Computes data corresponding to the specified decomposition (defined by export operation) and fills the pointers provided by calls to
This routine computes the data for the pointers of the (distributed) data to be exported (as defined the specified
operation). It is assumed that the exported data will be distributed in the same way as the matrix
A(as defined by
iparm(40)) used in
Only call this routine after the factorization phase (
) of the
cluster_sparse_solverhas been called. Neither
iparmshould be changed after the preceding call to
Only call this routine after all pointers to the data required for the specified operation have been provided by means of calling
- INTEGER*8for 64-bit architecturesArray with size of 64.Handle to internal data structure used in the prior calls tocluster_sparse_solver.Do not modifyptafter the calls tocluster_sparse_solver.
- INTEGERSpecifies a particular operation which defines what data are exported
Currently, foroperation=SPARSE_DPTLUQTa real (complex) unit vector is provided for the scaling matrixD. Do not turn on scaling(iparm(11)>0) or matching(iparm(13)>0) in theiparmduring the call tocluster_sparse_solverfor this value ofoperation.
- Exporting data from decomposition.P*A*Q=L*U
- Exporting data from decomposition from.P* (D-1A)*Q=L*U
- INTEGER*4MPI communicator. The solver uses the Fortran MPI communicator internally.
- INTEGERThe error indicator:
- no error
- ptis a null pointer
- pointers to some of the data required for the specifiedoperationwere not provided prior to callingcluster_sparse_solver_export
- unsupported internal code path, consider switching off non-defaultiparmparameters forcluster_sparse_solver
- unsupported case when the matrixAis distributed among processes with overlap in the preceding calls tocluster_sparse_solver
- internal memory error
cl_solver_export_f90.f90for an example using this functionality.