OpenMP学习的例子

#include <omp.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>

//

void omp_thread_test()
{
	int coreNum = omp_get_num_procs();
	printf("coreNum = %d\n", coreNum);
	omp_set_num_threads(coreNum);	//设置线程数，一般设置的线程数不超过CPU核心数，这里开4个线程执行并行代码段  
#pragma omp parallel for
	for (int i = 0; i < 4 * coreNum; i++)
		printf("i = %d, I am Thread %d\n", i, omp_get_thread_num());
}

//

double Test(int n)
{
	double t = 0.0;
	for (int i = 0; i < 100000000; ++i) {
		t += n*n;
	}
	//printf("n = %d ,  %f I am Thread %d\n", n, t, omp_get_thread_num());
	return t;
}

void omp_for_test1()
{
	double t_avg = 0.0;
	clock_t t2 = clock();
	omp_set_num_threads(8);
#pragma omp parallel for reduction(+: t_avg)
	for (int i = 0; i < 10; ++i) {
		t_avg += Test(i);
	}
	t_avg = t_avg / 10;

	clock_t t3 = clock();
	printf("t_avg = %f ,Parallel time: %d\n", t_avg, t3 - t2);
}

void omp_for_test2()
{
	double t_avg = 0.0;
	clock_t t2 = clock();
#pragma omp parallel for 
	for (int i = 0; i < 10; ++i)
	{
#pragma omp critical  
		t_avg += Test(i);
	}
	t_avg = t_avg / 10;

	clock_t t3 = clock();
	printf("t_avg = %f ,Parallel time: %d\n", t_avg, t3 - t2);
}


//
void no_opm_pi()
{
	double factor = 1;
	double sum = 0.0;
	int n = 1000000000;
	clock_t t2 = clock();
	for (int i = 0; i < n; i++)
	{
		sum += factor / (2 * i + 1);
		factor = -factor;

	}
	double pi_approx = 4.0*sum;
	clock_t t3 = clock();
	printf("%f   %d\n", pi_approx, t3 - t2);
}

void opm_for_pi()
{
	double factor = 1;
	double sum = 0.0;
	int n = 1000000000;
	clock_t t2 = clock();
#pragma omp parallel for num_threads(omp_get_num_procs()) reduction(+:sum) private(factor)
	for (int i = 0; i < n; i++)
	{
		if (i % 2 == 0)
			factor = 1.0;
		else
			factor = -1.0;
		sum += factor / (2 * i + 1);
	}
	double pi_approx = 4.0*sum;
	clock_t t3 = clock();
	printf("%f   %d\n", pi_approx, t3 - t2);
}
//
//
//
static long num_steps = 1000000000;//越大值越精确
const int NUM_THREADS = 8;
void no_omp_pi2()
{
	int i;
	double x = 0.0, pi = 0.0, sum = 0.0;
	double step = 1.0 / (double)num_steps;
	clock_t t2 = clock();
	for (i = 1; i <= num_steps; i++)
	{
		x = (i - 0.5)*step;
		sum = sum + 4.0 / (1.0 + x*x);
	}
	pi = step*sum;
	clock_t t3 = clock();
	printf("no_omp_pi2 : %1.12f %d\n", pi, t3 - t2);
}


void omp_for_pi2_1()
{
	int i;
	double pi = 0.0;
	double sum[NUM_THREADS];
	double step = 1.0 / (double)num_steps;
	omp_set_num_threads(NUM_THREADS);
	clock_t t2 = clock();
#pragma omp parallel private(i)  //并行域开始，每个线程(0和1)都会执行该代码
	{
		double x;
		int id = omp_get_thread_num();
		for (i = id, sum[id] = 0.0; i < num_steps; i = i + NUM_THREADS)
		{
			x = (i + 0.5)*step;
			sum[id] += 4.0 / (1.0 + x*x);
		}
	}
	for (i = 0; i < NUM_THREADS; i++)
		pi += sum[i] * step;
	clock_t t3 = clock();
	printf("omp_for_pi2_1: %1.12f %d\n", pi, t3 - t2);
}

void omp_for_pi2_2()
{
	int i;
	int id;
	double pi = 0.0, sum[NUM_THREADS] = { 0.0 };
	double step = 1.0 / (double)num_steps;
	omp_set_num_threads(NUM_THREADS);  //设置2线程
	clock_t t2 = clock();
#pragma omp parallel  //并行域开始，每个线程都会执行该代码
	{
		double x;

		id = omp_get_thread_num();
		sum[id] = 0;
#pragma omp for  //未指定chunk，迭代平均分配给各线程，连续划分
		for (i = 1; i <= num_steps; i++) {
			x = (i + 0.5)*step;
			sum[id] += 4.0 / (1.0 + x*x);
		}
	}
	for (i = 0; i < NUM_THREADS; i++)
		pi += sum[i] * step;
	clock_t t3 = clock();
	printf("omp_for_pi2_2: %1.12f %d\n", pi, t3 - t2);
}

void omp_for_pi2_3()
{
	int i;
	double pi = 0.0;
	double sum = 0.0;
	int id = 0;
	double x = 0.0;
	double step = 1.0 / (double)num_steps;
	clock_t t2 = clock();
	omp_set_num_threads(NUM_THREADS);
#pragma omp parallel private(x,sum,id) //该子句表示x,sum变量对于每个线程是私有的(!!!注意id要设为私有变量)
	{
		id = omp_get_thread_num();
		for (i = id, sum = 0.0; i <= num_steps; i = i + NUM_THREADS)
		{
			x = (i + 0.5)*step;
			sum += 4.0 / (1.0 + x*x);
		}
#pragma omp critical  //指定代码段在同一时刻只能由一个线程进行执行
		pi += sum*step;
	}
	clock_t t3 = clock();
	printf("omp_for_pi2_3: %1.12f %d\n", pi, t3 - t2);
}

void omp_for_pi2_4()
{
	int i;
	double pi = 0.0;
	double sum = 0.0;
	double x = 0.0;
	double step = 1.0 / (double)num_steps;
	omp_set_num_threads(NUM_THREADS);  //设置2线程
	clock_t t2 = clock();
#pragma omp parallel for reduction(+:sum) private(x) //每个线程保留一份私有拷贝sum，x为线程私有，最后对线程中所以sum进行+规约，并更新sum的全局值
	for (i = 1; i <= num_steps; i++) {
		x = (i - 0.5)*step;
		sum += 4.0 / (1.0 + x*x);
	}
	pi = sum * step;
	clock_t t3 = clock();
	printf("omp_for_pi2_4: %1.12f %d\n", pi, t3 - t2);
}

//
//
//
void omp_sleep()
{
	int coreNum = omp_get_num_procs();
	printf("coreNum = %d\n", coreNum);
	omp_set_num_threads(coreNum);	//设置线程数，一般设置的线程数不超过CPU核心数，这里开4个线程执行并行代码段  
#pragma omp parallel for
	for (int i = 0; i < coreNum; i++)
	{
		while (1)
		{
			int a = 5;
			int b = a;
		}
	}
}
//
//
//
int main(int argc, char* argv[])
{
	//omp_thread_test();
	//omp_for_test1();
	//omp_for_test2();
	//no_opm_pi();
	//opm_for_pi();
	//no_omp_pi2();
	//omp_for_pi2_1();
	//omp_for_pi2_2();
	//omp_for_pi2_3();
	//omp_for_pi2_4();
	omp_sleep();
	system("pause");
}