@smake/eigen 1.1.0 → 1.1.1

package/eigen/Eigen/src/IterativeLinearSolvers/BiCGSTAB.h

@@ -11,202 +11,207 @@
 #ifndef EIGEN_BICGSTAB_H
 #define EIGEN_BICGSTAB_H
 
-namespace Eigen { 
+// IWYU pragma: private
+#include "./InternalHeaderCheck.h"
+
+namespace Eigen {
 
 namespace internal {
 
 /** \internal Low-level bi conjugate gradient stabilized algorithm
-  * \param mat The matrix A
-  * \param rhs The right hand side vector b
-  * \param x On input and initial solution, on output the computed solution.
-  * \param precond A preconditioner being able to efficiently solve for an
-  *                approximation of Ax=b (regardless of b)
-  * \param iters On input the max number of iteration, on output the number of performed iterations.
-  * \param tol_error On input the tolerance error, on output an estimation of the relative error.
-  * \return false in the case of numerical issue, for example a break down of BiCGSTAB. 
-  */
-template<typename MatrixType, typename Rhs, typename Dest, typename Preconditioner>
-bool bicgstab(const MatrixType& mat, const Rhs& rhs, Dest& x,
-              const Preconditioner& precond, Index& iters,
-              typename Dest::RealScalar& tol_error)
-{
-  using std::sqrt;
-  using std::abs;
+ * \param mat The matrix A
+ * \param rhs The right hand side vector b
+ * \param x On input and initial solution, on output the computed solution.
+ * \param precond A preconditioner being able to efficiently solve for an
+ *                approximation of Ax=b (regardless of b)
+ * \param iters On input the max number of iteration, on output the number of performed iterations.
+ * \param tol_error On input the tolerance error, on output an estimation of the relative error.
+ * \return false in the case of numerical issue, for example a break down of BiCGSTAB.
+ */
+template <typename MatrixType, typename Rhs, typename Dest, typename Preconditioner>
+bool bicgstab(const MatrixType& mat, const Rhs& rhs, Dest& x, const Preconditioner& precond, Index& iters,
+              typename Dest::RealScalar& tol_error) {
   typedef typename Dest::RealScalar RealScalar;
   typedef typename Dest::Scalar Scalar;
-  typedef Matrix<Scalar,Dynamic,1> VectorType;
+  typedef Matrix<Scalar, Dynamic, 1> VectorType;
   RealScalar tol = tol_error;
   Index maxIters = iters;
 
   Index n = mat.cols();
-  VectorType r  = rhs - mat * x;
+  VectorType r = rhs - mat * x;
   VectorType r0 = r;
-  
-  RealScalar r0_sqnorm = r0.squaredNorm();
-  RealScalar rhs_sqnorm = rhs.squaredNorm();
-  if(rhs_sqnorm == 0)
-  {
+
+  RealScalar r0_norm = r0.stableNorm();
+  RealScalar r_norm = r0_norm;
+  RealScalar rhs_norm = rhs.stableNorm();
+  if (rhs_norm == 0) {
     x.setZero();
     return true;
   }
-  Scalar rho    = 1;
-  Scalar alpha  = 1;
-  Scalar w      = 1;
-  
+  Scalar rho(1);
+  Scalar alpha(0);
+  Scalar w(1);
+
   VectorType v = VectorType::Zero(n), p = VectorType::Zero(n);
-  VectorType y(n),  z(n);
+  VectorType y(n), z(n);
   VectorType kt(n), ks(n);
 
   VectorType s(n), t(n);
 
-  RealScalar tol2 = tol*tol*rhs_sqnorm;
-  RealScalar eps2 = NumTraits<Scalar>::epsilon()*NumTraits<Scalar>::epsilon();
+  RealScalar eps = NumTraits<Scalar>::epsilon();
   Index i = 0;
   Index restarts = 0;
 
-  while ( r.squaredNorm() > tol2 && i<maxIters )
-  {
+  while (r_norm > tol && i < maxIters) {
     Scalar rho_old = rho;
-
     rho = r0.dot(r);
-    if (abs(rho) < eps2*r0_sqnorm)
-    {
+    if (Eigen::numext::abs(rho) / Eigen::numext::maxi(r0_norm, r_norm) < eps * Eigen::numext::mini(r0_norm, r_norm)) {
       // The new residual vector became too orthogonal to the arbitrarily chosen direction r0
       // Let's restart with a new r0:
-      r  = rhs - mat * x;
+      r = rhs - mat * x;
       r0 = r;
-      rho = r0_sqnorm = r.squaredNorm();
-      if(restarts++ == 0)
-        i = 0;
+      rho = r.squaredNorm();
+      r0_norm = r.stableNorm();
+      alpha = Scalar(0);
+      w = Scalar(1);
+      if (restarts++ == 0) i = 0;
     }
-    Scalar beta = (rho/rho_old) * (alpha / w);
+    Scalar beta = (rho / rho_old) * (alpha / w);
     p = r + beta * (p - w * v);
-    
+
     y = precond.solve(p);
-    
-    v.noalias() = mat * y;
 
-    alpha = rho / r0.dot(v);
+    v.noalias() = mat * y;
+    Scalar theta = r0.dot(v);
+    // For small angles ∠(r0, v) < eps, random restart.
+    RealScalar v_norm = v.stableNorm();
+    if (Eigen::numext::abs(theta) / Eigen::numext::maxi(r0_norm, v_norm) < eps * Eigen::numext::mini(r0_norm, v_norm)) {
+      r = rhs - mat * x;
+      r0.setRandom();
+      r0_norm = r0.stableNorm();
+      rho = Scalar(1);
+      alpha = Scalar(0);
+      w = Scalar(1);
+      if (restarts++ == 0) i = 0;
+      continue;
+    }
+    alpha = rho / theta;
     s = r - alpha * v;
 
     z = precond.solve(s);
     t.noalias() = mat * z;
 
     RealScalar tmp = t.squaredNorm();
-    if(tmp>RealScalar(0))
+    if (tmp > RealScalar(0)) {
       w = t.dot(s) / tmp;
-    else
+    } else {
       w = Scalar(0);
+    }
     x += alpha * y + w * z;
     r = s - w * t;
+    r_norm = r.stableNorm();
     ++i;
   }
-  tol_error = sqrt(r.squaredNorm()/rhs_sqnorm);
+
+  tol_error = r_norm / rhs_norm;
   iters = i;
-  return true; 
+  return true;
 }
 
-}
+}  // namespace internal
 
-template< typename _MatrixType,
-          typename _Preconditioner = DiagonalPreconditioner<typename _MatrixType::Scalar> >
+template <typename MatrixType_, typename Preconditioner_ = DiagonalPreconditioner<typename MatrixType_::Scalar> >
 class BiCGSTAB;
 
 namespace internal {
 
-template< typename _MatrixType, typename _Preconditioner>
-struct traits<BiCGSTAB<_MatrixType,_Preconditioner> >
-{
-  typedef _MatrixType MatrixType;
-  typedef _Preconditioner Preconditioner;
+template <typename MatrixType_, typename Preconditioner_>
+struct traits<BiCGSTAB<MatrixType_, Preconditioner_> > {
+  typedef MatrixType_ MatrixType;
+  typedef Preconditioner_ Preconditioner;
 };
 
-}
+}  // namespace internal
 
 /** \ingroup IterativeLinearSolvers_Module
-  * \brief A bi conjugate gradient stabilized solver for sparse square problems
-  *
-  * This class allows to solve for A.x = b sparse linear problems using a bi conjugate gradient
-  * stabilized algorithm. The vectors x and b can be either dense or sparse.
-  *
-  * \tparam _MatrixType the type of the sparse matrix A, can be a dense or a sparse matrix.
-  * \tparam _Preconditioner the type of the preconditioner. Default is DiagonalPreconditioner
-  *
-  * \implsparsesolverconcept
-  *
-  * The maximal number of iterations and tolerance value can be controlled via the setMaxIterations()
-  * and setTolerance() methods. The defaults are the size of the problem for the maximal number of iterations
-  * and NumTraits<Scalar>::epsilon() for the tolerance.
-  * 
-  * The tolerance corresponds to the relative residual error: |Ax-b|/|b|
-  * 
-  * \b Performance: when using sparse matrices, best performance is achied for a row-major sparse matrix format.
-  * Moreover, in this case multi-threading can be exploited if the user code is compiled with OpenMP enabled.
-  * See \ref TopicMultiThreading for details.
-  * 
-  * This class can be used as the direct solver classes. Here is a typical usage example:
-  * \include BiCGSTAB_simple.cpp
-  * 
-  * By default the iterations start with x=0 as an initial guess of the solution.
-  * One can control the start using the solveWithGuess() method.
-  * 
-  * BiCGSTAB can also be used in a matrix-free context, see the following \link MatrixfreeSolverExample example \endlink.
-  *
-  * \sa class SimplicialCholesky, DiagonalPreconditioner, IdentityPreconditioner
-  */
-template< typename _MatrixType, typename _Preconditioner>
-class BiCGSTAB : public IterativeSolverBase<BiCGSTAB<_MatrixType,_Preconditioner> >
-{
+ * \brief A bi conjugate gradient stabilized solver for sparse square problems
+ *
+ * This class allows to solve for A.x = b sparse linear problems using a bi conjugate gradient
+ * stabilized algorithm. The vectors x and b can be either dense or sparse.
+ *
+ * \tparam MatrixType_ the type of the sparse matrix A, can be a dense or a sparse matrix.
+ * \tparam Preconditioner_ the type of the preconditioner. Default is DiagonalPreconditioner
+ *
+ * \implsparsesolverconcept
+ *
+ * The maximal number of iterations and tolerance value can be controlled via the setMaxIterations()
+ * and setTolerance() methods. The defaults are the size of the problem for the maximal number of iterations
+ * and NumTraits<Scalar>::epsilon() for the tolerance.
+ *
+ * The tolerance corresponds to the relative residual error: |Ax-b|/|b|
+ *
+ * \b Performance: when using sparse matrices, best performance is achied for a row-major sparse matrix format.
+ * Moreover, in this case multi-threading can be exploited if the user code is compiled with OpenMP enabled.
+ * See \ref TopicMultiThreading for details.
+ *
+ * This class can be used as the direct solver classes. Here is a typical usage example:
+ * \include BiCGSTAB_simple.cpp
+ *
+ * By default the iterations start with x=0 as an initial guess of the solution.
+ * One can control the start using the solveWithGuess() method.
+ *
+ * BiCGSTAB can also be used in a matrix-free context, see the following \link MatrixfreeSolverExample example \endlink.
+ *
+ * \sa class SimplicialCholesky, DiagonalPreconditioner, IdentityPreconditioner
+ */
+template <typename MatrixType_, typename Preconditioner_>
+class BiCGSTAB : public IterativeSolverBase<BiCGSTAB<MatrixType_, Preconditioner_> > {
   typedef IterativeSolverBase<BiCGSTAB> Base;
-  using Base::matrix;
   using Base::m_error;
-  using Base::m_iterations;
   using Base::m_info;
   using Base::m_isInitialized;
-public:
-  typedef _MatrixType MatrixType;
+  using Base::m_iterations;
+  using Base::matrix;
+
+ public:
+  typedef MatrixType_ MatrixType;
   typedef typename MatrixType::Scalar Scalar;
   typedef typename MatrixType::RealScalar RealScalar;
-  typedef _Preconditioner Preconditioner;
-
-public:
+  typedef Preconditioner_ Preconditioner;
 
+ public:
   /** Default constructor. */
   BiCGSTAB() : Base() {}
 
   /** Initialize the solver with matrix \a A for further \c Ax=b solving.
-    * 
-    * This constructor is a shortcut for the default constructor followed
-    * by a call to compute().
-    * 
-    * \warning this class stores a reference to the matrix A as well as some
-    * precomputed values that depend on it. Therefore, if \a A is changed
-    * this class becomes invalid. Call compute() to update it with the new
-    * matrix A, or modify a copy of A.
-    */
-  template<typename MatrixDerived>
+   *
+   * This constructor is a shortcut for the default constructor followed
+   * by a call to compute().
+   *
+   * \warning this class stores a reference to the matrix A as well as some
+   * precomputed values that depend on it. Therefore, if \a A is changed
+   * this class becomes invalid. Call compute() to update it with the new
+   * matrix A, or modify a copy of A.
+   */
+  template <typename MatrixDerived>
   explicit BiCGSTAB(const EigenBase<MatrixDerived>& A) : Base(A.derived()) {}
 
   ~BiCGSTAB() {}
 
   /** \internal */
-  template<typename Rhs,typename Dest>
-  void _solve_vector_with_guess_impl(const Rhs& b, Dest& x) const
-  {    
+  template <typename Rhs, typename Dest>
+  void _solve_vector_with_guess_impl(const Rhs& b, Dest& x) const {
     m_iterations = Base::maxIterations();
     m_error = Base::m_tolerance;
-    
+
     bool ret = internal::bicgstab(matrix(), b, x, Base::m_preconditioner, m_iterations, m_error);
 
-    m_info = (!ret) ? NumericalIssue
-           : m_error <= Base::m_tolerance ? Success
-           : NoConvergence;
+    m_info = (!ret) ? NumericalIssue : m_error <= Base::m_tolerance ? Success : NoConvergence;
   }
 
-protected:
-
+ protected:
 };
 
-} // end namespace Eigen
+}  // end namespace Eigen
 
-#endif // EIGEN_BICGSTAB_H
+#endif  // EIGEN_BICGSTAB_H

package/eigen/Eigen/src/IterativeLinearSolvers/ConjugateGradient.h

@@ -10,102 +10,95 @@
 #ifndef EIGEN_CONJUGATE_GRADIENT_H
 #define EIGEN_CONJUGATE_GRADIENT_H
 
-namespace Eigen { 
+// IWYU pragma: private
+#include "./InternalHeaderCheck.h"
+
+namespace Eigen {
 
 namespace internal {
 
 /** \internal Low-level conjugate gradient algorithm
-  * \param mat The matrix A
-  * \param rhs The right hand side vector b
-  * \param x On input and initial solution, on output the computed solution.
-  * \param precond A preconditioner being able to efficiently solve for an
-  *                approximation of Ax=b (regardless of b)
-  * \param iters On input the max number of iteration, on output the number of performed iterations.
-  * \param tol_error On input the tolerance error, on output an estimation of the relative error.
-  */
-template<typename MatrixType, typename Rhs, typename Dest, typename Preconditioner>
-EIGEN_DONT_INLINE
-void conjugate_gradient(const MatrixType& mat, const Rhs& rhs, Dest& x,
-                        const Preconditioner& precond, Index& iters,
-                        typename Dest::RealScalar& tol_error)
-{
-  using std::sqrt;
-  using std::abs;
+ * \param mat The matrix A
+ * \param rhs The right hand side vector b
+ * \param x On input and initial solution, on output the computed solution.
+ * \param precond A preconditioner being able to efficiently solve for an
+ *                approximation of Ax=b (regardless of b)
+ * \param iters On input the max number of iteration, on output the number of performed iterations.
+ * \param tol_error On input the tolerance error, on output an estimation of the relative error.
+ */
+template <typename MatrixType, typename Rhs, typename Dest, typename Preconditioner>
+EIGEN_DONT_INLINE void conjugate_gradient(const MatrixType& mat, const Rhs& rhs, Dest& x, const Preconditioner& precond,
+                                          Index& iters, typename Dest::RealScalar& tol_error) {
   typedef typename Dest::RealScalar RealScalar;
   typedef typename Dest::Scalar Scalar;
-  typedef Matrix<Scalar,Dynamic,1> VectorType;
-  
+  typedef Matrix<Scalar, Dynamic, 1> VectorType;
+
   RealScalar tol = tol_error;
   Index maxIters = iters;
-  
+
   Index n = mat.cols();
 
-  VectorType residual = rhs - mat * x; //initial residual
+  VectorType residual = rhs - mat * x;  // initial residual
 
   RealScalar rhsNorm2 = rhs.squaredNorm();
-  if(rhsNorm2 == 0) 
-  {
+  if (rhsNorm2 == 0) {
     x.setZero();
     iters = 0;
     tol_error = 0;
     return;
   }
   const RealScalar considerAsZero = (std::numeric_limits<RealScalar>::min)();
-  RealScalar threshold = numext::maxi(RealScalar(tol*tol*rhsNorm2),considerAsZero);
+  RealScalar threshold = numext::maxi(RealScalar(tol * tol * rhsNorm2), considerAsZero);
   RealScalar residualNorm2 = residual.squaredNorm();
-  if (residualNorm2 < threshold)
-  {
+  if (residualNorm2 < threshold) {
     iters = 0;
-    tol_error = sqrt(residualNorm2 / rhsNorm2);
+    tol_error = numext::sqrt(residualNorm2 / rhsNorm2);
     return;
   }
 
   VectorType p(n);
-  p = precond.solve(residual);      // initial search direction
+  p = precond.solve(residual);  // initial search direction
 
   VectorType z(n), tmp(n);
   RealScalar absNew = numext::real(residual.dot(p));  // the square of the absolute value of r scaled by invM
   Index i = 0;
-  while(i < maxIters)
-  {
-    tmp.noalias() = mat * p;                    // the bottleneck of the algorithm
-
-    Scalar alpha = absNew / p.dot(tmp);         // the amount we travel on dir
-    x += alpha * p;                             // update solution
-    residual -= alpha * tmp;                    // update residual
-    
+  while (i < maxIters) {
+    tmp.noalias() = mat * p;  // the bottleneck of the algorithm
+
+    Scalar alpha = absNew / p.dot(tmp);  // the amount we travel on dir
+    x += alpha * p;                      // update solution
+    residual -= alpha * tmp;             // update residual
+
     residualNorm2 = residual.squaredNorm();
-    if(residualNorm2 < threshold)
-      break;
-    
-    z = precond.solve(residual);                // approximately solve for "A z = residual"
+    if (residualNorm2 < threshold) break;
+
+    z = precond.solve(residual);  // approximately solve for "A z = residual"
 
     RealScalar absOld = absNew;
-    absNew = numext::real(residual.dot(z));     // update the absolute value of r
-    RealScalar beta = absNew / absOld;          // calculate the Gram-Schmidt value used to create the new search direction
-    p = z + beta * p;                           // update search direction
+    absNew = numext::real(residual.dot(z));  // update the absolute value of r
+    RealScalar beta = absNew / absOld;       // calculate the Gram-Schmidt value used to create the new search direction
+    p = z + beta * p;                        // update search direction
     i++;
   }
-  tol_error = sqrt(residualNorm2 / rhsNorm2);
+  tol_error = numext::sqrt(residualNorm2 / rhsNorm2);
   iters = i;
 }
 
-}
+}  // namespace internal
 
-template< typename _MatrixType, int _UpLo=Lower,
-          typename _Preconditioner = DiagonalPreconditioner<typename _MatrixType::Scalar> >
+template <typename MatrixType_, int UpLo_ = Lower,
+          typename Preconditioner_ = DiagonalPreconditioner<typename MatrixType_::Scalar> >
 class ConjugateGradient;
 
 namespace internal {
 
-template< typename _MatrixType, int _UpLo, typename _Preconditioner>
-struct traits<ConjugateGradient<_MatrixType,_UpLo,_Preconditioner> >
-{
-  typedef _MatrixType MatrixType;
-  typedef _Preconditioner Preconditioner;
+template <typename MatrixType_, int UpLo_, typename Preconditioner_>
+struct traits<ConjugateGradient<MatrixType_, UpLo_, Preconditioner_> > {
+  typedef MatrixType_ MatrixType;
+  typedef Preconditioner_ Preconditioner;
 };
 
