diff options
author | Benoit Steiner <benoit.steiner.goog@gmail.com> | 2014-06-13 09:56:51 -0700 |
---|---|---|
committer | Benoit Steiner <benoit.steiner.goog@gmail.com> | 2014-06-13 09:56:51 -0700 |
commit | 38ab7e6ed0491bd5a0c639f218d5ea4728bf1e81 (patch) | |
tree | 9f74f100b406a629c29676000d9ef46b5f2e7536 /unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h | |
parent | aa664eabb912a1b96e417e9a8d9c98f423b7fc23 (diff) |
Reworked the expression evaluation mechanism in order to make it possible to efficiently compute convolutions and contractions in the future:
* The scheduling of computation is moved out the the assignment code and into a new TensorExecutor class
* The assignment itself is now a regular node on the expression tree
* The expression evaluators start by recursively evaluating all their subexpressions if needed
Diffstat (limited to 'unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h')
-rw-r--r-- | unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h | 9 |
1 files changed, 9 insertions, 0 deletions
diff --git a/unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h b/unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h index c4cfe0cd8..58b1808a3 100644 --- a/unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h +++ b/unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h @@ -153,6 +153,15 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr const Dimensions& dimensions() const { return m_dimensions; } + EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void evalSubExprsIfNeeded() { + m_inputImpl.evalSubExprsIfNeeded(); + m_kernelImpl.evalSubExprsIfNeeded(); + } + EIGEN_DEVICE_FUNC EIGEN_STRONG_INLINE void cleanup() { + m_inputImpl.cleanup(); + m_kernelImpl.cleanup(); + } + void evalTo(typename XprType::Scalar* buffer) const { for (int i = 0; i < dimensions().TotalSize(); ++i) { buffer[i] += coeff(i); |