fill_bytes#

template<typename _DstTy, ::cuda::std::enable_if_t<__spannable<transformed_device_argument_t<_DstTy>>, int> = 0>
inline void cuda::fill_bytes(
stream_ref __stream,
_DstTy &&__dst,
::cuda::std::uint8_t __value
)#

Launches an operation to bytewise fill the memory into the provided stream.

The destination needs to be or launch_transform to a contiguous_range and convert to cuda::std::span. The element type of the destination is required to be trivially copyable.

The destination cannot reside in pagable host memory.

Destination needs to be or launch_transform to an instance of cuda::std::mdspan. It can also convert to cuda::std::mdspan, but the type needs to contain mdspan template arguments as member aliases named value_type, extents_type, layout_type and accessor_type. The resulting mdspan is required to be exhaustive. The element type of the destination is required to be trivially copyable.

The destination cannot reside in pagable host memory.

Parameters:
  • __stream – Stream that the copy should be inserted into

  • __dst – Destination memory to fill

  • __value – Value to fill into every byte in the destination

  • __stream – Stream that the copy should be inserted into

  • __dst – Destination memory to fill

  • __value – Value to fill into every byte in the destination