CUDA随机数生成

Question

我想在我的__device__函数中生成随机数,并将它保存在我的int Board [500]中,我找到了一些例子,但他们使用了一些名为curandState的类型.我只需要一个类似于rand()C++ 的函数.

Answer 1

这是我的代码,我的设备内存中有N3 [40000]数组,我在内核中为一个线程生成一些随机数(我的意思是"内核<<< 1,1 >>> ......"),然后我将它从CPU复制到我的N2 [40000]并打印出来,所以这里是代码

#include <iostream>
#include <Cuda.h>
#include<curand.h>
#include<curand_kernel.h>


int n = 200;
using namespace std;

__device__ float generate( curandState* globalState, int ind ) 
{
    //int ind = threadIdx.x;
    curandState localState = globalState[ind];
    float RANDOM = curand_uniform( &localState );
    globalState[ind] = localState;
    return RANDOM;
}

__global__ void setup_kernel ( curandState * state, unsigned long seed )
{
    int id = threadIdx.x;
    curand_init ( seed, id, 0, &state[id] );
}

__global__ void kernel(float* N, curandState* globalState, int n)
{
    // generate random numbers
    for(int i=0;i<40000;i++)
    {
        int k = generate(globalState, i) * 100000;
        while(k > n*n-1)
        {
            k-=(n*n-1);
        }
        N[i] = k;
    }
}

int main() 
{
    int N=40000;

    curandState* devStates;
    cudaMalloc ( &devStates, N*sizeof( curandState ) );

    // setup seeds
    setup_kernel <<< 1, N >>> ( devStates,unsigned(time(NULL)) );

    float N2[40000];
    float* N3;
    cudaMalloc((void**) &N3, sizeof(float)*N);

    kernel<<<1,1>>> (N3, devStates, n);

    cudaMemcpy(N2, N3, sizeof(float)*N, cudaMemcpyDeviceToHost);

    for(int i=0;i<N;i++)
    {
        cout<<N2[i]<<endl;
    }

    return 0;
}