SparseSolve(_:_:_:_:)
Solves the system AX=B for X, using the supplied SparseOpaqueFactorization_Complex_Float of A, and without any internal memory allocations.
Declaration
func SparseSolve(_ Factored: SparseOpaqueFactorization_Complex_Float, _ B: DenseMatrix_Complex_Float, _ X: DenseMatrix_Complex_Float, _ workspace: UnsafeMutableRawPointer)Parameters
- Factored:
A factorization of
A. - B:
The right-hand sides
Bto solve for. IfAhas dimensionm x n, thenBmust have dimensionm x nrhs, where nrhs is the number of right-hand sides to find solutions for. - X:
Matrix in which to return solutions. If
Ahas dimensionm x n, andBhas dimensionm x nrhs, thenXmust have dimensionn x nrhs. - workspace:
Scratch space of size
Factored.solveWorkspaceRequiredStatic + nrhs * Factored.solveWorkspaceRequiredPerRHS. This memory must be 16-byte aligned (any allocation returned bymalloc()has this property).
Discussion
If the factorization is A=QR and the system is underdetermined, the solution of minimum norm || x ||_2 is returned. If the factorization is A=QR and the system is overdetermined, the least squares solution arg min_x || Ax - b ||_2 is returned. In the case of a factorization of type=SparseCholeskyAtA, the factorization is in fact of A^T A, so the solution returned is for the system A^TAx=b.