ngsolve 6.2.2501.post47.dev1__cp313-cp313-macosx_10_15_universal2.whl → 6.2.2501.post48.dev1__cp313-cp313-macosx_10_15_universal2.whl

netgen/include/hcurlcurlfe.hpp

@@ -0,0 +1,2221 @@
+#ifndef FILE_HCURLCURLFE
+#define FILE_HCURLCURLFE
+
+/*********************************************************************/
+/* File:   hcurlcurlfe.hpp                                           */
+/* Author: Michael Neunteufel                                        */
+/* Date:   June 2018                                                 */
+/*********************************************************************/
+
+
+#include "finiteelement.hpp"
+#include "fe_interfaces.hpp"
+#include "hcurlfe.hpp"
+#include "hcurlfe_utils.hpp"
+#include "recursive_pol.hpp"
+#include "recursive_pol_trig.hpp"
+#include "recursive_pol_tet.hpp"
+
+namespace ngfem
+{
+  
+  template <typename T>
+  Mat<3,3,T> TensorCrossProduct(Mat<3,3,T> A, Mat<3,3,T> B)
+  {
+    // return 0.5 * ( Cof(A+B) - Cof(A-B) ); // more cancelation
+
+    Mat<3,3,T> prod;
+    prod.Col(0) = Cross(A.Col(1), B.Col(2)) - Cross(A.Col(2), B.Col(1));
+    prod.Col(1) = Cross(A.Col(2), B.Col(0)) - Cross(A.Col(0), B.Col(2));
+    prod.Col(2) = Cross(A.Col(0), B.Col(1)) - Cross(A.Col(1), B.Col(0));
+    return prod;
+  }
+  
+  template <typename T>
+  Mat<3,3,T> TensorCrossProduct(Vec<3,T> v, Mat<3,3,T> A)
+  {
+    Mat<3,3,T> result;
+    for (int j = 0; j < 3; j++)
+      result.Col(j) = Cross(v, A.Col(j));
+    return result;
+  }
+
+  template <typename T>
+  Mat<3,3,T> TensorCrossProduct(Mat<3,3,T> A, Vec<3,T> v)
+  {
+    Mat<3,3,T> result;
+    for (int j = 0; j < 3; j++)
+      result.Row(j) = Cross(A.Row(j), v);
+    return result;
+  }
+  
+  
+
+  template <int DIM>
+  class HCurlCurlFiniteElement : public FiniteElement
+  {
+  public:
+    using FiniteElement::FiniteElement;
+    using FiniteElement::ndof;
+    using FiniteElement::order;
+    
+    virtual void CalcMappedShape (const BaseMappedIntegrationPoint & bmip,
+                                  BareSliceMatrix<double> shape) const = 0;
+
+    virtual void EvaluateMappedShape (const BaseMappedIntegrationPoint & bmip,
+                                      BareSliceVector<double> coefs,
+                                      BareSliceMatrix<double> shape) const = 0;
+
+    virtual void CalcMappedCurlShape (const BaseMappedIntegrationPoint & bmip,
+      BareSliceMatrix<double> shape) const = 0;
+
+    virtual void CalcMappedIncShape (const BaseMappedIntegrationPoint & bmip,
+                                     BareSliceMatrix<double> shape) const = 0;
+
+    virtual void EvaluateMappedIncShape (const BaseMappedIntegrationPoint & bmip,
+                                         BareSliceVector<double> coefs,
+                                         BareSliceVector<double> inc) const = 0;
+
+    virtual void CalcMappedIncShape (const SIMD_BaseMappedIntegrationRule & bmir,
+                                     BareSliceMatrix<SIMD<double>> shape) const = 0;
+
+    virtual void EvaluateIncShape (const SIMD_BaseMappedIntegrationRule & ir,
+                                   BareSliceVector<> coefs,
+                                   BareSliceMatrix<SIMD<double>> values) const = 0;
+
+    virtual void AddTransIncShape (const SIMD_BaseMappedIntegrationRule & ir,
+                                   BareSliceMatrix<SIMD<double>> values,
+                                   BareSliceVector<> coefs) const = 0;
+
+    
+    virtual void CalcMappedShape (const SIMD_BaseMappedIntegrationRule & bmir, 
+                                         BareSliceMatrix<SIMD<double>> shapes) const = 0;
+    
+    virtual void Evaluate (const SIMD_BaseMappedIntegrationRule & ir,
+                                  BareSliceVector<> coefs,
+                                  BareSliceMatrix<SIMD<double>> values) const = 0;
+
+    virtual void AddTrans (const SIMD_BaseMappedIntegrationRule & ir,
+                                  BareSliceMatrix<SIMD<double>> values,
+                                  BareSliceVector<> coefs) const = 0;
+
+    virtual void CalcDualShape (const BaseMappedIntegrationPoint & bmip, BareSliceMatrix<> shape) const = 0;
+    virtual void CalcDualShape (const SIMD_BaseMappedIntegrationRule & bmir, BareSliceMatrix<SIMD<double>> shape) const = 0;
+    virtual void EvaluateDual (const SIMD_BaseMappedIntegrationRule & bmir, BareSliceVector<> coefs, BareSliceMatrix<SIMD<double>> values) const = 0;
+    virtual void AddDualTrans (const SIMD_BaseMappedIntegrationRule& bmir, BareSliceMatrix<SIMD<double>> values, BareSliceVector<double> coefs) const = 0;
+
+  };
+
+  template <int D,typename VEC,typename MAT>
+  void VecToSymMat(const VEC & vec, MAT & mat)
+  {
+    switch(D)
+    {
+    case 1:
+      mat(0) = vec(0);
+      break;
+    case 2:
+      mat(0) = vec(0);
+      mat(3) = vec(1);
+      mat(1) = mat(2) = vec(2);
+      break;
+    case 3:
+      mat(0) = vec(0);
+      mat(4) = vec(1);
+      mat(8) = vec(2);
+      mat(1) = mat(3) = vec(5);
+      mat(2) = mat(6) = vec(4);
+      mat(5) = mat(7) = vec(3);
+      break;
+    }
+  }
+
+  template <int H, int W, typename T>
+  Mat<H,W,T> DyadProd(Vec<H,T> a, Vec<W,T> b)
+  {
+    Mat<H,W,T> m;
+    for (int i = 0; i < H; i++)
+      for (int j = 0; j < W; j++)
+        m(i,j) = a(i)*b(j);
+    return m;
+  }
+  
+  template <int S, typename T>
+  Mat<S,S,T> SymDyadProd(Vec<S,T> a, Vec<S,T> b)
+  {
+    Mat<S,S,T> m;
+    for (int i = 0; i < S; i++)
+      for (int j = 0; j < S; j++)
+        m(i,j) = a(i)*b(j)+a(j)*b(i);
+    return m;
+  }
+
+
+  template <typename T>
+  Vec<6, AutoDiff<3,T>> SymDyadProd(AutoDiff<3,T> a, AutoDiff<3,T> b)
+  {
+    return Vec<6, AutoDiff<3,T>>(2*a.DValue(0)*b.DValue(0),2*a.DValue(1)*b.DValue(1),2*a.DValue(2)*b.DValue(2), a.DValue(1)*b.DValue(2)+a.DValue(2)*b.DValue(1), a.DValue(0)*b.DValue(2)+a.DValue(2)*b.DValue(0),a.DValue(1)*b.DValue(0)+a.DValue(0)*b.DValue(1));
+  }
+
+  template <typename T>
+  Vec<6, AutoDiff<3,T>> SymDyadProdAD(Vec<3,T> a, Vec<3,T> b)
+  {
+    return Vec<6, AutoDiff<3,T>>(2*a(0)*b(0),2*a(1)*b(1),2*a(2)*b(2), a(1)*b(2)+a(2)*b(1), a(0)*b(2)+a(2)*b(0),a(1)*b(0)+a(0)*b(1));
+  }
+
+  template <typename T>
+  Vec<3,AutoDiff<2,T>> SymDyadProd(AutoDiff<2,T> a, AutoDiff<2,T> b)
+  {
+    return Vec<3,AutoDiff<2,T>>(2*a.DValue(0)*b.DValue(0),2*a.DValue(1)*b.DValue(1),a.DValue(1)*b.DValue(0)+a.DValue(0)*b.DValue(1));
+  }
+
+  template <typename T>
+  Vec<3,AutoDiff<2,T>> SymDyadProdAD(Vec<2,T> a, Vec<2,T> b)
+  {
+    return Vec<3,AutoDiff<2,T>>(2*a(0)*b(0),2*a(1)*b(1),a(1)*b(0)+a(0)*b(1));
+  }
+
+  template <typename T>
+  AutoDiff<1,T> SymDyadProd(AutoDiff<1,T> a, AutoDiff<1,T> b)
+  {
+    return a.DValue(0)*b.DValue(0);
+  }
+
+
+  //------------------REGGE_SHAPE---------------------
+  template <int D, typename T> class T_REGGE_Shape;
+  template <typename T> class T_REGGE_Shape<1,T>
+  {
+    AutoDiff<1,T> u;
+  public:
+    T_REGGE_Shape  (AutoDiff<1,T> au) : u(au) { ; }
+    Vec<1,T> Shape() { return u.Value(); }
+    /*0 2
+      2 1*/
+    Vec<1,T> CurlShape() { return 0.0; }
+  };
+  
+  template <typename T> class T_REGGE_Shape<2,T>
+  {
+    Vec<3,AutoDiff<2,T>> u;
+  public:
+    T_REGGE_Shape  (Vec<3,AutoDiff<2,T>> au) : u(au) { ; }
+    Vec<3,T> Shape() { return Vec<3,T> (u(0).Value(), u(1).Value(), u(2).Value()); }
+    /*0 2
+      2 1*/
+    Vec<2,T> CurlShape() { return Vec<2,T> (u(2).DValue(0)-u(0).DValue(1), u(1).DValue(0)-u(2).DValue(1)); }
+  };
+  
+  template <typename T> class T_REGGE_Shape<3,T>
+  {
+    Vec<6,AutoDiff<3,T>> u;
+  public:
+    T_REGGE_Shape  (Vec<6,AutoDiff<3,T>> au) : u(au) { ; }
+    Vec<6,T> Shape() { return Vec<6,T> (u(0).Value(), u(1).Value(), u(2).Value(), u(3).Value(), u(4).Value(), u(5).Value()); }
