np.arange与C ++ iota的比较，iota较慢

Question

I wrote a small chunk of code to generate a sequence of integers, and compared it with np.arange() , and I found np.arange() is faster. Anyone can give me some hint if there is any faster C++ implementation of this ?

Here is the code:

#include <iostream>
#include <chrono>
#include <ctime>
#include <array>
#include <vector>
#include <numeric>
template <typename T>
std::vector<T> range(T start, T end) {
    size_t N = (int)floor(end - start) + 1;
    std::vector<T> vec(N);
    std::iota(vec.begin(), vec.end(), start);
    return vec;
}

int main() {
    auto start = std::chrono::system_clock::now();
    std::vector<int> x_range = range(0, 1024);
    auto end = std::chrono::system_clock::now();
    std::chrono::duration<double> elapsed_seconds = end - start;
    std::time_t end_time = std::chrono::system_clock::to_time_t(end);

    std::cout << "elapsed time: " << elapsed_seconds.count() << "s\n";
}

Python code is simple:

%timeit x = np.arange(0, 1024)

Python time: 1.51e-6 s C++ time: 1.6e-5 s

Python is 10 times faster than C++ implementation here.

Platform: Win10, Visual Studio Community 2017, in both O2 and Ox optimization mode. Both get more or less the same order of time- 1.x e-5 s .

Answer 1

Thanks everybody, although it is a simple piece of question, but I still learns a lot.

Here is new code which is different from code in the question, maybe the compassion will help some newbies understand better:

#include <iostream>
#include <chrono>
#include <ctime>
#include <array>
#include <vector>
#include <numeric>
#include <iostream>

template <typename T>
std::vector<T> range(T start, T end) {
    size_t N = (int)floor(end - start);
    std::vector<T> vec;
    vec.reserve(N);
    std::iota(vec.begin(), vec.end(), start);
    return vec;
}

int main() {
    auto start = std::chrono::system_clock::now();
    for (int i = 0; i < 100000000; ++i) {
        std::vector<int> x_range = range(0, 1000);
    }
    auto end = std::chrono::system_clock::now();
    std::chrono::duration<double> elapsed_seconds = end - start;
    std::time_t end_time = std::chrono::system_clock::to_time_t(end);

    std::cout << "elapsed time: " << elapsed_seconds.count() << "s\n";
    system("pause");

}

Now C++ implementation time: 7.14136e-8 s.