C++与OMP配合的最简线程池

最近在写算法时，发现OpenMP + C++ functional 可以瞬间构造一个线程池，并执行风马牛不相及的事情。看来，有必要好好思考一下今后算法优化的事情了！
在 g++下通过编译

#include <iostream>
#include <functional>
#include <vector>
using namespace std;

void fun (int a, int b)
{
	cout<< "fun exec :"<< a << '+' << b << '=' << a + b <<endl;
}

class C{
private:
	float m_c = 2.0f;
public:
	void mp( float d)
	{
		cout<<"c::mp exec :"<< m_c << 'x' << d << '=' << m_c * d <<endl;
	}
};

int main(int argc, char * argv[])
{
	const int task_groups = 5;
	C c [task_groups]; 
	vector<function<void (void) > > tasks;
	for (int i=0;i<task_groups;++i)
	{
		tasks.push_back(bind( fun , 10, i * 10 ) );
		tasks.push_back(bind( &C::mp , &c[i], i*2.0f ) );
		tasks.push_back(bind(
			[=] (void) {cout << "lambada :" <<i << endl;	}
			) );
	}
	size_t sz = tasks.size();
#pragma  omp parallel for
	for (size_t i=0;i<sz;++i)
	{
		tasks[i]();
	}
	return 0;
}

输出：

fun exec :10+0=10
c::mp exec :2x0=0
lambada :0
fun exec :10+10=20
c::mp exec :2x2=4
lambada :1
fun exec :10+20=30
c::mp exec :2x4=8
lambada :2
fun exec :10+30=40
c::mp exec :2x6=12
lambada :3
fun exec :10+40=50
c::mp exec :2x8=16
lambada :4

编译器选项：
-fopenmp
链接选项：
-lgomp