+    /*0 5 4
+      5 1 3
+      4 3 2*/
+    Vec<9,T> CurlShape() { return Vec<9,T> (u(4).DValue(1)-u(5).DValue(2), -u(4).DValue(0)+u(0).DValue(2), u(5).DValue(0)-u(0).DValue(1),
+					    u(3).DValue(1)-u(1).DValue(2), -u(3).DValue(0)+u(5).DValue(2), u(1).DValue(0)-u(5).DValue(1),
+					    u(2).DValue(1)-u(3).DValue(2), -u(2).DValue(0)+u(4).DValue(2), u(3).DValue(0)-u(4).DValue(1)); }
+  };
+  //---------------------------------------------------
+
+
+    // ***************** EpsGrad ****************************** */
+  // eps (nabla u)
+  
+  template <int D, typename T> class T_EpsGrad;
+  template <typename T> class T_EpsGrad<2,T>
+  {
+    AutoDiffDiff<2,T> u;
+  public:
+    T_EpsGrad  (AutoDiffDiff<2,T> au) : u(au) { ; }
+    Vec<3,T> Shape()
+    {
+      return Vec<3,T> (u.DDValue(0,0), u.DDValue(1,1), u.DDValue(0,1));
+    }
+    Vec<2,T> CurlShape() { return Vec<2,T> (0.0, 0.0); }
+  };
+  
+  template <int D, typename T>
+  auto EpsGrad (AutoDiffDiff<D,T> au) { return T_EpsGrad<D,T>(au); }
+
+  // ***************** wEpsGrad ****************************** */
+  // w*eps (nabla u)
+  
+  template <int D, typename T> class T_wEpsGrad;
+  template <typename T> class T_wEpsGrad<2,T>
+  {
+    AutoDiffDiff<2,T> u;
+    AutoDiff<1,T> w;
+  public:
+    T_wEpsGrad  (AutoDiffDiff<2,T> au, AutoDiff<1,T> aw) : u(au), w(aw) { ; }
+    Vec<6,T> Shape()
+    {
+      return w.Value()*Vec<6,T> (u.DDValue(0,0), u.DDValue(1,1), u.DDValue(2,2), u.DDValue(1,2), u.DDValue(0,2), u.DDValue(0,1));
+    }
+    Vec<9,T> CurlShape() { return Vec<9,T> (0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, 0.0); }
+  };
+  
+  template <int D, typename T>
+  auto wEpsGrad (AutoDiffDiff<D,T> au, AutoDiff<1,T> aw) { return T_wEpsGrad<D,T>(au, aw); }
+  
+  
+  // ***************** Eps_u_Gradv ****************************** */
+  // eps (u nabla v)
+  
+  template <int D, typename T> class T_Eps_u_Gradv;
+  template <typename T> class T_Eps_u_Gradv<2,T>
+  {
+    AutoDiffDiff<2,T> u, v;
+  public:
+    T_Eps_u_Gradv  (AutoDiffDiff<2,T> au, AutoDiffDiff<2,T> av) : u(au), v(av) { ; }
+    Vec<3,T> Shape() { return Vec<3,T> ((u.Value()*v.DDValue(0,0) + u.DValue(0)*v.DValue(0)),
+                                        (u.Value()*v.DDValue(1,1) + u.DValue(1)*v.DValue(1)),
+                                        u.Value()*v.DDValue(0,1) + 0.5 * (u.DValue(0)*v.DValue(1)+u.DValue(1)*v.DValue(0))); }
+    Vec<2,T> CurlShape()
+    {
+      T uxx = u.DDValue(0,0), uyy = u.DDValue(1,1), uxy = u.DDValue(0,1);
+      T ux = u.DValue(0), uy = u.DValue(1);
+      T vxx = v.DDValue(0,0), vyy = v.DDValue(1,1), vxy = v.DDValue(0,1);
+      T vx = v.DValue(0), vy = v.DValue(1);
+      
+      /*return -0.5 * Vec<2,T> (uyy*vx - uxy*vy + uy*vxy - ux*vyy,
+      -uxy*vx + uxx*vy - uy*vxx + ux*vxy);*/
+      return 0.5 * Vec<2,T>(ux*vxy - uy*vxx - uxy*vx + uxx*vy,
+                            ux*vyy + uxy*vy - uyy*vx - uy*vxy);
+    }
+  };
+  
+  template <int D, typename T>
+  auto Eps_u_Gradv (AutoDiffDiff<D,T> au, AutoDiffDiff<D,T> av) { return T_Eps_u_Gradv<D,T>(au, av); }
+  
+  
+  template <int D, typename T> class T_vEpsGradu;
+  template <typename T> class T_vEpsGradu<2,T>
+  {
+    AutoDiffDiff<2,T> u,v;
+  public:
+    T_vEpsGradu  (AutoDiffDiff<2,T> au, AutoDiffDiff<2,T> av) : u(au), v(av) { ; }
+    Vec<3,T> Shape() { return Vec<3,T> (u.DDValue(0,0)*v.Value(),
+                                      u.DDValue(1,1)*v.Value(),  (u.DDValue(1,0)*v.Value()));}
+    Vec<2,T> CurlShape()
+    {
+      T uxx = u.DDValue(0,0), uyy = u.DDValue(1,1), uxy = u.DDValue(0,1);
+      T vx = v.DValue(0), vy = v.DValue(1);
+
+      //return Vec<2,T> (uyy*vx- uxy*vy, uxx*vy- uxy*vx);
+      return Vec<2,T> (uxy*vx - vy*uxx, uyy*vx - uxy*vy);
+    }
+  };
+  
+  template <int D, typename T>
+  auto vEpsGradu (AutoDiffDiff<D,T> au, AutoDiffDiff<D,T> av) { return T_vEpsGradu<D,T>(au, av); }
+
+  
+  template <int D, typename T>
+  class ReggeAD
+  {
+    Mat<D,D,T> value;
+
+  public:
+    ReggeAD ()
+    {
+      value = T(0);
+    }
+    
+    ReggeAD (AutoDiff<D,T> a, AutoDiff<D,T> b)
+    {
+      Vec<D,T> Da, Db;
+      for(int i=0; i<D; i++)
+        {
+          Da(i) = a.DValue(i);
+          Db(i) = b.DValue(i);
+        }      
+      value = SymDyadProd(Da,Db);
+    }
+
+    auto Value() const { return value; }
+
+    Mat<D,D,T> & Value() { return value; }
+  };
+
+  template <int D, typename T>
+  auto MakeReggeAD(AutoDiff<D,T> a, AutoDiff<D,T> b)
+  {
+    return ReggeAD<D,T>(a, b);
+  }
+
+
+
+  template <int D, typename T>
+  ReggeAD<D,T> operator* (AutoDiff<D,T> s, ReggeAD<D,T> A)
+  {
+    ReggeAD<D,T> result;
+    result.Value() = s.Value()*A.Value();
+    return result;
+  }
+
+  
+  template <int D, typename T>
+  ReggeAD<D,T> operator* (T s, ReggeAD<D,T> A)
+  {
+    ReggeAD<D,T> result = A;
+    result.Value() *= s;
+    return result;
+  }
+
+  template <int D, typename T>
+  ReggeAD<D,T> operator+ (ReggeAD<D,T> A, ReggeAD<D,T> B)
+  {
+    ReggeAD<D,T> result = A;
+    result.Value() += B.Value();
+    return result;
+  }
+
+  template <int D, typename T>
+  ReggeAD<D,T> operator- (ReggeAD<D,T> A, ReggeAD<D,T> B) 
+  {
+    ReggeAD<D,T> result = A;
+    result.Value() -= B.Value();
+    return result;
+  }
+  
+  template <int D, typename T> class ReggeADD;
+
+  template <typename T>
+  class ReggeADD<3,T>
+  {
+    Mat<3,3,T> value;
+    Mat<3,3,T> curl;
+    Mat<3,3,T> inc;
+
+  public:
+    ReggeADD ()
+    {
+      value = T(0);
+      curl = T(0);
+      inc = T(0);
+    }
+    
+    ReggeADD (AutoDiffDiff<3,T> a, AutoDiffDiff<3,T> b)
+    {
+      auto Da = Vec<3,T>(a.DValue(0),a.DValue(1),a.DValue(2));
+      auto Db = Vec<3,T>(b.DValue(0),b.DValue(1),b.DValue(2));
+      value = SymDyadProd(Da,Db);
+      // curl(s*v) = nabla s x v + s curl(v) in 3D
+      //                     | nabla a_1 x b + a_1 curl(b) |
+      // curl(a \otimes b) = | nabla a_2 x b + a_2 curl(b) |
+      //                     | nabla a_3 x b + a_3 curl(b) |
+
+      // curl( nabla s ) = 0
+      //                                      | nabla d_x a x nabla b |
+      // ->  curl( nabla a \otimes nabla b) = | nabla d_y a x nabla b |
+      //                                      | nabla d_z a x nabla b |
+
+      Vec<3,T> Ddai [3] = { Vec<3,T>(a.DDValue(0,0), a.DDValue(0,1), a.DDValue(0,2)), Vec<3,T>(a.DDValue(1,0), a.DDValue(1,1), a.DDValue(1,2)), Vec<3,T>(a.DDValue(2,0), a.DDValue(2,1), a.DDValue(2,2)) };
+      Vec<3,T> Ddbi [3] = { Vec<3,T>(b.DDValue(0,0), b.DDValue(0,1), b.DDValue(0,2)), Vec<3,T>(b.DDValue(1,0), b.DDValue(1,1), b.DDValue(1,2)), Vec<3,T>(b.DDValue(2,0), b.DDValue(2,1), b.DDValue(2,2)) };
+
+      for (int i = 0; i < 3; i++)
+        curl.Row(i) = Cross(Ddai[i], Db) + Cross(Ddbi[i], Da);
+
+
+      //                                   | nabla d_x a x nabla b |
+      //  curl( nabla a \otimes nabla b) = | nabla d_y a x nabla b |
+      //                                   | nabla d_z a x nabla b |
+
+
+      // curl T curl ( nabla a \otimes nabla b):
+      //11: d_yd_z(a) d_yd_z(b) - d^2_z(a) d^2_y(b)   - d_y^2(a) d^2_z(b)   + d_yd_z(a)d_yd_z(b)
+      //12: d_xd_y(a) d_z^2(b)  - d_xd_z(a) d_yd_z(b) - d_yd_z(a) d_xd_z(b) + d_z^2(a)d_xd_y(b)
+      //13: d^2_y(a) d_xd_z(b)  - d_yd_z(a) d_xd_y(b) - d_xd_y(a) d_yd_z(b) + d_xd_z(a)d^2_y(b)
+      //22: d_xd_z(a) d_xd_z(b) - d^2_x(a) d^2_z(b)   - d^2_z(a) d^2_x(b)   + d_xd_z(a)d_xd_z(b)
+      //23: d_yd_z(a) d^2_x(b)  - d_xd_y(a) d_xd_z(b) - d_xd_z(a) d_xd_y(b) + d^2_x(a)d_yd_z(b)
+      //33: d^2_x(a) d^2_y(b)   - d_xd_y(a) d_xd_y(b) - d_xd_y(a) d_xd_y(b) + d^2_y(a)d^2_x(b)
+
+      //  = - hesse(a) x hesse(b) = -eps_imn eps_jlk d_md_l(a) d_nd_k b ??
+
+      /*inc(0,0) = a.DDValue(1,2)*b.DDValue(1,2) - a.DDValue(2,2)*b.DDValue(1,1) - a.DDValue(1,1)*b.DDValue(2,2) + a.DDValue(1,2)*b.DDValue(1,2);
+      inc(0,1) = a.DDValue(0,1)*b.DDValue(2,2) - a.DDValue(0,2)*b.DDValue(1,2) - a.DDValue(1,2)*b.DDValue(0,2) + a.DDValue(2,2)*b.DDValue(0,1);
+      inc(0,2) = a.DDValue(1,1)*b.DDValue(0,2) - a.DDValue(1,2)*b.DDValue(0,1) - a.DDValue(0,1)*b.DDValue(1,2) + a.DDValue(0,2)*b.DDValue(1,1);
+      inc(1,1) = a.DDValue(0,2)*b.DDValue(0,2) - a.DDValue(0,0)*b.DDValue(2,2) - a.DDValue(2,2)*b.DDValue(0,0) + a.DDValue(0,2)*b.DDValue(0,2);
+      inc(1,2) = a.DDValue(1,2)*b.DDValue(0,0) - a.DDValue(0,1)*b.DDValue(0,2) - a.DDValue(0,2)*b.DDValue(0,1) + a.DDValue(0,0)*b.DDValue(1,2);
+      inc(2,2) = a.DDValue(0,0)*b.DDValue(1,1) - a.DDValue(0,1)*b.DDValue(0,1) - a.DDValue(0,1)*b.DDValue(0,1) + a.DDValue(1,1)*b.DDValue(0,0);
+      // curl T curl ( nabla b \otimes nabla a):
+      inc(0,0) += b.DDValue(1,2)*a.DDValue(1,2) - b.DDValue(2,2)*a.DDValue(1,1) - b.DDValue(1,1)*a.DDValue(2,2) + b.DDValue(1,2)*a.DDValue(1,2);
+      inc(0,1) += b.DDValue(0,1)*a.DDValue(2,2) - b.DDValue(0,2)*a.DDValue(1,2) - b.DDValue(1,2)*a.DDValue(0,2) + b.DDValue(2,2)*a.DDValue(0,1);
+      inc(0,2) += b.DDValue(1,1)*a.DDValue(0,2) - b.DDValue(1,2)*a.DDValue(0,1) - b.DDValue(0,1)*a.DDValue(1,2) + b.DDValue(0,2)*a.DDValue(1,1);
+      inc(1,1) += b.DDValue(0,2)*a.DDValue(0,2) - b.DDValue(0,0)*a.DDValue(2,2) - b.DDValue(2,2)*a.DDValue(0,0) + b.DDValue(0,2)*a.DDValue(0,2);
+      inc(1,2) += b.DDValue(1,2)*a.DDValue(0,0) - b.DDValue(0,1)*a.DDValue(0,2) - b.DDValue(0,2)*a.DDValue(0,1) + b.DDValue(0,0)*a.DDValue(1,2);
+      inc(2,2) += b.DDValue(0,0)*a.DDValue(1,1) - b.DDValue(0,1)*a.DDValue(0,1) - b.DDValue(0,1)*a.DDValue(0,1) + b.DDValue(1,1)*a.DDValue(0,0);
+      // symmetry
+      inc(1,0) = inc(0,1);
+      inc(2,0) = inc(0,2);
+      inc(2,1) = inc(1,2);*/
+
+      Mat<3,3,T> hesse1, hesse2;
+      a.StoreHessian(hesse1.Data());
+      b.StoreHessian(hesse2.Data());
+      inc = -2*TensorCrossProduct(hesse1,hesse2);
+    }
+
+    auto Value() const { return value; }
+    auto Curl() const { return curl; }
+    auto Inc() const { return inc; }
+
+    Mat<3,3,T> & Value() { return value; }
+    Mat<3,3,T> & Curl() { return curl; }
+    Mat<3,3,T> & Inc() { return inc; }
+  };
+
+
+  template <typename T>
+  class ReggeADD<2,T>
+  {
+    Mat<2,2,T> value;
+    Vec<2,T> curl;
+    T inc;
+
+  public:
+    ReggeADD ()
+    {
+      value = T(0);
+      curl = T(0);
+      inc = T(0);
+    }
+    
+    ReggeADD (AutoDiffDiff<2,T> a, AutoDiffDiff<2,T> b)
+    {
+      auto Da = Vec<2,T>(a.DValue(0),a.DValue(1));
+      auto Db = Vec<2,T>(b.DValue(0),b.DValue(1));
+      value = SymDyadProd(Da,Db);
+      // curl(s*v) = v* nabla s^perp + s curl(v) in 2D
+      //                     | b * nabla a_1^perp + a_1 curl(b) |
+      // curl(a \otimes b) = | b * nabla a_2^perp + a_2 curl(b) |
+
+      // curl( nabla s ) = 0
+      //                                      | Db * nabla d_x a^perp |
+      // ->  curl( nabla a \otimes nabla b) = | Db * nabla d_y a^perp |
+
+      Vec<2,T> Ddai_p [2] = { Vec<2,T>(-a.DDValue(0,1), a.DDValue(0,0)), Vec<2,T>(-a.DDValue(1,1), a.DDValue(1,0)) };
+      Vec<2,T> Ddbi_p [2] = { Vec<2,T>(-b.DDValue(0,1), b.DDValue(0,0)), Vec<2,T>(-b.DDValue(1,1), b.DDValue(1,0)) };
+
+      for (int i=0; i<2; i++)
+        curl(i) = InnerProduct(Db,Ddai_p[i]) + InnerProduct(Da,Ddbi_p[i]);
+
+      //                                  | Db * nabla d_x a^perp |
+      // curl( nabla a \otimes nabla b) = | Db * nabla d_y a^perp |
+
+
+      // curl T curl ( nabla a \otimes nabla b) = d_x(Db * nabla d_y a^perp) - d_y(Db * nabla d_x a^perp)
+      //       =  (d_x Db) * nabla(d_y a)^perp - (d_y Db) * nabla(d_x a)^perp
+
+      inc = InnerProduct(Vec<2,T>(b.DDValue(0,0), b.DDValue(0,1)),Ddai_p[1]) - InnerProduct(Vec<2,T>(b.DDValue(1,0), b.DDValue(1,1)),Ddai_p[0]) + InnerProduct(Vec<2,T>(a.DDValue(0,0), a.DDValue(0,1)),Ddbi_p[1]) - InnerProduct(Vec<2,T>(a.DDValue(1,0), a.DDValue(1,1)),Ddbi_p[0]);
+    }
+
+    auto Value() const { return value; }
+    auto Curl() const { return curl; }
+    auto Inc() const { return inc; }
+
+    Mat<2,2,T> & Value() { return value; }
+    Vec<2,T> & Curl() { return curl; }
+    T & Inc() { return inc; }
+  };
+  
+  template <int D, typename T>
+  auto MakeReggeAD(AutoDiffDiff<D,T> a, AutoDiffDiff<D,T> b)
+  {
+    return ReggeADD<D,T>(a, b);
+  }
+
+  template <typename T>
+  ReggeADD<3,T> operator* (AutoDiffDiff<3,T> s, ReggeADD<3,T> A)
+  {
+    ReggeADD<3,T> result;
+    result.Value() = s.Value()*A.Value();
+
+    // s scalar, v vector
+    // curl(s*v) = nabla s x v + s curl(v) in 3D
+    Vec<3,T> gradient;
+    s.StoreGradient(gradient.Data());
+
+    result.Curl() = s.Value()*A.Curl();
+    for (int i = 0; i < 3; i++)
+      result.Curl().Row(i) += Cross(gradient, Vec<3,T>(A.Value().Row(i)));
+    
+    // inc(s A) = s inc(A) + 2sym(grad(s) x curl A) + hesse(s) x A, x...Tensor-Cross-Product
+    Mat<3,3,T> hesse;
+    s.StoreHessian(hesse.Data());
+    
+    result.Inc() = s.Value()*A.Inc() + TensorCrossProduct(gradient,A.Curl()) + Trans(TensorCrossProduct(gradient,A.Curl())) +  TensorCrossProduct(hesse,A.Value());
+
+    return result;
+  }
+
+  
+  template <int D, typename T>
+  ReggeADD<D,T> operator* (T s, ReggeADD<D,T> A)
+  {
+    ReggeADD<D,T> result = A;
+    result.Value() *= s;
+    result.Curl()  *= s;
+    result.Inc()   *= s;
+    return result;
+  }
+
+  template <int D, typename T>
+  ReggeADD<D,T> operator+ (ReggeADD<D,T> A, ReggeADD<D,T> B)
+  {
+    ReggeADD<D,T> result = A;
+    result.Value() += B.Value();
+    result.Curl()  += B.Curl();
+    result.Inc()   += B.Inc();
+    return result;
+  }
+
+  template <int D, typename T>
+  ReggeADD<D,T> operator- (ReggeADD<D,T> A, ReggeADD<D,T> B) 
+  {
+    ReggeADD<D,T> result = A;
+    result.Value() -= B.Value();
+    result.Curl()  -= B.Curl();
+    result.Inc()   -= B.Inc();
+    return result;
+  }
+
+
+  template <typename T>
+  ReggeADD<2,T> operator* (AutoDiffDiff<2,T> s, ReggeADD<2,T> A)
+  {
+    ReggeADD<2,T> result;
+    result.Value() = s.Value()*A.Value();
+
+    // s scalar, v vector
+    // curl(s*v) = v* nabla s^perp + s curl(v) in 2D
+    result.Curl() = A.Value()*Vec<2,T>(-s.DValue(1),s.DValue(0)) + s.Value()*A.Curl();
+    
+    // inc(sA) = A:( dydy s & -dxdy s \\ -dxdy s & dxdx s) + 2*nabla s^\perp*curl(A) + s*inc(A)
+    Mat<2,2,T> hesse;
+    hesse(0,0) = s.DDValue(1,1);
+    hesse(1,0) = -s.DDValue(1,0);
+    hesse(0,1) = -s.DDValue(0,1);
+    hesse(1,1) = s.DDValue(0,0);
+
+    result.Inc() = s.Value()*A.Inc() + InnerProduct(hesse,A.Value()) + 2*InnerProduct(Vec<2,T>(-s.DValue(1),s.DValue(0)),A.Curl());  
+
+    return result;
+  }
+
+
+
+  
+  template <ELEMENT_TYPE ET> class HCurlCurlFE;
+
+  
+  template <ELEMENT_TYPE ET>
+  class T_HCurlCurlFE : public HCurlCurlFiniteElement<ET_trait<ET>::DIM>,
+    public VertexOrientedFE<ET>
+  {
+  protected:
+    static constexpr int DIM = ET_trait<ET>::DIM;
+    enum { DIM_STRESS = (DIM*(DIM+1))/2 };
+    // enum { DIM_DMAT = 7*DIM-12 };
+    // enum { DIM_DDMAT = 8*DIM-15 };
+    enum { DIM_DMAT = (5*DIM*DIM-11*DIM+6)/2 };
+    enum { DIM_DDMAT = (7*DIM*DIM-19*DIM+12)/2 };
+    
+    using VertexOrientedFE<ET>::vnums;
+    using HCurlCurlFiniteElement<ET_trait<ET>::DIM>::ndof;
+    using HCurlCurlFiniteElement<ET_trait<ET>::DIM>::order;
+    
+
+    int order_edge[ET_trait<ET>::N_EDGE];
+    IVec<DIM-1> order_facet[ET_trait<ET>::N_FACET];
+    IVec<DIM> order_inner;
+
+    
+  public:
+    using VertexOrientedFE<ET>::SetVertexNumbers;
+    
+    T_HCurlCurlFE (int aorder)
+    {
+      order = aorder;
+      for (auto & of : order_facet) of = aorder;
+      order_inner = aorder;
+
+    }
+    
+    virtual ELEMENT_TYPE ElementType() const override { return ET; }
+    const HCurlCurlFE<ET> * Cast() const { return static_cast<const HCurlCurlFE<ET>*> (this); } 
+    
+    INLINE void SetOrderFacet (int nr, IVec<DIM-1,int> order) { order_facet[nr] = order; }
+    INLINE void SetOrderEdge (int nr, int order) { order_edge[nr] = order; }
+    INLINE void SetOrderInner (IVec<DIM,int> order) { order_inner = order; }
+
+    virtual void ComputeNDof()
+    {
+      cout << "Error, T_HCurlCurlFE<ET>:: ComputeNDof not available, only for ET == TRIG" << endl;
+    }
+
+
+    virtual void CalcMappedShape (const BaseMappedIntegrationPoint & bmip,
+                                  BareSliceMatrix<double> shapes) const override
+    {
+      Switch<4-DIM>
+        (bmip.DimSpace()-DIM,[this, &bmip, shapes](auto CODIM)
+         {
+           constexpr auto DIMSPACE = DIM+CODIM.value;
+           auto & mip = static_cast<const MappedIntegrationPoint<DIM, DIM+CODIM.value>&> (bmip);
+           
+           Cast() -> T_CalcShape (GetTIP(mip),SBLambda([shapes,DIMSPACE](int nr,auto val)
+                                                   {
+                                                     shapes.Row(nr).Range(DIMSPACE*DIMSPACE) = val.Value().AsVector();
+                                                   }));
+         });
+    }
+
+
+    
+    virtual void EvaluateMappedShape (const BaseMappedIntegrationPoint & bmip,
+                                      BareSliceVector<double> coefs,
+                                      BareSliceMatrix<double> shape) const override
+    {
+      Switch<4-DIM>
+        (bmip.DimSpace()-DIM,[this, &bmip, coefs, shape](auto CODIM)
+      {
+        auto & mip = static_cast<const MappedIntegrationPoint<DIM,DIM+CODIM.value>&> (bmip);
+
+        Mat<DIM+CODIM.value,DIM+CODIM.value> summat(0);
+        Cast() -> T_CalcShape (GetTIP(mip), SBLambda ([&summat,coefs] (int nr, auto val)
+                                                       {
+                                                         summat += coefs(nr) * val.Value();
+
+                                                       }));
+        for (int k = 0; k < sqr(DIM+CODIM.value); k++)
+          shape(k) = summat(k);
+      });
+
+    }
+
+    virtual void CalcMappedIncShape (const BaseMappedIntegrationPoint & bmip,
+                                  BareSliceMatrix<double> shapes) const override
+    {
+      auto & mip = static_cast<const MappedIntegrationPoint<DIM,DIM>&> (bmip);
+      if constexpr (ET == ET_TET || ET == ET_TRIG || ET == ET_QUAD)
+                     Cast() -> T_CalcShape (GetTIPHesse(mip),SBLambda([shapes](int nr,auto val)
+                                                                       {
+                                                                         if constexpr (DIM==3)
+                                                                           shapes.Row(nr).Range(DIM_DDMAT) = val.Inc().AsVector();
+                                                                         else
+                                                                           shapes.Row(nr).Range(DIM_DDMAT) = val.Inc();
+                                                                       }));
+      else
+        throw Exception("HCurlCurl::CalcMappedIncShape implemented only for TRIG and TET");
+      
+    }
+                            
+    virtual void EvaluateMappedIncShape (const BaseMappedIntegrationPoint & bmip,
+                                         BareSliceVector<double> coefs,
+                                         BareSliceVector<double> inc) const override
+    {
+      auto & mip = static_cast<const MappedIntegrationPoint<DIM,DIM>&> (bmip);
+
+      Mat<DIM*(DIM-1)/2,DIM*(DIM-1)/2> sum = 0.0;
+      if constexpr (ET == ET_TET || ET == ET_TRIG || ET == ET_QUAD)
+                     Cast() -> T_CalcShape (GetTIPHesse(mip),SBLambda([coefs, &sum](int nr,auto val)
+                                                                       {
+                                                                         sum += coefs(nr) * Mat<DIM*(DIM-1)/2,DIM*(DIM-1)/2>(val.Inc());
+                                                                       }));
+      else
+        throw Exception("HCurlCurl::EvaluateMappedIncShape implemented only for TRIG and TET");
+      
+      inc.Range(0,DIM_DDMAT) = sum.AsVector();
+    }
+
+    virtual void CalcMappedIncShape (const SIMD_BaseMappedIntegrationRule & bmir,
+                                     BareSliceMatrix<SIMD<double>> shapes) const override
+    {
+      auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM>&> (bmir);
+      for (size_t i = 0; i < mir.Size(); i++)
+        {
+          if constexpr (ET == ET_TET || ET == ET_TRIG || ET == ET_QUAD)
+                     {
+                       Cast() -> T_CalcShape (GetTIPHesse(mir[i]),SBLambda([shapes,i](int j,auto val)
+                                                                         {
+                                                                           if constexpr (DIM==3)
+                                                                                          shapes.Rows(j*sqr(DIM), (j+1)*sqr(DIM)).Col(i).Range(0,DIM_DDMAT) = val.Inc().AsVector();
+                                                                           else
+                                                                             shapes.Rows(j,j+1).Col(i).Range(0,DIM_DDMAT) = val.Inc();
+                                                                         }));
+                     }
+          else
+            throw Exception("HCurlCurl::CalcMappedIncShape implemented only for TRIG and TET");
+
+        }
+    }
+
+    void EvaluateIncShape (const SIMD_BaseMappedIntegrationRule & bmir,
+                           BareSliceVector<> coefs,
+                           BareSliceMatrix<SIMD<double>> values) const override
+    {
+      auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM>&> (bmir);
+      for (size_t i = 0; i < bmir.Size(); i++)
+        {
+          double *pcoefs = &coefs(0);
+          const size_t dist = coefs.Dist();  
+          if constexpr (ET == ET_TET && DIM == 3)
+                     {
+                       Mat<DIM,DIM,SIMD<double>> summat(0);
+                       Cast() -> T_CalcShape (GetTIPHesse(mir[i]),
+                                               SBLambda ([&summat,&pcoefs,dist] (size_t j, auto val)
+                                                         {
+                                                           summat += (*pcoefs)*val.Inc();
