CUDA:UnifiedMemory
Unified Memory is a component of the CUDA programming model, first introduced in CUDA 6.0, that defines a managed memory space in which all processors see a single coherent memory image with a common address space.
Device Properties
Unified Memory is supported only on devices with compute capability 3.0 or higher. A program may query whether a GPU device supports managed memory by using cudaGetDeviceProperties() and checking the new managedMemory property. The capability can also be determined using the individual attribute query function cudaDeviceGetAttribute() with the attribute cudaDevAttrManagedMemory.