vt::TensorIterator

template<typename Iterator, size_t N>
class TensorIterator : public thrust::iterator_adaptor<TensorIterator<Iterator, N>, Iterator>

TensorIterator class to represent an iterator of a tensor.

Template Parameters:

• Iterator – The raw iterator.

• N – Number of dimensions of the tensor.

Public Functions

inline __host__ __device__ TensorIterator (const Iterator &iter, const size_t *shape, const size_t *strides, const size_t start, const Order order, const bool contiguous)

Construct a new tensor Iterator object.

Parameters:

• iter – The raw iterator from thrust.

• shape – The shape of the tensor.

• strides – The strides of the tensor.

• start – The start index of the tensor.

• order – The order of the tensor.

• contiguous – The contiguous flag of the tensor.

inline __host__ __device__ void advance (typename thrust::iterator_difference< Iterator >::type n)

Advance the iterator by n steps. If the iterator is contiguous, used the default advance method, otherwise calculate the new index based on the strides and shape.

Parameters:

n – Number of steps.