparseXcsrgemm2_bufferSizeExt().

Note

If \(\alpha == 0\), then \(C = \beta \cdot D\) will be computed.

Note

If \(\beta == 0\), then \(C = \alpha \cdot A \cdot B\) will be computed.

Note

\(\alpha == beta == 0\) is invalid.

Note

Currently, only HIPSPARSE_MATRIX_TYPE_GENERAL is supported.

Note

This function is non blocking and executed asynchronously with respect to the host. It may return before the actual computation has finished.

Note

Please note, that for matrix products with more than 4096 non-zero entries per row, additional temporary storage buffer is allocated by the algorithm.