std::async方式循环提速

#include <future>
double CControl::GetSum(int data)
{
    double sum = 0;
    for (int j = 0; j < data; j++)
    {
        sum += j + 0.1 * pow(j, 4) + 88.8 * pow(j, 3);
    }
    return sum;
}

void Test()
{
　　int cycCount = 8;
    int addCount = 120000;
    CString strmsg;
    double t1, t2;
    DWORD dwStart = GetTickCount();
    double sum = 0;
    for (int i = 0; i < cycCount; i++)
    {
        for (int j = 0; j < addCount; j++)
        {
            sum += j + 0.1 * pow(j,4) + 88.8 * pow(j, 3);            
        }
    }
    t1 = GetTickCount() - dwStart;
    strmsg.Format("%0.3f %0.3f ", t1, sum);

    Sys_State state = m_SysState;
    dwStart = GetTickCount();
    double result = 0;
    std::vector<std::future<double>> fut(cycCount);
    //速度会比单流程方式快2-3倍
　　　　

const int32_t threadNum = static_cast<int32_t>(std::thread::hardware_concurrency());
strmsg.Format("%d", threadNum); //threadNum 等于CPU核心数，可以根据这个开这么多个异步线程

    for (int i = 0; i < cycCount; i++)
    {
        //函数方式
        //fut[i] = std::async(std::launch::async,GetSum, addCount);
        //lambda 方式，传参数只能用局部变量,不能用类成员变量
        fut[i] = std::async(std::launch::async, [addCount, state]{
            double sum = 0;
            for (int j = 0; j < addCount; j++)
            {
                sum += j + 0.1 * pow(j, 4) + 88.8 * pow(j, 3);
            }
            return sum; });
    }
    //auto v = fut[0].get();  
    for (int i = 0; i < cycCount; i++)
    {
        fut[i].wait();
        result += fut[i].get();
    }
    t2 = GetTickCount() - dwStart;
    strmsg.Format("%0.3f %0.3f result:%0.3f %0.3f", t1, t2, sum, result);
    AfxMessageBox(strmsg);
}