diff options
| author |  Deven Desai <deven.desai.amd@gmail.com> | 2018-06-06 10:12:58 -0400 |
|---|---|---|
| committer | Deven Desai <deven.desai.amd@gmail.com> | 2018-06-06 10:12:58 -0400 |
| commit | 8fbd47052bcafea612b8ae2841c1de5db738f042 (patch) | |
| tree | 1e8d3f8ab0bc9e48e18b0502b7d51500e72a7266 /Eigen/src/Core/ProductEvaluators.h | |
| parent | e206f8d4a401fe2060bada4d4b5d92e3bf3b561c (diff) | |
Adding support for using Eigen in HIP kernels.
This commit enables the use of Eigen on HIP kernels / AMD GPUs. Support has been added along the same lines as what already exists for using Eigen in CUDA kernels / NVidia GPUs.
Application code needs to explicitly define EIGEN_USE_HIP when using Eigen in HIP kernels. This is because some of the CUDA headers get picked up by default during Eigen compile (irrespective of whether or not the underlying compiler is CUDACC/NVCC, for e.g. Eigen/src/Core/arch/CUDA/Half.h). In order to maintain this behavior, the EIGEN_USE_HIP macro is used to switch to using the HIP version of those header files (see Eigen/Core and unsupported/Eigen/CXX11/Tensor)
Use the "-DEIGEN_TEST_HIP" cmake option to enable the HIP specific unit tests.
Diffstat (limited to 'Eigen/src/Core/ProductEvaluators.h')
| -rw-r--r-- | Eigen/src/Core/ProductEvaluators.h | 6 |
1 files changed, 6 insertions, 0 deletions
diff --git a/Eigen/src/Core/ProductEvaluators.h b/Eigen/src/Core/ProductEvaluators.h index fb637191d..cc75fbce3 100644 --- a/Eigen/src/Core/ProductEvaluators.h +++ b/Eigen/src/Core/ProductEvaluators.h @@ -137,6 +137,9 @@ struct Assignment<DstXprType, Product<Lhs,Rhs,Options>, internal::assign_op<Scal typename enable_if<(Options==DefaultProduct || Options==AliasFreeProduct)>::type> { typedef Product<Lhs,Rhs,Options> SrcXprType; + #if defined(EIGEN_HIPCC) + EIGEN_DEVICE_FUNC + #endif static EIGEN_STRONG_INLINE void run(DstXprType &dst, const SrcXprType &src, const internal::assign_op<Scalar,Scalar> &) { @@ -390,6 +393,9 @@ struct generic_product_impl<Lhs,Rhs,DenseShape,DenseShape,CoeffBasedProductMode> typedef typename Product<Lhs,Rhs>::Scalar Scalar; template<typename Dst> + #if defined(EIGEN_HIPCC) + EIGEN_DEVICE_FUNC + #endif static EIGEN_STRONG_INLINE void evalTo(Dst& dst, const Lhs& lhs, const Rhs& rhs) { // Same as: dst.noalias() = lhs.lazyProduct(rhs); |