TensorWrap shortcuts

group NVCV_CPP_CUDATOOLS_TENSORWRAPS

Specializes TensorWrap template classes to different dimensions.

The specializations have the last dimension as the only compile-time dimension as size of T. All other dimensions have run-time pitch and must be provided.

Template arguments:

• T data type of each element in TensorWrap

• N (optional) number of dimensions

See also

TensorWrap classes

Typedefs

template<typename T>
using Tensor1DWrap = TensorWrap<T, sizeof(T)>

template<typename T>
using Tensor2DWrap = TensorWrap<T, -1, sizeof(T)>

template<typename T>
using Tensor3DWrap = TensorWrap<T, -1, -1, sizeof(T)>

template<typename T>
using Tensor4DWrap = TensorWrap<T, -1, -1, -1, sizeof(T)>

template<typename T>
using Tensor5DWrap = TensorWrap<T, -1, -1, -1, -1, sizeof(T)>

template<typename T, int N>
using TensorNDWrap = std::conditional_t<N == 1, Tensor1DWrap<T>, std::conditional_t<N == 2, Tensor2DWrap<T>, std::conditional_t<N == 3, Tensor3DWrap<T>, std::conditional_t<N == 4, Tensor4DWrap<T>, std::conditional_t<N == 5, Tensor5DWrap<T>, void>>>>>