+                                                           pcoefs += dist;
+                                                         }));
+                       for (size_t k = 0; k < sqr(DIM); k++)
+                         values(k,i) = summat(k);
+                     }
+          if constexpr ((ET == ET_TRIG || ET == ET_QUAD) && DIM == 2)
+                     {
+                       SIMD<double> summat(0);
+                       Cast() -> T_CalcShape (GetTIPHesse(mir[i]),
+                                               SBLambda ([&summat,&pcoefs,dist] (size_t j, auto val)
+                                                         {
+                                                           summat += (*pcoefs)*val.Inc();
+                                                           pcoefs += dist;
+                                                         }));
+                       values(0,i) = summat;
+                     }
+        }
+          
+      /*if constexpr (ET == ET_TET || ET == ET_TRIG)
+          {
+            Switch<1>
+              (bmir.DimSpace()-DIM,[values,&bmir,coefs,this](auto CODIM)
+              {
+                constexpr auto DIMSPACE = DIM+CODIM.value;
+                auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIMSPACE>&> (bmir);
+                for (size_t i = 0; i < bmir.Size(); i++)
+                  {
+                    double *pcoefs = &coefs(0);
+                    const size_t dist = coefs.Dist();            
+                  
+
+                    if constexpr (DIMSPACE==3)
+                       {
+                         Mat<DIMSPACE,DIMSPACE,SIMD<double>> summat(0);
+                         Cast() -> T_CalcShape (GetTIPHesse(mir[i]),
+                                                 SBLambda ([&summat,&pcoefs,dist] (size_t j, auto val)
+                                                           {
+                                                             summat += (*pcoefs)*val.Inc();
+                                                             pcoefs += dist;
+                                                           }));
+                         for (size_t k = 0; k < sqr(DIMSPACE); k++)
+                           values(k,i) = summat(k);
+                       }
+                    if constexpr (DIMSPACE==2)
+                      {
+                        SIMD<double> summat(0);
+                        Cast() -> T_CalcShape (GetTIPHesse(mir[i]),
+                                                SBLambda ([&summat,&pcoefs,dist] (size_t j, auto val)
+                                                          {
+                                                            summat += (*pcoefs)*val.Inc();
+                                                            pcoefs += dist;
+                                                          }));
+                        values(0,i) = summat;
+                        }
+                   
+                  }
+              });
+              }*/
+    }
+
+    void AddTransIncShape (const SIMD_BaseMappedIntegrationRule & ir,
+                                   BareSliceMatrix<SIMD<double>> values,
+                                   BareSliceVector<> coefs) const override
+    {
+      throw ExceptionNOSIMD("HCurlCurl::AddTransIncShape not implemented yet");
+    }
+
+
+    
+    virtual void CalcDualShape (const BaseMappedIntegrationPoint & bmip, BareSliceMatrix<> shape) const override
+    {
+      shape.AddSize(ndof, sqr(bmip.DimSpace())) = 0.0;
+      Switch<4-DIM>
+        (bmip.DimSpace()-DIM,[this, &bmip, shape](auto CODIM)
+         {
+           auto & mip = static_cast<const MappedIntegrationPoint<DIM,DIM+CODIM.value>&> (bmip);
+
+           Cast() -> CalcDualShape2 (mip, SBLambda([shape] (size_t nr, auto val)
+                                                   {
+                                                     shape.Row(nr) = val.AsVector();
+                                                   }));
+         });
+    }
+
+    virtual void CalcDualShape (const SIMD_BaseMappedIntegrationRule& bmir, BareSliceMatrix<SIMD<double>> shapes) const override
+    {
+      Switch<4-DIM>
+        (bmir.DimSpace()-DIM,[this, &bmir, shapes](auto CODIM)
+         {
+           constexpr int DIMSPACE = DIM+CODIM.value;
+           auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM+CODIM.value>&> (bmir);
+
+           shapes.AddSize(ndof*sqr(DIMSPACE), mir.Size()) = 0.0;
+           for (size_t i = 0; i < mir.Size(); i++)
+             {
+               Cast() -> CalcDualShape2 (mir[i], SBLambda([shapes,i,DIMSPACE] (size_t j, auto val)
+                                                          {
+                                                            shapes.Rows(j*sqr(DIMSPACE), (j+1)*sqr(DIMSPACE)).Col(i).Range(0,sqr(DIMSPACE)) = val.AsVector();
+                                                          }));
+             }
+         });
+    }
+    
+    virtual void EvaluateDual (const SIMD_BaseMappedIntegrationRule & bmir, BareSliceVector<> coefs, BareSliceMatrix<SIMD<double>> values) const override
+    {
+      Switch<4-DIM>
+        (bmir.DimSpace()-DIM,[this,&bmir,coefs,values](auto CODIM)
+         {
+           constexpr int DIMSPACE = DIM+CODIM.value;
+           auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM+CODIM.value>&> (bmir);
+           for (size_t i = 0; i < mir.Size(); i++)
+             {
+               Mat<DIMSPACE,DIMSPACE,SIMD<double>> sum (SIMD<double>(0.0));
+               Cast() -> CalcDualShape2 (mir[i], SBLambda([&sum, coefs] (size_t j, auto val)
+                                                          {
+                                                            sum += coefs(j) * val;
+                                                          }));
+               for (size_t k = 0; k < sqr(DIMSPACE); k++)
+                 values(k, i) = sum(k);
+             }
+         });
+    }
+
+    virtual void AddDualTrans (const SIMD_BaseMappedIntegrationRule& bmir, BareSliceMatrix<SIMD<double>> values, BareSliceVector<double> coefs) const override
+    {
+      Switch<4-DIM>
+        (bmir.DimSpace()-DIM,[this,&bmir,coefs,values](auto CODIM)
+         {
+           constexpr int DIMSPACE = DIM+CODIM.value;
+           auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM+CODIM.value>&> (bmir);
+           for (size_t i = 0; i < mir.Size(); i++)
+             {
+               Mat<DIMSPACE,DIMSPACE,SIMD<double>> value;
+               for (size_t k = 0; k < sqr(DIMSPACE); k++)
+                 value(k) = values(k, i);
+               
+               Cast()-> CalcDualShape2 (mir[i], SBLambda([value, coefs] (size_t j, auto val)
+                                                         {
+                                                           coefs(j) += HSum(InnerProduct(val,value));
+                                                         }));
+             }
+         });
+    }
+
+    virtual void CalcMappedCurlShape (const BaseMappedIntegrationPoint & bmip,
+                                     BareSliceMatrix<double> shape) const override
+    {
+      auto mip = static_cast<const MappedIntegrationPoint<DIM,DIM> &>(bmip);
+
+      if constexpr (ET == ET_TET || ET == ET_TRIG)
+                     Cast() -> T_CalcShape (GetTIPHesse(mip),SBLambda([shape](int nr,auto val)
+                                                                       {
+                                                                         if constexpr (DIM==3)
+                                                                           shape.Row(nr).Range(DIM_DMAT) = val.Curl().AsVector();
+                                                                         else
+                                                                           shape.Row(nr).Range(DIM_DMAT) = val.Curl();
+                                                                       }));
+      else
+        throw Exception("HCurlCurl::CalcMappedCurlShape implemented only for TRIG and TET");
+    }
+      
+    virtual void CalcMappedShape (const SIMD_BaseMappedIntegrationRule & bmir, 
+                                         BareSliceMatrix<SIMD<double>> shapes) const override
+    {
+      Switch<4-DIM>
+        (bmir.DimSpace()-DIM,[this, &bmir, shapes](auto CODIM)
+         {
+           constexpr auto DIMSPACE = DIM+CODIM.value;
+           auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIM+CODIM.value>&> (bmir);
+           for (size_t i = 0; i < mir.Size(); i++)
+             {
+               this->Cast() -> T_CalcShape (GetTIP(mir[i]),
+                                             SBLambda ([i,shapes,DIMSPACE] (size_t j, auto val) 
+                                                       {
+                                                         shapes.Rows(j*sqr(DIMSPACE), (j+1)*sqr(DIMSPACE)).Col(i).Range(0,sqr(DIMSPACE)) = val.Value().AsVector();
+                                                       }));
+             }
+         });
+    }
+
+
+    virtual void Evaluate (const SIMD_BaseMappedIntegrationRule & bmir,
+                                  BareSliceVector<> coefs,
+                                  BareSliceMatrix<SIMD<double>> values) const override
+    {
+      Switch<4-DIM>
+        (bmir.DimSpace()-DIM,[values,&bmir, coefs,this](auto CODIM)
+       {
+         constexpr auto DIMSPACE = DIM+CODIM.value;
+         auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIMSPACE>&> (bmir);
+         for (size_t i = 0; i < bmir.Size(); i++)
+           {
+             double *pcoefs = &coefs(0);
+             const size_t dist = coefs.Dist();
+
+             Mat<DIMSPACE,DIMSPACE,SIMD<double>> summat(0);
+             Cast() -> T_CalcShape (GetTIP(mir[i]),
+                                    SBLambda ([&summat,&pcoefs,dist] (size_t j, auto val)
+                                              {
+                                                summat += (*pcoefs)*val.Value();
+                                                pcoefs += dist;
+                                              }));
+             for (size_t k = 0; k < sqr(DIMSPACE); k++)
+               values(k,i) = summat(k);              
+           }
+        });
+    }
+
+    virtual void AddTrans (const SIMD_BaseMappedIntegrationRule & bmir,
+                                  BareSliceMatrix<SIMD<double>> values,
+                                  BareSliceVector<> coefs) const override
+    {
+      Switch<4-DIM>
+      (bmir.DimSpace()-DIM,[this,&bmir,coefs,values](auto CODIM)
+       {
+         constexpr int DIMSPACE = DIM+CODIM.value;
+         auto & mir = static_cast<const SIMD_MappedIntegrationRule<DIM,DIMSPACE>&> (bmir);
+         for (size_t i = 0; i < mir.Size(); i++)
+           {
+             Mat<DIMSPACE,DIMSPACE,SIMD<double>> vali;
+             vali.AsVector()= values.Col(i);
+
+             double *pcoefs = &coefs(0);
+             const size_t dist = coefs.Dist();
+             Cast()->T_CalcShape (GetTIP(mir[i]),
+                                SBLambda ([vali,&pcoefs,dist] (size_t j, auto s)
+                                          {
+                                            *pcoefs += HSum(InnerProduct(s.Value(), vali));
+                                             pcoefs += dist;
+                                          }));
+           }
+       });
+      
+    }
+   
+  };
+
+
+
+  
+#ifdef FILE_HCURLCURLFE_CPP
+#define HCURLCURLFE_EXTERN
+#else
+#define HCURLCURLFE_EXTERN extern
+#endif
+  
+  HCURLCURLFE_EXTERN template class HCurlCurlFiniteElement<2>;
+  HCURLCURLFE_EXTERN template class HCurlCurlFiniteElement<3>;
+
+  template <> class HCurlCurlFE<ET_SEGM> : public T_HCurlCurlFE<ET_SEGM> 
+  {
+    
+  public:
+    using T_HCurlCurlFE<ET_SEGM> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+      ndof += order_inner[0]+1;
+      order = max2(order,order_inner[0]);
+
+    }
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<1,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x;
+      Tx lami[2] ={ x, 1-x };
+      int ii = 0;
+
+      IVec<2> e = ET_trait<ET_SEGM>::GetEdgeSort (0, vnums);
+      Tx ls = lami[e[0]], le = lami[e[1]];
+
+      auto symdyadic = MakeReggeAD(ls, le);
+
+      LegendrePolynomial::Eval(order_inner[0], ls-le, SBLambda([symdyadic, &ii, shape] (size_t nr, auto val)
+                            {
+                              shape[ii++] = 0.5*val*symdyadic;
+                            }));
+    }
+
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      auto & ip = mip.IP();
+      typedef typename std::remove_const<typename std::remove_reference<decltype(mip.IP()(0))>::type>::type T;    
+      T x = ip(0);
+      T lam[2] = { x, 1-x };
+
+      int ii = 0;
+
+      IVec<2> e = ET_trait<ET_SEGM>::GetEdgeSort (0, vnums);
+      T xi = lam[e[0]]-lam[e[1]];
+
+      auto tv = mip.GetJacobian()*Vec<1,T>(1);
+      auto tt = DyadProd(tv,tv);
+
+      LegendrePolynomial::Eval(order_inner[0], xi, SBLambda([shape,mip,tt,&ii] (size_t nr, T val)
+                            {
+                              shape[ii++] = 1/mip.GetMeasure()*val*tt;
+                            }));
+
+    }
+  };
+
+  
+  template <> class HCurlCurlFE<ET_TRIG> : public T_HCurlCurlFE<ET_TRIG> 
+  {
+    
+  public:
+    using T_HCurlCurlFE<ET_TRIG> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+      for (int i=0; i<3; i++)
+      {
+        ndof += order_facet[i][0]+1;
+        order = max2(order, order_facet[i][0]);
+      }
+      int ninner = 3*order_inner[0]*(order_inner[0]+1)/2 ;
+      order = max2(order, order_inner[0]);
+      
+      ndof += ninner;
+
+    }
+    
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<2,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x, y = ip.y;
+      Tx lami[3] ={ x, y, 1-x-y };
+      int ii = 0;
+
+      //    /*int maxorder_facet =
+      //      max2(order_facet[0][0],max2(order_facet[1][0],order_facet[2][0]));
+      //    ArrayMem<Tx,20> ha(maxorder_facet+1);
+      //    ArrayMem<Tx,20> u(order_inner[0]+2), v(order_inner[0]+2);
+      
+      //    for (int i = 0; i < 3; i++)
+      //      {
+      //        IVec<2> e = ET_trait<ET_TRIG>::GetEdgeSort(i,vnums);
+      //        Tx ls = llami[e[0]], le = llami[e[1]];
+      
+      //        // edge functions are all curl-free!
