小编sil*_*lva的帖子

我刚刚发现 libcu++ 库并尝试使用这些cuda::atomic变量。我编写了以下程序，但它给了我意想不到的结果：

#include <atomic>
#include <cuda/atomic>
#include <stdio.h>


#define gpuErrchk(ans) { gpuAssert((ans), __FILE__, __LINE__); }
inline void gpuAssert(cudaError_t code, const char *file, int line, bool abort=true)
{
   if (code != cudaSuccess) 
   {
      fprintf(stderr,"GPUassert: %s %s %d\n", cudaGetErrorString(code), file, line);
      if (abort) exit(code);
   }
}


__global__ void atomic_test()
{
    cuda::atomic<int, cuda::thread_scope_block> x{0};
    x.fetch_add(1, cuda::memory_order_seq_cst);
    __syncthreads();
    int y = x.load(cuda::memory_order_acquire);
    printf("(%d %d) - Value of x is %d\n", blockIdx.x, threadIdx.x, y); 
}

int main()
{
    atomic_test<<<2, 32>>>();
    gpuErrchk( cudaDeviceSynchronize() …