CUDA并发规约求和

CUDA并发规约求和

https://www.cnblogs.com/viviman/archive/2012/11/21/2780286.html

https://blog.csdn.net/abcjennifer/article/details/43528407

https://blog.csdn.net/taonull/article/details/52606179

https://www.cnblogs.com/little-hobbit/p/4366493.html

规约求和

规约求和的核函数代码如下：
__global__ void RowSum(float* A, float* B){
    int bid = blockIdx.x;
    int tid = threadIdx.x;

    __shared__ s_data[128];
    
    //read data to shared memory
    s_data[tid] = A[bid*128 + tid];
    __synctheads();        //sync

    for(int i=64; i>0; i/=2){
        if(tid<i)
            s_data[tid] = s_data[tid] + s_data[tid+i] ;
        __synctheads();
    }
    if(tid==0)
        B[bid] = s_data[0];
}