+      //        IntegratedLegendreMonomialExt::CalcTrigExt(maxorder_facet+2,
+      //                                                   le-ls, 1-le-ls, ha);
+      
+      //        for (int l = 0; l <= order_facet[i][0]; l++)
+      //          shape[ii++] = EpsGrad (ha[l]);
+      //          }*/
+      
+      for (int i = 0; i < 3; i++)
+        {
+          IVec<2> e = ET_trait<ET_TRIG>::GetEdgeSort (i, vnums);
+          Tx ls = lami[e[1]], le = lami[e[0]];
+
+          auto symdyadic = MakeReggeAD(ls, le);
+
+          LegendrePolynomial::EvalScaled(order_facet[i][0], ls-le,ls+le, SBLambda([symdyadic, &ii, shape] (size_t nr, auto val)
+                            {
+                              shape[ii++] = -val*symdyadic;
+                            }));
+        }
+
+
+      if (order_inner[0] > 0)
+        {
+	  IVec<4> f = ET_trait<ET_TRIG>::GetFaceSort(0, vnums); 
+	  Tx ls = lami[f[0]], le = lami[f[1]], lt = lami[f[2]];
+	  
+          auto symdyadic1 = lt*MakeReggeAD(ls, le);
+          auto symdyadic2 = ls*MakeReggeAD(lt, le);
+          auto symdyadic3 = le*MakeReggeAD(ls, lt);
+
+          
+	  DubinerBasis::Eval(order_inner[0]-1, ls,le,
+			      SBLambda([symdyadic1,symdyadic2,symdyadic3, &ii, shape] (size_t nr, auto val)
+				       {
+					 shape[ii++] = 2*val*symdyadic1;
+					 shape[ii++] = 2*val*symdyadic2;
+                                         shape[ii++] = 2*val*symdyadic3;
+				       }));
+	}
+      
+    };
+
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      auto & ip = mip.IP();
+      typedef typename std::remove_const<typename std::remove_reference<decltype(mip.IP()(0))>::type>::type T;    
+      T x = ip(0), y = ip(1);
+      T lam[3] = { x, y, 1-x-y };
+      Vec<2,T> pnts[3] = { { 1, 0 }, { 0, 1 } , { 0, 0 } };
+      int facetnr = ip.FacetNr();
+
+      int ii = 0;
+
+
+      if (ip.VB() == BND)
+        { // facet shapes
+          for (int i = 0; i < 3; i++)
+            {
+              int p = order_facet[i][0];
+              
+              if (i == facetnr)
+                {             
+                  IVec<2> e = ET_trait<ET_TRIG>::GetEdgeSort (i, vnums);
+                  
+                  T xi = lam[e[0]]-lam[e[1]];
+                  Vec<2,T> tauref = pnts[e[0]] - pnts[e[1]];
+                  
+                  
+                  auto tv = mip.GetJacobian()*tauref;
+
+                  auto tt = DyadProd(tv,tv);
+                  LegendrePolynomial::Eval
+                    (p, xi,
+                     SBLambda([&] (size_t nr, T val)
+                              {
+                                shape[nr+ii] = 1/mip.GetMeasure()*val*tt;
+                              }));
+                }
+              ii += (p+1);
+            }
+        }
+      else
+        {
+          for (int i = 0; i < 3; i++)
+            ii += order_facet[i][0]+1;
+        }
+      if (ip.VB() == VOL)
+        {
+          auto p = order_inner[0]-1;
+          if( p >= 0 )
+            {
+              IVec<4> f =  ET_trait<ET_TRIG>::GetFaceSort(0, vnums);
+
+              DubinerBasis::Eval (p, lam[f[0]], lam[f[1]],
+                                   SBLambda([&] (size_t nr, T val)
+                                            {
+                                              shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<2,2>({{1,0},{0,0}})*Trans(mip.GetJacobian());
+                                              shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<2,2>({{0,0},{0,1}})*Trans(mip.GetJacobian());
+                                              shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<2,2>({{0,1},{1,0}})*Trans(mip.GetJacobian());
+                                            }));
+            }
+        }
+    }
+
+    
+  };
+  
+  template <> class HCurlCurlFE<ET_QUAD> : public T_HCurlCurlFE<ET_QUAD> 
+  {
+    
+  public:
+    using T_HCurlCurlFE<ET_QUAD> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+      for (int i=0; i<4; i++)
+      {
+        ndof += order_facet[i][0]+1;
+        order = max2(order, order_facet[i][0]);
+      }
+      int ninner = order_inner[0]*order_inner[0] + (order_inner[0]+2)*order_inner[0]*2  +1;//+ 2*order_inner[0];
+      order = max2(order, order_inner[0]);
+      order += 1;
+      ndof += ninner;
+
+    }
+
+
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<2,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x, y = ip.y;
+      Tx lx[4] ={ 1-x, x, x, 1-x };
+      Tx ly[4] ={ 1-y, 1-y, y, y };
+      Tx lami[4] = {(1-x)*(1-y),x*(1-y),x*y,(1-x)*y};  
+      Tx sigma[4] = {(1-x)+(1-y),x+(1-y),x+y,(1-x)+y};  
+      int ii = 0;
+
+      ArrayMem<Tx,20> v(order+2), u(order+2);
+      
+     
+      for (int i = 0; i < 4; i++)
+        {
+          IVec<2> e = ET_trait<ET_QUAD>::GetEdgeSort (i, vnums);
+          Tx xi  = sigma[e[1]]-sigma[e[0]];
+          Tx lam_e = lami[e[0]]+lami[e[1]];  
+          auto symdyadic = MakeReggeAD(xi, xi);
+
+
+          //IntLegNoBubble::
+            LegendrePolynomial::
+            EvalMult (order_facet[i][0], 
+                      xi, 0.25*lam_e, SBLambda ([&](int i, auto val)
+                                           {
+                                             shape[ii++] = val*symdyadic;
+                                           }));
+        }
+
+
+
+      int oi = order_inner[0];
+
+      auto symdyadic = MakeReggeAD(0.5*x,0.5*y); //(0,0.5,  0.5,0) * P(y) * P(x)
+      
+      Tx eta = ly[2]-ly[1];
+      Tx xi = lx[1]-lx[0];
+      LegendrePolynomial (oi, eta, v);
+      LegendrePolynomial (oi, xi, u);
+
+      for (int i = 0; i <= oi; i++)
+        for (int j = 0; j <= oi; j++)
+          {
+            shape[ii++] = u[i]*v[j]*symdyadic;
+          }
+
+      
+      auto symdyad = lx[1]*lx[0]*MakeReggeAD(y,y);//x*(1-x)*(0,0,  0,1) * P(y) * P(x)
+      for (int i = 0; i < oi; i++)
+        for (int j = 0; j <= oi; j++)
+          {
+            shape[ii++] = u[i]*v[j]*symdyad;
+          }
+
+      symdyad = ly[2]*ly[1]*MakeReggeAD(x,x); //y*(1-y)*(1,0,  0,0) * P(x) * P(y)
+      
+      for (int j = 0; j < oi; j++)
+        for (int i = 0; i <= oi; i++)
+          {
+            shape[ii++] = u[i]*v[j]*symdyad;
+          }
+
+      //old version
+      //ArrayMem<Tx,20> u(order+2);
+      /*for (int i = 0; i < 4; i++)
+        {
+          IVec<2> e = ET_trait<ET_QUAD>::GetEdgeSort (i, vnums);
+          Tx xi = llx[e[1]]+lly[e[1]]-llx[e[0]]-lly[e[0]];
+          Tx eta = llx[e[0]]*lly[e[0]]+llx[e[1]]*lly[e[1]];
+
+	  IntegratedLegendreMonomialExt::Calc(order_facet[i][0]+2,xi,u);
+
+          
+          for (int l = 0; l <= order_facet[i][0]; l++)
+            shape[ii++] = Eps_u_Gradv (eta, u[l]);
+        }
+      
+      IntegratedLegendreMonomialExt::Calc(oi+3,llx[0]-llx[1],u);
+      IntegratedLegendreMonomialExt::Calc(oi+3,lly[0]-lly[2],v);
+
+      for(int i = 0; i <= oi-1; i++)
+        for(int j = 0; j <= oi-1; j++)
+          shape[ii++] = EpsGrad(u[i]*v[j]);
+
+      for(int i = 0; i <= oi+1; i++)
+        for(int j = 0; j <= oi-1; j++)
+        {
+          shape[ii++] = vEpsGradu(u[i],v[j]);
+          shape[ii++] = vEpsGradu(v[i],u[j]);
+        }
+      shape[ii++] = Eps_u_Gradv(lx[0], ly[0]);
+
+      for(int i = 0; i <= oi-1; i++)
+      {
+        shape[ii++] = Eps_u_Gradv(u[i], ly[0]);
+        shape[ii++] = Eps_u_Gradv(v[i], lx[0]);
+        }*/
+
+      
+      
+    };
+
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      auto & ip = mip.IP();
+      typedef typename std::remove_const<typename std::remove_reference<decltype(mip.IP()(0))>::type>::type T; 
+
+      T x = ip(0), y = ip(1);
+      T lx[4] = { 1-x, x, x, 1-x };
+      T ly[4] = { 1-y, 1-y, y, y };
+      T sigma[4] = {(1-x)+(1-y),x+(1-y),x+y,(1-x)+y};
+      
+      Vec<2,T> pnts[4] = {  { 0, 0 }, { 1, 0 }, { 1, 1 }, { 0, 1 } };
+      int facetnr = ip.FacetNr();
+
+      int ii = 0;
+      
+      ArrayMem<T,20> v(order+2), u(order+2);
+
+      
+      if (mip.IP().VB() == BND)
+        { // facet shapes
+          for (int i = 0; i < 4; i++)
+            {
+              int p = order_facet[i][0];
+              
+              if (i == facetnr)
+                {             
+                  IVec<2> e = ET_trait<ET_QUAD>::GetEdgeSort (i, vnums);
+                  
+                  //T xi = lam[e[0]]-lam[e[1]];
+                  T xi  = sigma[e[1]]-sigma[e[0]];
+                  Vec<2,T> tauref = pnts[e[0]] - pnts[e[1]];
+                  
+                  
+                  auto tv = mip.GetJacobian()*tauref;
+
+                  auto tt = DyadProd(tv,tv);
+                  LegendrePolynomial::Eval
+                    (p, xi,
+                     SBLambda([&] (size_t nr, T val)
+                              {
+                                shape[nr+ii] = 1/mip.GetMeasure()*val*tt;
+                                }));
+                  /*IVec<2> e = ET_trait<ET_QUAD>::GetEdgeSort (i, vnums);
+                  AutoDiff<2,T> xi  = sigma[e[1]]-sigma[e[0]];
+                  AutoDiff<2,T> lam_e = lami[e[0]]+lami[e[1]];  
+                  Vec<3, AutoDiff<2,T>> symdyadic = SymDyadProd(xi,xi);
+
+
+                  IntLegNoBubble::
+                    EvalMult (order_edge[i], 
+                              xi, lam_e, SBLambda ([&](int nr, auto val)
+                                                   {
+                                                     VecToSymMat<2>(T_REGGE_Shape<2,T>(val*symdyadic).Shape(),tmp);
+                                                     shape[nr + ii] = mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+                                                     }));*/
+                  /*AutoDiff<2,T> xi  = sigma[e[1]]-sigma[e[0]];
+                  AutoDiff<2,T> lam_e = lami[e[0]]+lami[e[1]];  
+                  Vec<3, AutoDiff<2,T>> symdyadic = SymDyadProd(xi,xi);
+
+
+                  IntLegNoBubble::
+                    EvalMult (p,xi, lam_e, SBLambda ([&](int nr, auto val)
+                                                   {
+                                                     VecToSymMat<2>(T_REGGE_Shape<2,T>(val*symdyadic).Shape(),tmp);
+                                                     shape[nr + ii] = 1/mip.GetMeasure()*tmp;
