Eigen-Contrib  5.0.1
 
Loading...
Searching...
No Matches
Eigen::gpu::SparseLDLT< Scalar_, UpLo_ > Class Template Reference

#include <contrib/Eigen/src/GPU/GpuSparseLDLT.h>

Detailed Description

template<typename Scalar_, int UpLo_ = Lower>
class Eigen::gpu::SparseLDLT< Scalar_, UpLo_ >

GPU sparse LDL^T factorization (symmetric indefinite / Hermitian indefinite).

Wraps cuDSS with CUDSS_MTYPE_SYMMETRIC (real) or CUDSS_MTYPE_HERMITIAN (complex). Uses pivoting for numerical stability.

Template Parameters
Scalar_float, double, complex<float>, or complex<double>
UpLo_Lower (default) or Upper — which triangle of A is stored
+ Inheritance diagram for Eigen::gpu::SparseLDLT< Scalar_, UpLo_ >:

Public Member Functions

 SparseLDLT (Context &ctx)
 
template<typename InputType>
 SparseLDLT (Context &ctx, const SparseMatrixBase< InputType > &A)
 
- Public Member Functions inherited from Eigen::gpu::internal::SparseSolverBase< Scalar_, SparseLDLT< Scalar_, Lower > >
SparseLDLT< Scalar_, Lower > & analyzePattern (const SparseMatrixBase< InputType > &A)
 
SparseLDLT< Scalar_, Lower > & compute (const SparseMatrixBase< InputType > &A)
 
const SparseSolverConfig & config () const
 
SparseLDLT< Scalar_, Lower > & factorize (const SparseMatrixBase< InputType > &A)
 
SparseLDLT< Scalar_, Lower > & setConfig (const SparseSolverConfig &cfg)
 
DeviceMatrix< Scalar > solve (const DeviceMatrix< Scalar > &d_B) const
 
DenseMatrix solve (const MatrixBase< Rhs > &B) const
 
 SparseSolverBase (Context &ctx)
 

Constructor & Destructor Documentation

◆ SparseLDLT() [1/2]

template<typename Scalar_, int UpLo_ = Lower>
Eigen::gpu::SparseLDLT< Scalar_, UpLo_ >::SparseLDLT ( Context & ctx)
inlineexplicit

Borrow ctx's stream so solves chain with other work on the same Context. ctx must outlive this object.

◆ SparseLDLT() [2/2]

template<typename Scalar_, int UpLo_ = Lower>
template<typename InputType>
Eigen::gpu::SparseLDLT< Scalar_, UpLo_ >::SparseLDLT ( Context & ctx,
const SparseMatrixBase< InputType > & A )
inline

Borrow ctx's stream and factor A immediately.


The documentation for this class was generated from the following file: