aboutsummaryrefslogtreecommitdiffhomepage
path: root/unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h
diff options
context:
space:
mode:
authorGravatar Deven Desai <deven.desai.amd@gmail.com>2018-07-17 14:16:48 -0400
committerGravatar Deven Desai <deven.desai.amd@gmail.com>2018-07-17 14:16:48 -0400
commitf124f0796533081cb7b061c4aa8667df6ed58863 (patch)
tree20efd29fdfaf3da79cad39f468d330927b62ebc2 /unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h
parent82f0ce27261df3b21037d93d4595655b3df754a6 (diff)
applying EIGEN_DECLARE_TEST to *gpu* tests
Also, a few minor fixes for GPU tests running in HIP mode. 1. Adding an include for hip/hip_runtime.h in the Macros.h file For HIP __host__ and __device__ are macros which are defined in hip headers. Their definitions need to be included before their use in the file. 2. Fixing the compile failure in TensorContractionGpu introduced by the commit to "Fuse computations into the Tensor contractions using output kernel" 3. Fixing a HIP/clang specific compile error by making the struct-member assignment explicit
Diffstat (limited to 'unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h')
-rw-r--r--unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h10
1 files changed, 5 insertions, 5 deletions
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h
index a4f92ee44..b9956cd43 100644
--- a/unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h
+++ b/unsupported/Eigen/CXX11/src/Tensor/TensorContractionGpu.h
@@ -1215,16 +1215,16 @@ EigenFloatContractionKernel16x16(const LhsMapper lhs, const RhsMapper rhs,
}
-template<typename Indices, typename LeftArgType, typename RightArgType>
-struct TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType>, GpuDevice> :
- public TensorContractionEvaluatorBase<TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType>, GpuDevice> > {
+template<typename Indices, typename LeftArgType, typename RightArgType, typename OutputKernelType>
+struct TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType, OutputKernelType>, GpuDevice> :
+ public TensorContractionEvaluatorBase<TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType, OutputKernelType>, GpuDevice> > {
typedef GpuDevice Device;
- typedef TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType>, Device> Self;
+ typedef TensorEvaluator<const TensorContractionOp<Indices, LeftArgType, RightArgType, OutputKernelType>, Device> Self;
typedef TensorContractionEvaluatorBase<Self> Base;
- typedef TensorContractionOp<Indices, LeftArgType, RightArgType> XprType;
+ typedef TensorContractionOp<Indices, LeftArgType, RightArgType, OutputKernelType> XprType;
typedef typename internal::remove_const<typename XprType::Scalar>::type Scalar;
typedef typename XprType::Index Index;
typedef typename XprType::CoeffReturnType CoeffReturnType;