以下函数用于近似计算 1 / √x
计算结果有一定的误差,在对精度有要求的地方还是老老实实用 1.0 / sqrt(x) 吧
float Q_rsqrt(float number)
{
long i;
float x2, y;
const float threehalfs = 1.5F;
x2 = number * 0.5F;
y = number;
i = * (long *) &y;
i = 0x5f3759df - ( i >> 1);
y = * (float *) &i;
y = y * (threehalfs - (x2 * y * y)); // 1st iteration
// y = y * (threehalfs - (x2 * y * y)); // 2nd iteration, can be removed
return y;
}
该方法的运行速度已经比标准库慢了许多
仅仅看看而以
int main()
{
clock_t start_time=clock();
for (int i=0; i<1000000; ++i) {
Q_rsqrt(34);
}
clock_t end_time=clock();
std::cout << "Q_rsqrt Running time is: "<<static_cast<double>(end_time-start_time)/CLOCKS_PER_SEC*1000<<"ms"<< std::endl;
clock_t start_time2=clock();
for (int i=0; i<1000000; ++i) {
1.0 / sqrt(34);
}
clock_t end_time2=clock();
std::cout << "sqrt Running time is: "<<static_cast<double>(end_time2-start_time2)/CLOCKS_PER_SEC*1000<<"ms"<< std::endl;
}
Q_rsqrt Running time is: 19.618ms
1.0 / sqrt(34) Running time is: 3.872ms