Specialisers

Specialisation operators. More...

Classes
struct	sp::BroadcastingTensorSpecialiser
	Set the BroadcastsLastDimension flag on BroadcastingTensor, if possible. More...
struct	sp::PtrScalarSpecialiser< ForceLoc >
	Specialiser for coping with scalar parameters that are provided as pointers: something APIs love to do. More...
struct	sp::ScalarSpecialiser
	Specialises input scalar types to compile-time-constant sp::Scalar derivatives, if possible. More...
struct	sp::TensorSpecialiser
	Attempt to increase the strictness of the optimisation flags of all Tensor arguments. More...
struct	sp::TensorDescriptorSpecialiser
	Convert a TensorDescriptor into a Tensor by finding the type/rank at runtime. More...
struct	sp::TensorFlattenSpecialiser
	A non-default Specialiser that flattens N-dimensional packed tensors into 1-dimensional tensors that cover the same memory. More...
struct	sp::TensorFlatteningTrampoline< T >
	A trampoline struct you can use as a Specialiser target to easily chain TensorFlatten with the default specialisers. More...
struct	sp::VariantOutputPtrSpecialiser
	Convert a VariantOutputPtr to a strongly-typed pointer. More...
struct	sp::VariantScalarSpecialiser
	Specialise CUDA API variant scalars into proper scalars. More...

Detailed Description

Specialisation operators.

Each one provides a way of branching on some kind of input argument to pick different template instantiations of the target Specialisable. The interface for specialisers is defined in Specialiser.