devtools/eigen
1
0
Fork 0
mirror of https://gitlab.com/libeigen/eigen.git synced 2026-04-10 11:34:33 +08:00
Code Issues Packages Projects Releases Wiki Activity

Enabled the vectorized evaluation of several tensor expressions that was previously disabled by mistake

Browse Source
This commit is contained in:
Benoit Steiner
2015-07-01 11:32:04 -07:00
parent 44eedd8915
commit 925d0d375a
11 changed files with 42 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
unsupported/Eigen/CXX11/src/Tensor/TensorConvolution.h
View File

@@ -510,7 +510,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
Scalar* local = (Scalar*)m_device.allocate(kernel_sz);
typedef TensorEvalToOp<const KernelArgType> EvalTo;
EvalTo evalToTmp(local, m_kernelArg);
internal::TensorExecutor<const EvalTo, Device, TensorEvaluator<KernelArgType, Device>::PacketAccess>::run(evalToTmp, m_device);
internal::TensorExecutor<const EvalTo, Device>::run(evalToTmp, m_device);
m_kernel = local;
m_local_kernel = true;
@@ -815,7 +815,7 @@ struct TensorEvaluator<const TensorConvolutionOp<Indices, InputArgType, KernelAr
Scalar* local = (Scalar*)m_device.allocate(kernel_sz);
typedef TensorEvalToOp<const KernelArgType> EvalTo;
EvalTo evalToTmp(local, m_kernelArg);
internal::TensorExecutor<const EvalTo, GpuDevice, TensorEvaluator<KernelArgType, GpuDevice>::PacketAccess>::run(evalToTmp, m_device);
internal::TensorExecutor<const EvalTo, GpuDevice>::run(evalToTmp, m_device);
m_kernel = local;
m_local_kernel = true;