diff options
author | Gael Guennebaud <g.gael@free.fr> | 2016-01-09 08:37:01 +0100 |
---|---|---|
committer | Gael Guennebaud <g.gael@free.fr> | 2016-01-09 08:37:01 +0100 |
commit | b557662e589a76265f73b99d7ca54a988d7bdb59 (patch) | |
tree | ad6a7118992e68f6038fff39898ff7c6e29df2e0 /unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h | |
parent | 8b9dc9f0dfb44c2c4ad6d02fb88ecce0986cd154 (diff) | |
parent | e76904af1b0f6aef80b68492534404f534ce0240 (diff) |
merge
Diffstat (limited to 'unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h')
-rw-r--r-- | unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h | 16 |
1 files changed, 8 insertions, 8 deletions
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h b/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h index 2ecdccdfd..fd7064459 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorReduction.h @@ -342,7 +342,7 @@ template <typename Self, typename Op, typename Device> struct InnerReducer { static const bool HasOptimizedImplementation = false; - static EIGEN_DEVICE_FUNC void run(const Self&, Op&, const Device&, typename Self::CoeffReturnType*, typename Self::Index, typename Self::Index) { + static void run(const Self&, Op&, const Device&, typename Self::CoeffReturnType*, typename Self::Index, typename Self::Index) { assert(false && "Not implemented"); } }; @@ -352,7 +352,7 @@ template <typename Self, typename Op, typename Device> struct OuterReducer { static const bool HasOptimizedImplementation = false; - static EIGEN_DEVICE_FUNC void run(const Self&, Op&, const Device&, typename Self::CoeffReturnType*, typename Self::Index, typename Self::Index) { + static void run(const Self&, Op&, const Device&, typename Self::CoeffReturnType*, typename Self::Index, typename Self::Index) { assert(false && "Not implemented"); } }; @@ -506,7 +506,7 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType>, Device> typedef typename internal::remove_const<typename XprType::CoeffReturnType>::type CoeffReturnType; typedef typename internal::remove_const<typename XprType::PacketReturnType>::type PacketReturnType; - EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE bool evalSubExprsIfNeeded(CoeffReturnType* data) { + EIGEN_STRONG_INLINE bool evalSubExprsIfNeeded(CoeffReturnType* data) { m_impl.evalSubExprsIfNeeded(NULL); // Use the FullReducer if possible. @@ -527,7 +527,7 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType>, Device> } // Attempt to use an optimized reduction. -#if defined(EIGEN_USE_GPU) && defined(__CUDACC__) +#if 0 else if (RunningOnGPU && data && (m_device.majorDeviceVersion() >= 3)) { bool reducing_inner_dims = true; for (int i = 0; i < NumReducedDims; ++i) { @@ -537,12 +537,12 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType>, Device> reducing_inner_dims &= m_reducedDims[NumInputDims - 1 - i]; } } - if (internal::InnerReducer<Self, Op, GpuDevice>::HasOptimizedImplementation && + if (internal::InnerReducer<Self, Op, Device>::HasOptimizedImplementation && (reducing_inner_dims || ReducingInnerMostDims)) { const Index num_values_to_reduce = internal::array_prod(m_reducedDims); const Index num_coeffs_to_preserve = internal::array_prod(m_dimensions); Op reducer(m_reducer); - internal::InnerReducer<Self, Op, GpuDevice>::run(*this, reducer, m_device, data, num_values_to_reduce, num_coeffs_to_preserve); + internal::InnerReducer<Self, Op, Device>::run(*this, reducer, m_device, data, num_values_to_reduce, num_coeffs_to_preserve); return false; } @@ -554,12 +554,12 @@ struct TensorEvaluator<const TensorReductionOp<Op, Dims, ArgType>, Device> preserving_inner_dims &= m_reducedDims[i]; } } - if (internal::OuterReducer<Self, Op, GpuDevice>::HasOptimizedImplementation && + if (internal::OuterReducer<Self, Op, Device>::HasOptimizedImplementation && preserving_inner_dims) { const Index num_values_to_reduce = internal::array_prod(m_reducedDims); const Index num_coeffs_to_preserve = internal::array_prod(m_dimensions); Op reducer(m_reducer); - internal::OuterReducer<Self, Op, GpuDevice>::run(*this, reducer, m_device, data, num_values_to_reduce, num_coeffs_to_preserve); + internal::OuterReducer<Self, Op, Device>::run(*this, reducer, m_device, data, num_values_to_reduce, num_coeffs_to_preserve); return false; } } |