cuda对global memery的双调排序

__device__ void swap(int& a, int& b) {
    int t = a;
    a = b;
    b = t;
}

__global__ void sort(int* a, int flag_j, int flag_i, int count)
{
    unsigned int tid = blockIdx.x * blockDim.x + threadIdx.x;
    unsigned int tid_comp = tid ^ flag_j;

    if (tid > count)
        return;


    if (tid_comp > tid) {
        if ((tid & flag_i) == 0) { //ascending
            if (a[tid] > a[tid_comp]) {
                swap(a[tid], a[tid_comp]);
            }
        }
        else { //desending
            if (a[tid] < a[tid_comp]) {
                swap(a[tid], a[tid_comp]);
            }
        }
    }

}

int main() 
{


    int count = 800;
    std::default_random_engine e;

    int* a, * b;
    cudaMallocHost((void**)&a, sizeof(int) * count);
    cudaMallocHost((void**)&b, sizeof(int) * count);

    for (int i = 0; i < count; i++)
    {
        a[i] = e() % 100000;