OpenACC 计算圆周率（简单版）

▶ 书上的计算圆周率的简单程序，主要是使用了自定义函数

#include <stdio.h>
#include <stdlib.h>
#include <math.h>
#include <openacc.h>

#define N   100

#pragma acc routine seq
float ff(const float x)
{    
    return 4.0f / (1.0f + x * x);
}

int main()
{
    const float h = 1.0f / N;
    float sumf = 0, result;
           
#pragma acc parallel loop reduction(+:sumf)
    for (int i = 0; i < N; i++)
        sumf += ff(h * (i - 0.5f));

    result = h * sumf;    
    printf("
N = %d, myPi = %f, diff = %e
", N, result, result / 3.141592653589793238 - 1);
    //getchar();
    return 0;
}

● 输出结果

D:CodeOpenACCOpenACCProjectOpenACCProject>pgcc main.c -acc -Minfo -o main_acc.exe
ff:
     10, Generating acc routine seq
         Generating Tesla code
     11, FMA (fused multiply-add) instruction(s) generated
main:
     19, Accelerator kernel generated
         Generating Tesla code
         20, #pragma acc loop gang, vector(100) /* blockIdx.x threadIdx.x */
             Generating reduction(+:sumf)
     19, Generating implicit copy(sumf)

D:CodeOpenACCOpenACCProjectOpenACCProject>main_acc.exe
launch CUDA kernel  file=D:CodeOpenACCOpenACCProjectOpenACCProjectmain.c function=main line=19 device=0 threadid=1 num_gangs=1 num_workers=1 vector_length=100 grid=1 block=100 shared memory=1024
launch CUDA kernel  file=D:CodeOpenACCOpenACCProjectOpenACCProjectmain.c function=main line=19 device=0 threadid=1 num_gangs=1 num_workers=1 vector_length=256 grid=1 block=256 shared memory=1024

N = 100, myPi = 3.161500, diff = 6.336546e-03
PGI: "acc_shutdown" not detected, performance results might be incomplete.
 Please add the call "acc_shutdown(acc_device_nvidia)" to the end of your application to ensure that the performance results are complete.

Accelerator Kernel Timing data
D:CodeOpenACCOpenACCProjectOpenACCProjectmain.c
  main  NVIDIA  devicenum=0
    time(us): 11
    19: compute region reached 1 time
        19: kernel launched 1 time
            grid: [1]  block: [100]
            elapsed time(us): total=1000 max=1000 min=1000 avg=1000
        19: reduction kernel launched 1 time
            grid: [1]  block: [256]
             device time(us): total=0 max=0 min=0 avg=0
    19: data region reached 2 times
        19: data copyin transfers: 1
             device time(us): total=4 max=4 min=4 avg=4
        23: data copyout transfers: 1
             device time(us): total=7 max=7 min=7 avg=7