|
CUDPP 2.0
CUDA Data-Parallel Primitives Library
|
Wrapper class for templatized dynamic shared memory arrays. More...
#include <sharedmem.h>
Public Member Functions | |
| __device__ T * | getPointer () |
| Return a pointer to the runtime-sized shared memory array. | |
Wrapper class for templatized dynamic shared memory arrays.
This struct uses template specialization on the type T to declare a differently named dynamic shared memory array for each type (
extern __shared__ T s_type[]
).
Currently there are specializations for the following types: int, uint, char, uchar, short, ushort, long, unsigned long, bool, float, and double. One can also specialize it for user defined types.
| __device__ T* SharedMemory< T >::getPointer | ( | ) | [inline] |
Return a pointer to the runtime-sized shared memory array.
1.7.4