cuda kernel函数:
//16bit 对应 ushort
//一般是8bit 对应char
__global__ void fun(cuda::PtrStepSz<ushort> matInCuda){
int tx = threadIdx.x + blockDim.x * blockIdx.x;
int ty = threadIdx
cuda kernel函数:
//16bit 对应 ushort
//一般是8bit 对应char
__global__ void fun(cuda::PtrStepSz<ushort> matInCuda){
int tx = threadIdx.x + blockDim.x * blockIdx.x;
int ty = threadIdx