CUDA

Collaboration diagram for CUDA:

Macros
#define	native_host
	portable __host__ for CUDA
#define	native_device
	portable __device__ for CUDA
#define	native_hd   native_host native_device
	applies both native_host and native_device
#define	native_global
	portable __global__ for CUDA

Detailed Description

Macro Definition Documentation

native_device

#define native_device

portable __device__ for CUDA

indicates the function should be available on the device

Definition at line 1246 of file attributes.h.

native_global

#define native_global

portable __global__ for CUDA

indicates the function is a global entry point for a compute kernel

Definition at line 1247 of file attributes.h.

native_hd

#define native_hd   native_host native_device

applies both native_host and native_device

most functions in the library should be defined this way

may eventually add similar annotations for sycl and the like

Definition at line 1250 of file attributes.h.

native_host

#define native_host

portable __host__ for CUDA

indicates the function should be available on the host

Definition at line 1245 of file attributes.h.