Hey, finally the copyCoeff stuff is not only used to implement swap anymore :)

Add an internal pseudo expression allowing to optimize operators like +=, *= using
the copyCoeff stuff.
This allows to easily enforce aligned load for the destination matrix everywhere.

1 files changed, 113 insertions, 0 deletions

diff --git a/Eigen/src/Core/SelfCwiseBinaryOp.h b/Eigen/src/Core/SelfCwiseBinaryOp.h
new file mode 100644
index 000000000..ac0ccd963
--- /dev/null
+++ b/Eigen/src/Core/SelfCwiseBinaryOp.h
@@ -0,0 +1,113 @@
+// This file is part of Eigen, a lightweight C++ template library
+// for linear algebra.
+//
+// Copyright (C) 2009 Gael Guennebaud <g.gael@free.fr>
+//
+// Eigen is free software; you can redistribute it and/or
+// modify it under the terms of the GNU Lesser General Public
+// License as published by the Free Software Foundation; either
+// version 3 of the License, or (at your option) any later version.
+//
+// Alternatively, you can redistribute it and/or
+// modify it under the terms of the GNU General Public License as
+// published by the Free Software Foundation; either version 2 of
+// the License, or (at your option) any later version.
+//
+// Eigen is distributed in the hope that it will be useful, but WITHOUT ANY
+// WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
+// FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License or the
+// GNU General Public License for more details.
+//
+// You should have received a copy of the GNU Lesser General Public
+// License and a copy of the GNU General Public License along with
+// Eigen. If not, see <http://www.gnu.org/licenses/>.
+
+#ifndef EIGEN_SELFCWISEBINARYOP_H
+#define EIGEN_SELFCWISEBINARYOP_H
+
+/** \class SelfCwiseBinaryOp
+  *
+  * \internal
+  *
+  * \brief Internal helper class for optimizing operators like +=, -=
+  */
+template<typename BinaryOp, typename MatrixType>
+struct ei_traits<SelfCwiseBinaryOp<BinaryOp,MatrixType> > : ei_traits<MatrixType> {};
+
+template<typename BinaryOp, typename MatrixType> class SelfCwiseBinaryOp
+  : public MatrixBase<SelfCwiseBinaryOp<BinaryOp,MatrixType> >
+{
+  public:
+
+    EIGEN_GENERIC_PUBLIC_INTERFACE(SelfCwiseBinaryOp)
+    typedef typename ei_packet_traits<Scalar>::type Packet;
+
+    using Base::operator=;
+
+    inline SelfCwiseBinaryOp(MatrixType& xpr, const BinaryOp& func = BinaryOp()) : m_matrix(xpr), m_functor(func) {}
+
+    inline int rows() const { return m_matrix.rows(); }
+    inline int cols() const { return m_matrix.cols(); }
+    inline int stride() const { return m_matrix.stride(); }
+
+    // note that this function is needed by assign to correctly align loads/stores
+    // TODO make Assign use .data()
+    inline Scalar& coeffRef(int row, int col)
+    {
+      return m_matrix.const_cast_derived().coeffRef(row, col);
+    }
+
+    // note that this function is needed by assign to correctly align loads/stores
+    // TODO make Assign use .data()
+    inline Scalar& coeffRef(int index)
+    {
+      return m_matrix.const_cast_derived().coeffRef(index);
+    }
+
+    template<typename OtherDerived>
+    void copyCoeff(int row, int col, const MatrixBase<OtherDerived>& other)
+    {
+      OtherDerived& _other = other.const_cast_derived();
+      ei_internal_assert(row >= 0 && row < rows()
+                         && col >= 0 && col < cols());
+      Scalar& tmp = m_matrix.coeffRef(row,col);
+      tmp = m_functor(tmp, _other.coeff(row,col));
+    }
+
+    template<typename OtherDerived>
+    void copyCoeff(int index, const MatrixBase<OtherDerived>& other)
+    {
+      OtherDerived& _other = other.const_cast_derived();
+      ei_internal_assert(index >= 0 && index < m_matrix.size());
+      Scalar& tmp = m_matrix.coeffRef(index);
+      tmp = m_functor(tmp, _other.coeff(index));
+    }
+
+    template<typename OtherDerived, int StoreMode, int LoadMode>
+    void copyPacket(int row, int col, const MatrixBase<OtherDerived>& other)
+    {
+      OtherDerived& _other = other.const_cast_derived();
+      ei_internal_assert(row >= 0 && row < rows()
+                        && col >= 0 && col < cols());
+      m_matrix.template writePacket<StoreMode>(row, col,
+        m_functor.packetOp(m_matrix.template packet<StoreMode>(row, col),_other.template packet<LoadMode>(row, col)) );
+    }
+
+    template<typename OtherDerived, int StoreMode, int LoadMode>
+    void copyPacket(int index, const MatrixBase<OtherDerived>& other)
+    {
+      OtherDerived& _other = other.const_cast_derived();
+      ei_internal_assert(index >= 0 && index < m_matrix.size());
+      m_matrix.template writePacket<StoreMode>(index,
+        m_functor.packetOp(m_matrix.template packet<StoreMode>(index),_other.template packet<LoadMode>(index)) );
+    }
+
+  protected:
+    MatrixType& m_matrix;
+    const BinaryOp& m_functor;
+
+  private:
+    SelfCwiseBinaryOp& operator=(const SelfCwiseBinaryOp&);
+};
+
+#endif // EIGEN_SELFCWISEBINARYOP_H