static void execMetaPredicateStrided(cudaStream_t * stream, Nd4jPointer *extras, const int opTypeA, const int opNumA, const int opTypeB, const int opNumB, Nd4jLong N, T *dx, Nd4jLong xStride, T *dy, Nd4jLong yStride, T *dz, Nd4jLong zStride, T *extraA, T *extraB, T scalarA, T scalarB);
template<typename OpType>
static __device__ void transformCuda(Nd4jLong n, T *dx, T *dy, Nd4jLong incx, Nd4jLong incy, T *params, T *result, Nd4jLong incz,int *allocationPointer, UnifiedSharedMemory *manager, Nd4jLong *tadOnlyShapeInfo);
static __device__ void transformCuda(const int opTypeA, const int opNumA, const int opTypeB, const int opNumB, Nd4jLong n, T *dx, T *dy, Nd4jLong incx, Nd4jLong incy, T *params, T *result, Nd4jLong incz,int *allocationPointer, UnifiedSharedMemory *manager, Nd4jLong *tadOnlyShapeInfo);