+                                                     }));*/
+                }
+              ii += (p+1);
+            }
+        }
+      else
+        {
+          for (int i = 0; i < 4; i++)
+            ii += order_facet[i][0]+1;
+        }
+      
+      if (mip.IP().VB() == VOL)
+        {
+          auto p = order_inner[0];
+         
+          T eta = ly[2]-ly[1];
+          T xi = lx[1]-lx[0];
+          LegendrePolynomial (p, eta, v);
+          LegendrePolynomial (p, xi, u);
+          
+          for (int i = 0; i <= p; i++)
+            for (int j = 0; j <= p; j++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{0,1},{1,0}}))*Trans(mip.GetJacobian());
+              }
+          
+          
+          //auto symdyad = lx[1]*lx[0]*SymDyadProd(Vec<2,T>(0,1),Vec<2,T>(0,1));//x*(1-x)*(0,0,  0,1) * P(y) * P(x)
+          for (int i = 0; i < p; i++)
+            for (int j = 0; j <= p; j++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{0,0},{0,1}}))*Trans(mip.GetJacobian());
+              }
+          
+          //symdyad = ly[2]*ly[1]*SymDyadProd(Vec<2,T>(1,0),Vec<2,T>(1,0)); //y*(1-y)*(1,0,  0,0) * P(x) * P(y)
+          
+          for (int j = 0; j < p; j++)
+            for (int i = 0; i <= p; i++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{1,0},{0,0}}))*Trans(mip.GetJacobian());
+              }
+      
+          //IVec<4> f = ET_trait<ET_QUAD>::GetFaceSort(0, vnums);
+          
+          /*IntegratedLegendreMonomialExt::Calc(p+3,lx[0]-lx[1],u);
+          IntegratedLegendreMonomialExt::Calc(p+3,ly[0]-ly[2],v);
+
+          Mat<2,2,T> tmp;
+          
+          for(int i = 0; i <= p-1; i++)
+              for(int j = 0; j <= p-1; j++)
+                {
+                  VecToSymMat<2>(EpsGrad(u[i]*v[j]).Shape(),tmp);
+                  shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+                }
+          
+          for(int i = 0; i <= p+1; i++)
+              for(int j = 0; j <= p-1; j++)
+                {
+                  VecToSymMat<2>(vEpsGradu(u[i],v[j]).Shape(),tmp);
+                  shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+                  VecToSymMat<2>(vEpsGradu(v[i],u[j]).Shape(),tmp);
+                  shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+                }
+
+          VecToSymMat<2>(Eps_u_Gradv(lx[0], ly[0]).Shape(),tmp);
+          shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+          
+          for(int i = 0; i <= p-1; i++)
+            {
+              VecToSymMat<2>(Eps_u_Gradv(u[i], ly[0]).Shape(),tmp);
+              shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+              VecToSymMat<2>(Eps_u_Gradv(v[i], lx[0]).Shape(),tmp);
+              shape[ii++] = 1/mip.GetMeasure()*mip.GetJacobian()*tmp*Trans(mip.GetJacobian());
+              }*/
+        }
+    }
+  };
+
+
+  template <> class HCurlCurlFE<ET_PRISM> : public T_HCurlCurlFE<ET_PRISM> 
+  {
+  public:
+    enum { incrorder_xx1 = 0};
+    enum { incrorder_zz1 = 0};
+    enum { incrorder_xx2 = 0};
+    enum { incrorder_zz2 = 0};
+    enum { incrorder_xx1_bd = 0};
+    enum { incrorder_zz1_bd = 0};
+    enum { incrorder_xx2_bd = 0};
+    enum { incrorder_zz2_bd = 0};
+    using T_HCurlCurlFE<ET_PRISM> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+
+      for (int i=0; i < 9; i++)
+        {
+          ndof += order_edge[i]+1;
+          order = max2(order,order_edge[i]);
+        }
+
+      for (int i=0; i<2; i++)
+        {
+          ndof += 3*(order_facet[i][0])*(order_facet[i][0]+1)/2;
+          order = max2(order, order_facet[i][0]);
+        }
+
+      for (int i=2; i<5; i++)
+        {
+          ndof += order_facet[i][0]*order_facet[i][0] + (order_facet[i][0]+2)*order_facet[i][0]*2 +1;
+          order = max2(order, order_facet[i][0]);
+        }
+      int p = order_inner[0];
+      int ninner =  3*p*(p+1)/2*p + (p-1)*(p)/2*(p+1) + (p+1)*p*(p+1);
+      ndof += ninner;
+
+      order = 1+max2(order, p);
+    }
+    
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<3,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x, y = ip.y, z = ip.z;
+      Tx lx[6] ={ x, y, 1-x-y, x, y, 1-x-y };
+      Tx lz[6] ={ 1-z,1-z,1-z,z,z,z };
+
+      int ii = 0;
+      
+
+      const FACE * faces = ElementTopology::GetFaces(ET_PRISM);
+
+      ArrayMem<Tx,20> leg_u(order+2), leg_v(order+3);
+      ArrayMem<Tx,20> leg_w(order+2);
+
+      //horizontal edge shapes
+      for (int i = 0; i < 6; i++)
+        {
+          IVec<2> e = ET_trait<ET_PRISM>::GetEdgeSort (i, vnums);
+          Tx ls = lx[e[1]], le = lx[e[0]], lm = lz[e[0]];
+
+          auto symdyadic = lm*MakeReggeAD(ls,le);
+
+          LegendrePolynomial::EvalScaled(order_edge[i], ls-le,ls+le, SBLambda([symdyadic, &ii, shape] (size_t nr, auto val)
+                            {
+                              shape[ii++] = -val*symdyadic;
+                            }));
+        }
+
+
+
+      //vertical edge shapes
+      for (int i = 6; i < 9; i++)
+        {
+          IVec<2> e = ET_trait<ET_PRISM>::GetEdgeSort (i, vnums);
+          Tx ls = lx[e[0]], lm1 = lz[e[0]], lm2 = lz[e[1]];
+          auto symdyadic = ls*MakeReggeAD(lm1,lm1);
+          LegendrePolynomial (order_edge[i],lm1-lm2, leg_v);
+
+          for (int j=0; j <= order_edge[i]; j++)
+            shape[ii++] = leg_v[j]*symdyadic;
+        }
+      
+
+
+      //horizontal face shaps
+      for(int fa = 0; fa < 2; fa++)
+        {
+          if (order_facet[fa][0] > 0)
+            {
+              IVec<4> f = ET_trait<ET_PRISM>::GetFaceSort(fa, vnums);
+              Tx ls = lx[f[0]], le = lx[f[1]], lt = lx[f[2]], lm = lz[f[0]];
+              
+              auto symdyadic1 = lm*lt*MakeReggeAD(ls,le);
+              auto symdyadic2 = lm*ls*MakeReggeAD(lt,le);
+              auto symdyadic3 = lm*le*MakeReggeAD(ls,lt);
+              
+              DubinerBasis::Eval(order_facet[fa][0]-1, ls,le,
+                                 SBLambda([symdyadic1,symdyadic2,symdyadic3, &ii, shape] (size_t nr, auto val)
+                                          {
+                                            shape[ii++] = val*symdyadic1;
+                                            shape[ii++] = val*symdyadic2;
+                                            shape[ii++] = val*symdyadic3;
+                                          }));
+            }
+        }
+
+
+      //vertical face shaps
+      for(int fa = 2; fa < 5; fa++)
+        {
+          int of = order_facet[fa][0];
+          
+          int fmax = 0;
+          for(int j = 1; j < 4; j++)
+            if(vnums[faces[fa][j]] > vnums[faces[fa][fmax]]) fmax = j;
+          int fz,ftrig;
+          fz = 3 - fmax;
+          ftrig = fmax^1;          
+          fmax = faces[fa][fmax];
+          fz = faces[fa][fz];
+          ftrig = faces[fa][ftrig];
+          
+          Tx eta = lz[fz]-lz[fmax];
+          Tx xi = lx[ftrig]-lx[fmax];
+
+          LegendrePolynomial (of, eta, leg_v);
+          LegendrePolynomial (of, xi, leg_u);
+
+          auto W = uDv_minus_vDu(lx[ftrig],lx[fmax]);
+          Tx W_AD;
+          W_AD.DValue(0) = W.Value()(0);
+          W_AD.DValue(1) = W.Value()(1);
+          W_AD.DValue(2) = W.Value()(2);
+          auto symdyadic = MakeReggeAD(eta,0.25*W_AD);   //^= (0,1, 1,0) * P(x)*P(y)
+          for (int j = 0; j <= of; j++)
+            for (int k = 0; k <= of; k++)
+              shape[ii++] = leg_v[j]*leg_u[k]*symdyadic;
+
+          
+          auto symdyad = 0.25*lx[ftrig]*lx[fmax]*MakeReggeAD(eta,eta);  //^= x*(1-x)*(0,0, 0,1) * P(x) * P(y)
+          for (int i = 0; i < of; i++)
+            for (int j = 0; j <= of; j++)
+                shape[ii++] = leg_u[i]*leg_v[j]*symdyad;
+
+          symdyad = 0.25*lz[fz]*lz[fmax]*MakeReggeAD(lx[ftrig],lx[fmax]);    //^= y*(1-y)*(1,0, 0,0) * P(x)*P(y)
+          for (int j = 0; j < of; j++)
+            for (int i = 0; i <= of; i++)
+                shape[ii++] = leg_u[i]*leg_v[j]*symdyad;
+        }
+      
+      //inner shapes
+      int p = order_inner[0];
+      if (p > 0)
+        {
+          
+          IVec<4> f = ET_trait<ET_PRISM>::GetFaceSort(0, vnums);
+
+          Tx ls = lx[f[0]], le = lx[f[1]], lt = lx[f[2]], lm = lz[0], ln = lz[3];
+
+          auto symdyadic1 = lm*ln*lt*MakeReggeAD(ls,le);
+          auto symdyadic2 = lm*ln*ls*MakeReggeAD(lt,le);
+          auto symdyadic3 = lm*ln*le*MakeReggeAD(ls,lt);
+
+          Tx eta = lz[0]-lz[4];
+          LegendrePolynomial (p, eta, leg_w);
+
+          // Reg(T) x [0,1]
+          DubinerBasis::Eval(p-1, ls,le,
+                             SBLambda([symdyadic1,symdyadic2,symdyadic3, &ii, shape,p,leg_w] (size_t nr, auto val)
+                                      {
+                                        for(int j=0; j < p; j++)
+                                          {
+                                            shape[ii++] = leg_w[j]*val*symdyadic1;
+                                            shape[ii++] = leg_w[j]*val*symdyadic2;
+                                            shape[ii++] = leg_w[j]*val*symdyadic3;
+                                          }
+                                      }));
+          
+
+          // H1(T) x [0,1]
+          auto symdyadic = ls*le*lt*MakeReggeAD(eta,eta);
+          DubinerBasis::Eval(p-2, ls,le,
+                             SBLambda([symdyadic, &ii, shape,p,leg_w] (size_t nr, auto val)
+                                      {
+                                        for(int j=0; j <= p; j++)
+                                          {
+                                            shape[ii++] = val*leg_w[j]*symdyadic;
+                                          }
+                                      }));
+
+
+          // Nedelec_1 x [0,1]
+          DubinerBasis::EvalMult(p-2, lx[f[0]], lx[f[1]],lx[f[0]]*lx[f[1]]*lx[f[2]], 
+                                 SBLambda([&](int nr, auto val)
+                                          {
+                                            auto tmp = Du(val);
+                                            Tx tmp_AD;
+                                            tmp_AD.DValue(0) = tmp.Value()(0);
+                                            tmp_AD.DValue(1) = tmp.Value()(1);