-}
+}  // namespace internal
 
 /** \ingroup IterativeLinearSolvers_Module
   * \brief A conjugate gradient solver for sparse (or dense) self-adjoint problems
@@ -113,25 +106,25 @@ struct traits<ConjugateGradient<_MatrixType,_UpLo,_Preconditioner> >
   * This class allows to solve for A.x = b linear problems using an iterative conjugate gradient algorithm.
   * The matrix A must be selfadjoint. The matrix A and the vectors x and b can be either dense or sparse.
   *
-  * \tparam _MatrixType the type of the matrix A, can be a dense or a sparse matrix.
-  * \tparam _UpLo the triangular part that will be used for the computations. It can be Lower,
+  * \tparam MatrixType_ the type of the matrix A, can be a dense or a sparse matrix.
+  * \tparam UpLo_ the triangular part that will be used for the computations. It can be Lower,
   *               \c Upper, or \c Lower|Upper in which the full matrix entries will be considered.
   *               Default is \c Lower, best performance is \c Lower|Upper.
-  * \tparam _Preconditioner the type of the preconditioner. Default is DiagonalPreconditioner
+  * \tparam Preconditioner_ the type of the preconditioner. Default is DiagonalPreconditioner
   *
   * \implsparsesolverconcept
   *
   * The maximal number of iterations and tolerance value can be controlled via the setMaxIterations()
   * and setTolerance() methods. The defaults are the size of the problem for the maximal number of iterations
   * and NumTraits<Scalar>::epsilon() for the tolerance.
-  * 
+  *
   * The tolerance corresponds to the relative residual error: |Ax-b|/|b|
-  * 
-  * \b Performance: Even though the default value of \c _UpLo is \c Lower, significantly higher performance is
-  * achieved when using a complete matrix and \b Lower|Upper as the \a _UpLo template parameter. Moreover, in this
+  *
+  * \b Performance: Even though the default value of \c UpLo_ is \c Lower, significantly higher performance is
+  * achieved when using a complete matrix and \b Lower|Upper as the \a UpLo_ template parameter. Moreover, in this
   * case multi-threading can be exploited if the user code is compiled with OpenMP enabled.
   * See \ref TopicMultiThreading for details.
-  * 
+  *
   * This class can be used as the direct solver classes. Here is a typical usage example:
     \code
     int n = 10000;
@@ -146,71 +139,67 @@ struct traits<ConjugateGradient<_MatrixType,_UpLo,_Preconditioner> >
     // update b, and solve again
     x = cg.solve(b);
     \endcode
-  * 
+  *
   * By default the iterations start with x=0 as an initial guess of the solution.
   * One can control the start using the solveWithGuess() method.
-  * 
-  * ConjugateGradient can also be used in a matrix-free context, see the following \link MatrixfreeSolverExample example \endlink.
+  *
+  * ConjugateGradient can also be used in a matrix-free context, see the following \link MatrixfreeSolverExample example
+  \endlink.
   *
   * \sa class LeastSquaresConjugateGradient, class SimplicialCholesky, DiagonalPreconditioner, IdentityPreconditioner
   */
-template< typename _MatrixType, int _UpLo, typename _Preconditioner>
-class ConjugateGradient : public IterativeSolverBase<ConjugateGradient<_MatrixType,_UpLo,_Preconditioner> >
-{
+template <typename MatrixType_, int UpLo_, typename Preconditioner_>
+class ConjugateGradient : public IterativeSolverBase<ConjugateGradient<MatrixType_, UpLo_, Preconditioner_> > {
   typedef IterativeSolverBase<ConjugateGradient> Base;
-  using Base::matrix;
   using Base::m_error;
-  using Base::m_iterations;
   using Base::m_info;
   using Base::m_isInitialized;
-public:
-  typedef _MatrixType MatrixType;
+  using Base::m_iterations;
+  using Base::matrix;
+
+ public:
+  typedef MatrixType_ MatrixType;
   typedef typename MatrixType::Scalar Scalar;
   typedef typename MatrixType::RealScalar RealScalar;
-  typedef _Preconditioner Preconditioner;
+  typedef Preconditioner_ Preconditioner;
 
