C++ 致命错误 C1001:使用 openMP 的编译器发生内部错误 [英] C++ fatal error C1001: An internal error has occurred in the compiler with openMP

问题描述

我有一个解决数独谜题的程序，我已经让它按顺序工作，但现在我正在尝试使用 openMP 并行化它.函数 solvePuzzle() 包含算法，我想在其中并行化 for 循环，但是当我在 for 循环之前添加 #pragma omp parallel for 语句时，我收到此错误: fatal error C1001: 编译器出现内部错误.

函数的代码是solvePuzzle():

 bool sudoku::solvePuzzle(int grid[CELL][CELL]) {整数行，列；if (!findEmptyCell(grid, row, col))返回真；#pragma omp 并行用于for (int num = 1; num <= 9; num++) {if (checkAccuracy(grid, row, col, num)) {网格[行][列] = num;if (solvePuzzle(grid))返回真；网格[行][列] = EMPTY_CELL;}}返回假；}

如果有帮助，这是带有 main 的驱动程序:

#include "SudokuGrid.h"使用命名空间标准；int main() {数独数独；时钟_t t1, t2;t1 = 时钟()；if (sudoku.readPuzzle("1.txt")) {sudoku.printGrid(sudoku.grid);}别的 {cout<<不正确的文件"<<结束;}cout<<结束;if (sudoku.solvePuzzle(sudoku.grid) == true)sudoku.printGrid(sudoku.grid);别的printf("无解");t2 = 时钟()；printf("执行时间 = %1d ms\n", (t2 - t1));返回0；}

构建时完全错误:

1>------ 构建开始:项目:数独，配置:调试 Win32 ------1>数独网格.cpp1>c:\users\john\source\repos\sudoku\sudoku\sudokugrid.cpp(8): 致命错误C1001: 编译器发生内部错误.1>(编译器文件'f:\dd\vctools\compiler\utc\src\p2\main.c'，第258行)1>要解决此问题，请尝试在附近简化或更改程序上面列出的位置.1>请在Visual C++上选择技术支持命令1>帮助菜单，或打开技术支持帮助文件了解更多信息1>1> 已完成构建项目Sudoku.vcxproj" -- 失败.========== 构建:0 成功，1 失败，0 最新，0 跳过 ==========

解决方案

总而言之，问题是从并行化的 for 循环内部返回.

#pragma omp parallel forfor (int num = 1; num <= 9; num++) {...if (solvePuzzle(grid))返回真；//坏的...}}

<块引用>

OpenMP 要求循环结构处理每次迭代.不允许跳出循环(使用 return、goto、break、throw 或其他方式)."

-- https://bisqwit.iki.fi/story/howto/openmp/

所以解决方案是让循环遍历所有迭代(另见跳出循环).

bool solvePuzzle(int grid[CELL][CELL]) {布尔解决 = 假；整数行，列；if (!findEmptyCell(grid, row, col))返回真；#pragma omp 并行用于for (int num = 1; num <= 9; num++) {#pragma omp flush (已解决)if (!solved && checkAccuracy(grid, row, col, num)) {网格[行][列] = num;if (solvePuzzle(grid)) {解决了=真；#pragma omp flush (已解决)} 别的 {网格[行][列] = EMPTY_CELL;}}}返回解决；}

复制

演示

I have a program that solves sudoku puzzles and I've gotten it to work sequentially, but now I'm trying to parallelize it using openMP. The function solvePuzzle() includes the algorithm and I want to parallelize the for loop within it however when I add #pragma omp parallel for statement before my for loop i get this error: fatal error C1001: An internal error has occurred in the compiler.

The code for the function is solvePuzzle():

    bool sudoku::solvePuzzle(int grid[CELL][CELL]) {
        int row, col;

        if (!findEmptyCell(grid, row, col))
            return true; 
    #pragma omp parallel for
        for (int num = 1; num <= 9; num++) {
            if (checkAccuracy(grid, row, col, num)) {
                grid[row][col] = num;

                if (solvePuzzle(grid))
                    return true;

                grid[row][col] = EMPTY_CELL;
            }
        }
        return false;
    }

Here is the driver with main if it helps:

#include "SudokuGrid.h"

using namespace std;

int main() {

    sudoku sudoku;
    clock_t t1, t2;
    t1 = clock();

    if (sudoku.readPuzzle("1.txt")) {
        sudoku.printGrid(sudoku.grid);
    }
    else {
        cout << "Incorrect file" << endl;
    }
    cout << endl;
    if (sudoku.solvePuzzle(sudoku.grid) == true)
        sudoku.printGrid(sudoku.grid);
    else
        printf("No solution exists");

    t2 = clock();
    printf("Time to execute = %1d ms\n", (t2 - t1));

    return 0;
}

Full error when built:

1>------ Build started: Project: Sudoku, Configuration: Debug Win32 ------
1>SudokuGrid.cpp
1>c:\users\john\source\repos\sudoku\sudoku\sudokugrid.cpp(8): fatal error 
C1001: An internal error has occurred in the compiler.
1>(compiler file 'f:\dd\vctools\compiler\utc\src\p2\main.c', line 258)
1> To work around this problem, try simplifying or changing the program near 
the locations listed above.
1>Please choose the Technical Support command on the Visual C++
1> Help menu, or open the Technical Support help file for more information
1>
1>Done building project "Sudoku.vcxproj" -- FAILED.
========== Build: 0 succeeded, 1 failed, 0 up-to-date, 0 skipped ==========

解决方案

To summarize, the problem is returning from inside the parallelized for loop.

#pragma omp parallel for
    for (int num = 1; num <= 9; num++) {
            ...
            if (solvePuzzle(grid))
                return true;           // BAD

            ...
        }
    }

"OpenMP requires that a loop construct processes each iteration. Breaking out of the loop (using return, goto, break, throw or other means) is not allowed."

-- https://bisqwit.iki.fi/story/howto/openmp/

So the solution is to let the loop loop through all iterations (See also Breaking Out of Loops).

bool solvePuzzle(int grid[CELL][CELL]) {
    bool solved = false;
    int row, col;

    if (!findEmptyCell(grid, row, col))
        return true;

    #pragma omp parallel for
    for (int num = 1; num <= 9; num++) {
        #pragma omp flush (solved)
        if (!solved && checkAccuracy(grid, row, col, num)) {
            grid[row][col] = num;

            if (solvePuzzle(grid)) {
                solved = true;
                #pragma omp flush (solved)
            } else {
                grid[row][col] = EMPTY_CELL;
            }
        }
    }
    return solved;
}

Repro

Demo

这篇关于C++ 致命错误 C1001:使用 openMP 的编译器发生内部错误的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！