关于OpenMP小程序测试结果的疑问

[dimire]

今天装上了intel C++ compiler 9.0编译器，并且编了个小的测试程序
#include <stdio.h>
#include "omp.h"

int main(int argc, char* argv[])
{
int i;
#pragma omp parallel for
for (i = 0; i < 10; i++ )
{
printf("i = %d\n", i);
}
return 0;
}

这个程序理想的打印结果应该是乱序的0到9数字，可我每次运行的时候，能输出正常的结果，但有时候又会出现重复输出两个数字的结果，比如
i = 0
i = 1
i = 2
i = 3
i = 3
i = 5
i = 4
i = 6
i = 7
i = 8
i = 9
Press any key to continue

i = 0
i = 0
i = 5
i = 6
i = 7
i = 1
i = 8
i = 2
i = 3
i = 9
i = 4
Press any key to continue
这是什么原因呢？

[intel_jeff]

你选择的链接库类型不对，应该选择多线程版本，即/ML或/MLd

看我在4核机器上的结果：

============================
链接多线程版本的库，屏幕输出正确
============================
C:\>icl /Qopenmp /MD hello.c
Intel(R) C++ Compiler for applications running on IA-32, Version 10.1 Build 20080212 Package ID: w_cc_p_10.1.020
Copyright (C) 1985-2008 Intel Corporation. All rights reserved.

hello.c
hello.c(7): (col. 1) remark: OpenMP DEFINED LOOP WAS PARALLELIZED.
Microsoft (R) Incremental Linker Version 7.10.3077
Copyright (C) Microsoft Corporation. All rights reserved.

-out:hello.exe
-nodefaultlib:libguide_stats.lib
-nodefaultlib:libguide40_stats.lib
-defaultlib:libguide40.lib
hello.obj

C:\>hello
i = 0
i = 6
i = 7
i = 1
i = 2
i = 3
i = 4
i = 5
i = 8
i = 9

============================
链接单线程版本的库，屏幕输入有重复，这是由于单线程库中的printf是不可重入的，但是某些时刻却被多个线程同时调用，因此出现重复打印
============================
C:\>icl /Qopenmp /ML hello.c
Intel(R) C++ Compiler for applications running on IA-32, Version 10.1 Build 20080212 Package ID: w_cc_p_10.1.020
Copyright (C) 1985-2008 Intel Corporation. All rights reserved.

hello.c
hello.c(7): (col. 1) remark: OpenMP DEFINED LOOP WAS PARALLELIZED.
Microsoft (R) Incremental Linker Version 7.10.3077
Copyright (C) Microsoft Corporation. All rights reserved.

-out:hello.exe
-nodefaultlib:libguide_stats.lib
-nodefaultlib:libguide40_stats.lib
-defaultlib:libguide.lib
hello.obj

C:\>hello
i = 0
i = 3
i = 6
i = 6
i = 8
i = 1
i = 4
i = 7
i = 9
i = 2
i = 5

[intel_jeff]

楼上有笔误，应该是
你使用了 /ML 或 /MLd （即单线程版本库）
应该使用多线程版本库，如 /MD, /MDd, /MT, /MTd

[dimire]

问题解决了，谢谢大家指点：）

[intel_www]

检查使用的C runtime library是不是多线程版本。

[dimire]

对于上面的那个程序，我把变量i的定义放在了for语句中用icc编译就有错,而用VC编译器就能没问题
int main(int argc, char* argv[])
{
//int i;
#pragma omp parallel for
for (int i = 0; i < 10; i++ )
{
printf("i = %d\n", i);
}
return 0;
}

显示错误信息如下：
icl openmp.c
icl: warning: problem with Microsoft compilation of 'D:\openmp\openmp.c'
D:\openmp\openmp.c(126): error: expected an expression
for (int i = 0; i < 10; i++ )
^
D:\openmp\openmp.c(126): error: identifier "i" is undefined
for (int i = 0; i < 10; i++ )
^
D:\openmp\openmp.c(126): error: OpenMP for-init does not conform
for (int i = 0; i < 10; i++ )
^
D:\openmp\openmp.c(126): error: OpenMP for-test does not conform
for (int i = 0; i < 10; i++ )
^
D:\openmp\openmp.c(126): error: OpenMP for-incr does not conform
for (int i = 0; i < 10; i++ )

是不是intel编译器不支持这样的变量定义？我看别人写的Openmp测试程序，有在for语句中定义并初始化变量的。请问大家有没遇到这种情况？

[intel_www]

[Quote=引用 5 楼 dimire 的回复:]
对于上面的那个程序，我把变量i的定义放在了for语句中用icc编译就有错,而用VC编译器就能没问题
int main(int argc, char* argv[])
{
//int i;
#pragma omp parallel for
for (int i = 0; i < 10; i++ )
{
printf("i = %d\n", i);
}
return 0;
}

显示错误信息如下：
icl openmp.c
icl: warning: problem with Microsoft compilation of 'D:\openmp\openmp.c'
D:\openmp\openmp.c(126): error: expe…
[/Quote]

这涉及C和C++语言标准的问题。你把文件后缀名改成CPP就行了。