-  enum {
-    UpLo = _UpLo
-  };
-
-public:
+  enum { UpLo = UpLo_ };
 
+ public:
   /** Default constructor. */
   ConjugateGradient() : Base() {}
 
   /** Initialize the solver with matrix \a A for further \c Ax=b solving.
-    * 
-    * This constructor is a shortcut for the default constructor followed
-    * by a call to compute().
-    * 
-    * \warning this class stores a reference to the matrix A as well as some
-    * precomputed values that depend on it. Therefore, if \a A is changed
-    * this class becomes invalid. Call compute() to update it with the new
-    * matrix A, or modify a copy of A.
-    */
-  template<typename MatrixDerived>
+   *
+   * This constructor is a shortcut for the default constructor followed
+   * by a call to compute().
+   *
+   * \warning this class stores a reference to the matrix A as well as some
+   * precomputed values that depend on it. Therefore, if \a A is changed
+   * this class becomes invalid. Call compute() to update it with the new
+   * matrix A, or modify a copy of A.
+   */
+  template <typename MatrixDerived>
   explicit ConjugateGradient(const EigenBase<MatrixDerived>& A) : Base(A.derived()) {}
 
   ~ConjugateGradient() {}
 
   /** \internal */
-  template<typename Rhs,typename Dest>
-  void _solve_vector_with_guess_impl(const Rhs& b, Dest& x) const
-  {
+  template <typename Rhs, typename Dest>
+  void _solve_vector_with_guess_impl(const Rhs& b, Dest& x) const {
     typedef typename Base::MatrixWrapper MatrixWrapper;
     typedef typename Base::ActualMatrixType ActualMatrixType;
     enum {
-      TransposeInput  =   (!MatrixWrapper::MatrixFree)
-                      &&  (UpLo==(Lower|Upper))
-                      &&  (!MatrixType::IsRowMajor)
-                      &&  (!NumTraits<Scalar>::IsComplex)
+      TransposeInput = (!MatrixWrapper::MatrixFree) && (UpLo == (Lower | Upper)) && (!MatrixType::IsRowMajor) &&
+                       (!NumTraits<Scalar>::IsComplex)
     };
-    typedef typename internal::conditional<TransposeInput,Transpose<const ActualMatrixType>, ActualMatrixType const&>::type RowMajorWrapper;
-    EIGEN_STATIC_ASSERT(EIGEN_IMPLIES(MatrixWrapper::MatrixFree,UpLo==(Lower|Upper)),MATRIX_FREE_CONJUGATE_GRADIENT_IS_COMPATIBLE_WITH_UPPER_UNION_LOWER_MODE_ONLY);
-    typedef typename internal::conditional<UpLo==(Lower|Upper),
-                                           RowMajorWrapper,
-                                           typename MatrixWrapper::template ConstSelfAdjointViewReturnType<UpLo>::Type
-                                          >::type SelfAdjointWrapper;
+    typedef std::conditional_t<TransposeInput, Transpose<const ActualMatrixType>, ActualMatrixType const&>
+        RowMajorWrapper;
+    EIGEN_STATIC_ASSERT(internal::check_implication(MatrixWrapper::MatrixFree, UpLo == (Lower | Upper)),
+                        MATRIX_FREE_CONJUGATE_GRADIENT_IS_COMPATIBLE_WITH_UPPER_UNION_LOWER_MODE_ONLY);
+    typedef std::conditional_t<UpLo == (Lower | Upper), RowMajorWrapper,
+                               typename MatrixWrapper::template ConstSelfAdjointViewReturnType<UpLo>::Type>
+        SelfAdjointWrapper;
 
     m_iterations = Base::maxIterations();
     m_error = Base::m_tolerance;
@@ -220,10 +209,9 @@ public:
     m_info = m_error <= Base::m_tolerance ? Success : NoConvergence;
   }
 
-protected:
-
+ protected:
 };
 
-} // end namespace Eigen
+}  // end namespace Eigen
 
-#endif // EIGEN_CONJUGATE_GRADIENT_H
+#endif  // EIGEN_CONJUGATE_GRADIENT_H