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.
This commit is contained in:
Benoit Steiner
2014-05-16 15:08:05 -07:00
parent 0320f7e3a7
commit 7402fea0a8
17 changed files with 719 additions and 65 deletions

View File

@@ -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_;