+                                            tmp_AD.DValue(2) = tmp.Value()(2);
+                                            auto symdyadic = MakeReggeAD(tmp_AD,eta);
+                                            for(int j=0; j <= p; j++)
+                                              shape[ii++] = leg_w[j]*symdyadic;
+                                          }));
+          
+          DubinerBasis::EvalMult(p-1, lx[f[0]], lx[f[1]], lx[f[0]], 
+                                 SBLambda([&ii,shape,p,leg_w,eta,f,lx](int nr, auto val)
+                                          {
+                                            auto tmp = wuDv_minus_wvDu (lx[f[1]], lx[f[2]], val);
+                                            Tx tmp_AD;
+                                            tmp_AD.DValue(0) = tmp.Value()(0);
+                                            tmp_AD.DValue(1) = tmp.Value()(1);
+                                            tmp_AD.DValue(2) = tmp.Value()(2);
+                                            auto symdyadic = MakeReggeAD(tmp_AD,eta);
+                                            for(int j=0; j <= p; j++)
+                                              shape[ii++] = leg_w[j]*symdyadic;
+                                          }));
+          
+          LegendrePolynomial::EvalMult(p-1, lx[f[2]]-lx[f[1]], lx[f[2]], 
+                                       SBLambda([&ii,shape,p,leg_w,eta,lx,f] (int j, auto val)
+                                                {
+                                                  auto tmp = wuDv_minus_wvDu (lx[f[1]], lx[f[0]], val);
+                                                  Tx tmp_AD;
+                                                  tmp_AD.DValue(0) = tmp.Value()(0);
+                                                  tmp_AD.DValue(1) = tmp.Value()(1);
+                                                  tmp_AD.DValue(2) = tmp.Value()(2);
+                                                  auto symdyadic = MakeReggeAD(tmp_AD,eta);
+                                                  for(int j=0; j <= p; j++)
+                                                    shape[ii++] = leg_w[j]*symdyadic;
+                                                }));
+          
+        }
+
+    }
+
+
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      throw Exception ("Hcurlcurlfe calcdualshape2 not implementend for element type ET_PRISM");
+    }
+
+  };
+
+
+  
+  template <> class HCurlCurlFE<ET_TET> : public T_HCurlCurlFE<ET_TET> 
+  {
+  public:
+    using T_HCurlCurlFE<ET_TET> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+
+      for (int i=0; i<6; i++)
+      {
+        ndof += order_edge[i]+1;
+        order = max2(order, order_edge[i]);
+      }
+      
+      for (int i=0; i<4; i++)
+      {
+        ndof += 3*(order_facet[i][0])*(order_facet[i][0]+1)/2;
+        order = max2(order, order_facet[i][0]);
+      }
+     
+      int p = order_inner[0];
+      int ninner = p > 1 ? 6*(p+1)*(p)*(p-1)/6 : 0;
+      ndof += ninner; 
+
+      order = max2(order, p);
+    }
+
+
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<3,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x, y = ip.y, z = ip.z;
+      Tx lam[4] = {x, y, z, 1-x-y-z};
+      int ii = 0;
+
+      for (int i = 0; i < 6; i++)
+        {
+          IVec<2> e = ET_trait<ET_TET>::GetEdgeSort (i, vnums);
+          Tx ls = lam[e[1]], le = lam[e[0]];
+
+          auto symdyadic = MakeReggeAD(ls, le);
+          LegendrePolynomial::EvalScaled(order_edge[i], ls-le,ls+le, SBLambda([symdyadic, &ii, shape] (size_t nr, auto val)
+                            {
+                              shape[ii++] = -val*symdyadic;
+                            }));
+        }
+
+      
+      for(int fa = 0; fa < 4; fa++)
+        {
+          if (order_facet[fa][0] > 0)
+            {
+              IVec<4> f = ET_trait<ET_TET>::GetFaceSort(fa, vnums);
+              Tx ls = lam[f[0]], le = lam[f[1]], lt = lam[f[2]];
+              
+              auto symdyadic1 = lt*MakeReggeAD(ls, le);
+              auto symdyadic2 = ls*MakeReggeAD(lt, le);
+              auto symdyadic3 = le*MakeReggeAD(ls, lt);
+              
+              DubinerBasis::Eval(order_facet[fa][0]-1, ls,le,
+                                 SBLambda([symdyadic1,symdyadic2,symdyadic3, &ii, shape] (size_t nr, auto val)
+                                          {
+                                            shape[ii++] = val*symdyadic1;
+                                            shape[ii++] = val*symdyadic2;
+                                            shape[ii++] = val*symdyadic3;
+                                          }));
+            }
+        }
+
+      if (order_inner[0] > 1)
+        {
+          Tx li = lam[0], lj = lam[1], lk = lam[2], ll = lam[3];
+
+          auto symdyadic1 = li*lj*MakeReggeAD(lk, ll);
+          auto symdyadic2 = lj*lk*MakeReggeAD(ll, li);
+          auto symdyadic3 = lk*ll*MakeReggeAD(li, lj);
+          auto symdyadic4 = ll*li*MakeReggeAD(lj, lk);
+          auto symdyadic5 = li*lk*MakeReggeAD(lj, ll);
+          auto symdyadic6 = lj*ll*MakeReggeAD(li, lk);
+
+          
+          DubinerBasis3D::Eval (order_inner[0]-2, lam[0], lam[1], lam[2], SBLambda([&ii, shape, symdyadic1, symdyadic2, symdyadic3, symdyadic4, symdyadic5, symdyadic6](size_t j, auto val)
+                                                                                   {
+                                                                                     shape[ii++] = val*symdyadic1;
+                                                                                     shape[ii++] = val*symdyadic2;
+                                                                                     shape[ii++] = val*symdyadic3;
+                                                                                     shape[ii++] = val*symdyadic4;
+                                                                                     shape[ii++] = val*symdyadic5;
+                                                                                     shape[ii++] = val*symdyadic6;
+                                                                                   }));
+        }
+      
+    }
+
+    
+
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      auto & ip = mip.IP();
+      typedef typename std::remove_const<typename std::remove_reference<decltype(mip.IP()(0))>::type>::type T;    
+      T x = ip(0), y = ip(1), z = ip(2);
+      T lam[4] = { x, y, z, 1-x-y-z };
+      Vec<3,T> pnts[4] = { { 1, 0, 0 }, { 0, 1, 0 }, { 0, 0, 1 } , { 0, 0, 0 } };
+      int facetnr = ip.FacetNr();
+
+      int ii = 0;
+
+      if (ip.VB() == BBND)
+        { // facet shapes
+          for (int i = 0; i < 6; i++)
+            {
+              int p = order_edge[i];
+              
+              if (i == facetnr)
+                {             
+                  IVec<2> e = ET_trait<ET_TET>::GetEdgeSort (i, vnums);
+                  
+                  T xi = lam[e[1]]-lam[e[0]];
+                  Vec<3,T> tauref = pnts[e[1]] - pnts[e[0]];
+                  Vec<3,T> tau = mip.GetJacobian()*tauref;
+                  Mat<3,3,T> tt = DyadProd(tau,tau);
+                  LegendrePolynomial::Eval
+                    (p, xi,
+                     SBLambda([&] (size_t nr, T val)
+                              {
+                                shape[nr+ii] = 1/mip.GetMeasure()*val*tt;
+                              }));
+                }
+              ii += (p+1);
+            }
+        }
+      else
+        {
+          for (int i = 0; i < 6; i++)
+            ii += order_edge[i]+1;
+        }
+      if (ip.VB() == BND)
+        {
+          for (int i = 0; i < 4; i++)
+            {
+              auto p = order_facet[i][0]-1;
+              if( p >= 0 && i == facetnr )
+                {
+                  IVec<4> fav = ET_trait<ET_TET>:: GetFaceSort(facetnr, vnums);
+                  Vec<3,T> adxi = pnts[fav[0]] - pnts[fav[2]];
+                  Vec<3,T> adeta = pnts[fav[1]] - pnts[fav[2]];
+                  T xi = lam[fav[0]];
+                  T eta = lam[fav[1]];
+                  
+                  Matrix<T> F(3,2);
+                  F.Col(0) = adxi;
+                  F.Col(1) = adeta;
+		 
+                  Matrix<T> Ftmp(2,2);
+                  Ftmp = Trans(F)*F;
+                  auto det = sqrt(Ftmp(0,0)*Ftmp(1,1)-Ftmp(1,0)*Ftmp(0,1));
+                                              
+                  DubinerBasis::Eval (p, xi, eta,
+                                       SBLambda([&] (size_t nr, T val)
+                                                {
+                                                  shape[ii++] = 1/(det*mip.GetMeasure())*val*Mat<3,3,T>(mip.GetJacobian()*F*Matrix<>({{1,0},{0,0}})*Trans(mip.GetJacobian()*F));
+                                                  shape[ii++] = 1/(det*mip.GetMeasure())*val*Mat<3,3,T>(mip.GetJacobian()*F*Matrix<>({{0,0},{0,1}})*Trans(mip.GetJacobian()*F));
+                                                  shape[ii++] = 1/(det*mip.GetMeasure())*val*Mat<3,3,T>(mip.GetJacobian()*F*Matrix<>({{0,1},{1,0}})*Trans(mip.GetJacobian()*F));
+                                                }));
+                }
+              else
+                ii += 3*(order_facet[i][0])*(order_facet[i][0]+1)/2;
+            }
+        }
+      else
+        {
+          for (int i = 0; i < 4; i++)
+            ii += 3*(order_facet[i][0])*(order_facet[i][0]+1)/2;
+        }
+      
+      if (ip.VB() == VOL && order_inner[0] >= 2)
+        {
+          DubinerBasis3D::Eval (order_inner[0]-2, lam[0], lam[1], lam[2], SBLambda([&](size_t j, T val)
+                               {
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{1,0,0},{0,0,0},{0,0,0}}))*Trans(mip.GetJacobian());
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{0,0,0},{0,1,0},{0,0,0}}))*Trans(mip.GetJacobian());
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{0,0,0},{0,0,0},{0,0,1}}))*Trans(mip.GetJacobian());
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{0,0,0},{0,0,1},{0,1,0}}))*Trans(mip.GetJacobian());
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{0,0,1},{0,0,0},{1,0,0}}))*Trans(mip.GetJacobian());
