First, load the CUDALink application:

This loads memory:

Information about memory can be retrieved using CUDAMemoryInformation:

Memory added must be freed with CUDAMemoryUnload:

Adding memory as Real or Complex gets the type based on whether the device supports double precision or not:

In this case, the CUDA device has double-precision support:

The behavior can be forced to change by setting the "TargetPrecision":

Memory added must be freed with CUDAMemoryUnload:

Images can be added with type "UnsignedByte":

Getting the memory returns an image with the same properties as the original. Memory is retrieved using CUDAMemoryGet:

The "TypeInformation" contains the image information:

Memory added must be freed with CUDAMemoryUnload:

Images can be added with specified type:

Getting the memory returns an image:

When adding graphics objects, the object is rasterized:

When getting the memory, an image is returned:

Memory added must be freed with CUDAMemoryUnload:

This adds two to an input list:

Internally, running a CUDAFunction will load and (if the input is not CUDAMemory) will unload the memory. To demonstrate, this color negates an input image:

This loads the function using CUDAFunctionLoad:

This defines input parameters:

This runs the CUDAFunction: