C++，重载运算符*=会降低性能

Question

我创建了一个包含三个整数的结构并重载了运算符*=， 然后创建一个大小为 100000000 的向量，并随机填充 0 和 1。

然后测试了一个函数的两个版本，将元素乘以 2 求和。 当我使用使用重载 operator*= 的函数时，我得到了将近 60% 的开销。

猜猜发生了什么事？

测试于：

• Windows 10
• Visual Studio 2019 (MSVC 14.29.30133)
• 项目配置：发布 x64（O2，启用全程序优化）。

谢谢

检查 main、without_overloading 和 with_overloading 函数。

#include <random>
#include <vector>
#include <iostream>
#include <ctime>

/*
I took the code for calculating times from this post:

https://stackoverflow.com/questions/17432502/how-can-i-measure-cpu-time-and-wall-clock-time-on-both-linux-windows/17440673#17440673
*/
//  Windows
#ifdef _WIN32
#include <Windows.h>
double get_wall_time() {
    LARGE_INTEGER time, freq;
    if (!QueryPerformanceFrequency(&freq)) {
        //  Handle error
        return 0;
    }
    if (!QueryPerformanceCounter(&time)) {
        //  Handle error
        return 0;
    }
    return (double)time.QuadPart / freq.QuadPart;
}
double get_cpu_time() {
    FILETIME a, b, c, d;
    if (GetProcessTimes(GetCurrentProcess(), &a, &b, &c, &d) != 0) {
        //  Returns total user time.
        //  Can be tweaked to include kernel times as well.
        return
            (double)(d.dwLowDateTime |
                ((unsigned long long)d.dwHighDateTime << 32)) * 0.0000001;
    }
    else {
        //  Handle error
        return 0;
    }
}
//  Posix/Linux
#else
#include <time.h>
#include <sys/time.h>
double get_wall_time() {
    struct timeval time;
    if (gettimeofday(&time, NULL)) {
        //  Handle error
        return 0;
    }
    return (double)time.tv_sec + (double)time.tv_usec * .000001;
}
double get_cpu_time() {
    return (double)clock() / CLOCKS_PER_SEC;
}
#endif

//=================================================================================
const size_t size = 100000000;

struct vec {
    int x, y, z;
    vec& operator*=(int rhs) {
        x *= rhs;
        y *= rhs;
        z *= rhs;
        return *this;
    }
};
    
void without_overloading(std::vector<vec> const &points) {
    double wall0 = get_wall_time();
    double cpu0 = get_cpu_time();

    vec sum{ 0,0,0 };
    for (size_t i = 0; i < size; ++i) {
        sum.x += points[i].x * 2;
        sum.y += points[i].y * 2;
        sum.z += points[i].z * 2;
    }

    double wall1 = get_wall_time();
    double cpu1 = get_cpu_time();

    std::cout << "Wall Time = " << wall1 - wall0 << std::endl;
    std::cout << "CPU Time  = " << cpu1 - cpu0 << std::endl;
    std::cout << sum.x << " " << sum.y << " " << sum.z << std::endl;
}

void with_overloading(std::vector<vec> &points) {
    double wall0 = get_wall_time();
    double cpu0 = get_cpu_time();

    vec sum{ 0,0,0 };
    for (size_t i = 0; i < size; ++i) {
        points[i] *= 2;
        sum.x += points[i].x;
        sum.y += points[i].y;
        sum.z += points[i].z;
    }

    double wall1 = get_wall_time();
    double cpu1 = get_cpu_time();

    std::cout << "Overloading Wall Time = " << wall1 - wall0 << std::endl;
    std::cout << "Overloading CPU Time  = " << cpu1 - cpu0 << std::endl;
    std::cout << sum.x << " " << sum.y << " " << sum.z << std::endl;
}

int main() {
    std::srand(std::time(nullptr));
    std::vector<vec> points(size);
    for (int i = 0; i < size; ++i) {
        points[i] = { rand() % 2, rand() % 2, rand() % 2 };
    }

    without_overloading(points);
    with_overloading(points);
}