Vectorized the evaluation of tensor expression (using SSE, AVX, NEON, ...)

Added the ability to parallelize the evaluation of a tensor expression over multiple cpu cores.
Added the ability to offload the evaluation of a tensor expression to a GPU.

1 files changed, 7 insertions, 1 deletions

diff --git a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h
index f5c027d1c..d8ff3f584 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/Tensor.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/Tensor.h
@@ -75,9 +75,15 @@ class Tensor : public TensorBase<Tensor<Scalar_, NumIndices_, Options_> >
     typedef typename internal::traits<Self>::StorageKind StorageKind;
     typedef typename internal::traits<Self>::Index Index;
     typedef Scalar_ Scalar;
-    typedef typename internal::packet_traits<Scalar>::type PacketScalar;
+    typedef typename internal::packet_traits<Scalar>::type Packet;
     typedef typename NumTraits<Scalar>::Real RealScalar;
     typedef typename Base::CoeffReturnType CoeffReturnType;
+    typedef typename Base::PacketReturnType PacketReturnType;
+
+    enum {
+      IsAligned = bool(EIGEN_ALIGN),
+      PacketAccess = true,
+    };
 
     static const int Options = Options_;
     static const std::size_t NumIndices = NumIndices_;