+                                 shape[ii++] = 1/mip.GetMeasure()*val*mip.GetJacobian()*Mat<3,3>(Matrix<>({{0,1,0},{1,0,0},{0,0,0}}))*Trans(mip.GetJacobian());
+                               }));
+          
+        }
+    }
+  };
+  
+
+
+  template <> class HCurlCurlFE<ET_HEX> : public T_HCurlCurlFE<ET_HEX> 
+  {
+  public:
+    using T_HCurlCurlFE<ET_HEX> :: T_HCurlCurlFE;
+
+    virtual void ComputeNDof()
+    {
+      order = 0;
+      ndof = 0;
+      for (int i=0; i < 12; i++)
+        {
+          ndof += order_edge[i]+1;
+          order = max2(order,order_edge[i]);
+        }
+      for (int i=0; i<6; i++)
+      {
+        ndof += order_facet[i][0]*order_facet[i][0] + 2*(order_facet[i][0]+2)*order_facet[i][0]+1;
+        order = max2(order, order_facet[i][0]);
+      }
+      int p = order_inner[0];
+      ndof += 3*(p*(p+1)*(p+1) + p*p*(p+1) );
+
+      order = 1 + max2(order, p);
+    }
+
+
+    template <typename Tx, typename TFA> 
+    void T_CalcShape (TIP<3,Tx> ip, TFA & shape) const
+    {
+      Tx x = ip.x, y = ip.y, z = ip.z;
+      Tx lx[2] ={ 1-x, x};
+      Tx ly[2] ={ 1-y, y};
+      Tx lz[2] ={ 1-z, z};
+      Tx lami[8]={(1-x)*(1-y)*(1-z),x*(1-y)*(1-z),x*y*(1-z),(1-x)*y*(1-z),
+                  (1-x)*(1-y)*z,x*(1-y)*z,x*y*z,(1-x)*y*z}; 
+      Tx sigma[8]={(1-x)+(1-y)+(1-z),x+(1-y)+(1-z),x+y+(1-z),(1-x)+y+(1-z),
+                   (1-x)+(1-y)+z,x+(1-y)+z,x+y+z,(1-x)+y+z};
+      int ii = 0;
+      
+      const FACE * faces = ElementTopology::GetFaces(ET_HEX);
+
+      ArrayMem<Tx,20> leg_u(order+2), leg_v(order+2), leg_w(order+2);
+      
+      // edges
+      for (int i = 0; i < 12; i++)
+        {
+          int p = order_edge[i]; 
+          IVec<2> e = ET_trait<ET_HEX>::GetEdgeSort (i, vnums);
+          Tx xi  = sigma[e[1]]-sigma[e[0]];
+          Tx lam_e = lami[e[0]]+lami[e[1]];
+          auto symdyadic = MakeReggeAD(xi,xi);
+          
+          //IntLegNoBubble::
+          LegendrePolynomial::
+            EvalMult (p, xi, 0.25*lam_e, SBLambda ([&](int i, auto val)
+                                              {
+                                                shape[ii++] = val*symdyadic;
+                                              }));
+        }
+      
+      
+      for (int i = 0; i<6; i++)
+        {
+          int p = order_facet[i][0];
+          
+          Tx lam_f(0);
+          for (int j = 0; j < 4; j++)
+            lam_f += lami[faces[i][j]];
+          
+          IVec<4> f = ET_trait<ET_HEX>::GetFaceSort (i, vnums);	  
+          Tx xi  = sigma[f[0]] - sigma[f[1]]; 
+          Tx eta = sigma[f[0]] - sigma[f[3]];
+          
+          LegendrePolynomial (p, eta, leg_u);
+          LegendrePolynomial (p, xi, leg_v);
+          
+          auto symdyadic = 0.25*lam_f*MakeReggeAD(eta,xi);
+          for (int j = 0; j <= p; j++)
+            for (int k = 0; k <= p; k++)
+              shape[ii++] = leg_u[j]*leg_v[k]*symdyadic;
+
+          symdyadic = 0.25*lam_f*(1-eta*eta)*MakeReggeAD(xi,xi);
+          for (int j = 0; j < p; j++)
+            for (int k = 0; k <= p; k++)
+              shape[ii++] = leg_u[j]*leg_v[k]*symdyadic;
+
+          symdyadic = 0.25*lam_f*(1-xi*xi)*MakeReggeAD(eta,eta);
+          for (int k = 0; k < p; k++)
+            for (int j = 0; j <= p; j++)
+              shape[ii++] = leg_u[j]*leg_v[k]*symdyadic;
+          
+        }
+
+      int p = order_inner[0];
+      if (p > 0)
+        {
+          Tx xi  = sigma[0] - sigma[1];
+          Tx eta = sigma[0] - sigma[3];
+          Tx nv = sigma[0] - sigma[4];
+          
+          LegendrePolynomial (p, xi,  leg_u);
+          LegendrePolynomial (p, eta, leg_v);
+          LegendrePolynomial (p, nv,  leg_w);
+
+          auto symdyadic1 = lz[0]*lz[1]*MakeReggeAD(eta,xi);
+          auto symdyadic2 = lx[0]*lx[1]*MakeReggeAD(nv,eta);
+          auto symdyadic3 = ly[0]*ly[1]*MakeReggeAD(xi,nv);
+          for (int i = 0; i <= p; i++)
+            for (int j = 0; j <= p; j++)
+              for (int k = 0; k < p; k++)
+                {
+                  shape[ii++] = leg_u[i]*leg_v[j]*leg_w[k]*symdyadic1;
+                  shape[ii++] = leg_v[i]*leg_w[j]*leg_u[k]*symdyadic2;
+                  shape[ii++] = leg_w[i]*leg_u[j]*leg_v[k]*symdyadic3;
+                }
+
+          symdyadic1 = ly[0]*ly[1]*lz[0]*lz[1]*MakeReggeAD(xi,xi);
+          symdyadic2 = lz[0]*lz[1]*lx[0]*lx[1]*MakeReggeAD(eta,eta);
+          symdyadic3 = lx[0]*lx[1]*ly[0]*ly[1]*MakeReggeAD(nv,nv);
+
+          for (int i = 0; i <= p; i++)
+            for (int j = 0; j < p; j++)
+              for (int k = 0; k < p; k++)
+                {
+                  shape[ii++] = leg_u[i]*leg_v[j]*leg_w[k]*symdyadic1;
+                  shape[ii++] = leg_v[i]*leg_w[j]*leg_u[k]*symdyadic2;
+                  shape[ii++] = leg_w[i]*leg_u[j]*leg_v[k]*symdyadic3;
+                }
+        }
+    }
+    
+    template <typename MIP, typename TFA>
+    void CalcDualShape2 (const MIP & mip, TFA & shape) const
+    {
+      auto & ip = mip.IP();
+      typedef typename std::remove_const<typename std::remove_reference<decltype(mip.IP()(0))>::type>::type T; 
+
+      T x = ip(0), y = ip(1), z = ip(2);
+      // T lx[4] = { 1-x, x, x, 1-x };
+      // T ly[4] = { 1-y, 1-y, y, y };
+      // T lz[4] = { 1-z, 1-z, z, z };
+      // T lam[4] = { 1-x-y+x*y, x*(1-y), x*y, y*(1-x) };
+      T sigma[8]={(1-x)+(1-y)+(1-z),x+(1-y)+(1-z),x+y+(1-z),(1-x)+y+(1-z),
+                   (1-x)+(1-y)+z,x+(1-y)+z,x+y+z,(1-x)+y+z};
+
+      /*Vec<2,AutoDiff<2,T>> adip = ip;
+      auto tip = TIP<2,AutoDiffDiff<2,T>>(adip);
+      AutoDiffDiff<2,T> xxx = tip.x, yyy = tip.y;
+      AutoDiff<2,T> xx(xxx.Value(), &xxx.DValue(0));
+      AutoDiff<2,T> yy(yyy.Value(), &yyy.DValue(0));
+      AutoDiff<2,T> lami[4] = {(1-xx)*(1-yy),xx*(1-yy),xx*yy,(1-xx)*yy};  
+      AutoDiff<2,T> sigma[4] = {(1-xx)+(1-yy),xx+(1-yy),xx+yy,(1-xx)+yy}; */ 
+      
+      Vec<3,T> pnts[8] = {  { 0, 0, 0 }, { 1, 0, 0 }, { 1, 1, 0 }, { 0, 1, 0 }, { 0, 0, 1 }, { 1, 0, 1 }, { 1, 1, 1 }, { 0, 1, 1 } };
+      int facetnr = ip.FacetNr();
+
+      int ii = 0;
+      
+      ArrayMem<T,20> v(order+2), u(order+2), w(order+2);
+
+      if (mip.IP().VB() == BBND)
+        { // edge shapes
+          for (int i = 0; i < 12; i++)
+            {
+              int p = order_edge[i];
+              
+              if (i == facetnr)
+                {             
+                  IVec<2> e = ET_trait<ET_HEX>::GetEdgeSort (i, vnums);
+                  
+                  T xi  = sigma[e[1]]-sigma[e[0]];
+                  Vec<3,T> tauref = pnts[e[0]] - pnts[e[1]];
+                  
+                  
+                  auto tv = mip.GetJacobian()*tauref;
+
+                  auto tt = DyadProd(tv,tv);
+                  LegendrePolynomial::Eval
+                    (p, xi,
+                     SBLambda([&] (size_t nr, T val)
+                              {
+                                shape[nr+ii] = 1/mip.GetMeasure()*val*tt;
+                                }));
+                 
+                }
+              ii += (p+1);
+            }
+        }
+      else
+        {
+          for (int i = 0; i < 12; i++)
+            ii += order_edge[i]+1;
+        }
+      if (mip.IP().VB() == BND)
+        {
+          for (int i = 0; i < 6; i++)
+            {
+              int p = order_facet[i][0];
+              
+              if (i == facetnr)
+                {
+                  IVec<4> f = ET_trait<ET_HEX>::GetFaceSort (i, vnums);	  
+                  Vec<3,T> tauref1 = pnts[f[0]] - pnts[f[1]];
+                  Vec<3,T> tauref2 = pnts[f[0]] - pnts[f[3]];
+                  T xi  = sigma[f[0]] - sigma[f[1]]; 
+                  T eta = sigma[f[0]] - sigma[f[3]];
+                  //Vec<6, T> symdyadic = SymDyadProd(GetGradient(etaa),GetGradient(xia));
+                  auto tv1 = mip.GetJacobian()*tauref1;
+                  auto tv2 = mip.GetJacobian()*tauref2;
+                  auto symdyadic = SymDyadProd(tv1,tv2);
+                  
+                  LegendrePolynomial (p, eta, u);
+                  LegendrePolynomial (p, xi, v);
+                  for (int j = 0; j <= p; j++)
+                    for (int k = 0; k <= p; k++)
+                      shape[ii + j*(p+1) + k] = u[j]*v[k]*symdyadic;
+                      
+                  /*                  T eta = ly[2]-ly[1];
+          T xi = lx[1]-lx[0];
+          LegendrePolynomial (p, eta, v);
+          LegendrePolynomial (p, xi, u);
+          
+          for (int i = 0; i <= p; i++)
+            for (int j = 0; j <= p; j++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{0,1},{1,0}}))*Trans(mip.GetJacobian());
+              }
+          
+          
+          //auto symdyad = lx[1]*lx[0]*SymDyadProd(Vec<2,T>(0,1),Vec<2,T>(0,1));//x*(1-x)*(0,0,  0,1) * P(y) * P(x)
+          for (int i = 0; i < p; i++)
+            for (int j = 0; j <= p; j++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{0,0},{0,1}}))*Trans(mip.GetJacobian());
+              }
+          
+          //symdyad = ly[2]*ly[1]*SymDyadProd(Vec<2,T>(1,0),Vec<2,T>(1,0)); //y*(1-y)*(1,0,  0,0) * P(x) * P(y)
+          
+          for (int j = 0; j < p; j++)
+            for (int i = 0; i <= p; i++)
+              {
+                shape[ii++] = 1/mip.GetMeasure()*u[i]*v[j]*mip.GetJacobian()*Mat<2,2>(Matrix<>({{1,0},{0,0}}))*Trans(mip.GetJacobian());
+                }*/
+                }
+              ii += p*p + (p+2)*p*2 + 1;
+            }
+        }
+      else 
+        {
+          for (int i = 0; i < 6; i++)
+            ii += order_facet[i][0]*order_facet[i][0] + (order_facet[i][0]+2)*order_facet[i][0]*2 + 1;
+        }
+
+      if (mip.IP().VB() == VOL)
+        {
+          if (order_inner[0])
+            throw Exception ("Hcurlcurlfe calcdualshape2 not implementend for element type ET_HEX for high-order");
+        }
+    }
+
+  };
+
+
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_SEGM>;
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_TRIG>;
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_QUAD>;
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_TET>;
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_PRISM>;
+  HCURLCURLFE_EXTERN template class T_HCurlCurlFE<ET_HEX>;
+}
+
+#endif