为数组赋值时为什么会崩溃？ [英] why it crashes when assigning new values to arrays?

问题描述

可能重复：
无法分配内存

我的下面的代码运行良好：

 双重重量[600] [800] [3]; 
 double mean [600] [800] [3]; 
 double sd [600] [800] [3]; 
 double u_diff [600] [800] [3]; （int j = 0; j <800; j ++）
 
（int i = 0; i <600; i ++）
 {
 （int k = 0; k <3; m ++）
 {
 weight [i] [j] [k] = 0; 
 mean [i] [j] [k] = 0; 
 sd [i] [j] [k] = 6; 
 
 
 
 $ / code $ / pre 
 
但是当我改变它这个表单： 
 
 
  int init = 6; 
 int C = 3; （int j = 0; j <800; j ++）
 
（int i = 0; i <600; i ++）
 {
 （int k = 0; k <3; k ++）
 {
 weight [i] [j] [k] = 1 / C; b $ b {
 
 mean [i] [j] [k] = rand（）; 
 sd [i] [j] [k] = init; 
 
 
 
 
 
 
崩溃。我甚至尝试过为重量，平均值和SD分别工作。我怀疑它可能是数据类型，如下所示： 
 
 
  double value = rand（）; 
 weight [i] [j] [m] = value; 
  
但是错误仍然存​​在。这里有什么问题？ 
解决方案
 
我还得到了第一个崩溃的版本（cygwin，4.5.3）。
 
 
这个问题与有限的堆栈大小有关，大小约为2 MB。
 
 
为什么不会崩溃可能是由于优化： 
 
由于'rand'在另一个片段中，优化器/编译器不可能在
中告诉数组根本没有被使用 - 这很可能是可见的< b 
 
 
  gcc -std = c99 tst.c -O&& ./a.exe  - 不会产生
 gcc -std = c99 tst.c&& ./a.exe  - 分段错误
  
为了解决这个错误， malloc 
的堆（或者研究一下极小的数组80x60x3也许会限制这个限制？）
 
 
  // tst。 c 
 //编译并运行gcc -std = c99 tst.c -DOK = 0 -DW = 80 -DH = 60&& ./a.exe // ok 
 //或gcc -std = c99 tst.c -DOK = 0 -DW = 800 -DH = 600&& ./a.exe //崩溃
 //或gcc -std = c99 tst.c -DOK = 1 -DW = 800 -DH = 600&& ./a.exe // ok 
 #include< stdlib.h> 
 int main（）
 {
 #if OK 
 double * weight =（double *）malloc（W * H * 3 * sizeof（double））; //没有崩溃
 #else 
双重重量[W * H * 3]; //当W * H很大时崩溃，W * H很小的时候崩溃
 #endif 
 int z = 0; （int j = 0; j< H; j ++）
 {
 for（int i = 0; i< W; i ++） （int m = 0; m <3; m ++）
 {
 weight [z ++] = 0; 
} 
} 
} 
 return 0; 
} 
  
 

  
Possible Duplicate:

  Could not allocate memory  
My following code runs fine:
double weight [600] [800][3];
double mean [600] [800][3];
double sd [600] [800][3];
double u_diff [600] [800][3];

for ( int i = 0; i < 600; i ++ )
{
    for ( int j = 0; j < 800; j ++ )
    {
        for ( int k=0; k < 3; m ++ )
        {
            weight [i][j][k] = 0;
            mean[i][j][k] = 0; 
            sd[i][j][k] = 6;        
        }       
    }
}
But when I change it into this form:
int init = 6;
int C = 3;

for ( int i = 0; i < 600; i ++ )
{
    for ( int j = 0; j < 800; j ++ )
    {
        for ( int k =0; k < 3; k ++ )
        {
            weight [i][j][k] = 1/C;
            mean[i][j][k] = rand(); 
            sd[i][j][k] = init;         
        }       
    }
}
it crashes. I even tried working for "weight", "mean" and "sd" seperately. I doubt it might be of datatype, changed like: 
double value = rand();
weight[i][j][m] = value;
but the error still remains. What is wrong here? 
解决方案
I got also the first version to crash (cygwin, 4.5.3).

The problem has to do with limited stack size, which has been around 2 MB.

Why it wouldn't crash is probably due to optimization:

due to 'rand' in the other fragment, the optimizer/compiler couldn't possibly

tell that the array is not used at all -- which would very likely be visible

from the first fragment.
gcc -std=c99 tst.c -O  && ./a.exe -- produces nothing
gcc -std=c99 tst.c && ./a.exe -- segmentation fault
To get around the error, just allocate the large arrays from the heap with malloc
(or study the limit by having considerably smaller array 80x60x3 perhaps?)  
// tst.c
// compile and run with gcc -std=c99 tst.c -DOK=0 -DW=80 -DH=60 && ./a.exe    // ok
//               or     gcc -std=c99 tst.c -DOK=0 -DW=800 -DH=600 && ./a.exe  // crash
//               or     gcc -std=c99 tst.c -DOK=1 -DW=800 -DH=600 && ./a.exe  // ok
#include <stdlib.h>
int main()
{
#if OK
    double *weight =(double*)malloc(W*H*3*sizeof(double));      // no crash
#else
    double weight[W*H*3];   // crash when W*H is large, nocrash when W*H is small
#endif
    int z=0;
    for ( int i = 0; i < W; i ++ )
    {
        for ( int j = 0; j < H; j ++ )
        {
            for ( int m =0; m < 3; m ++ )
            {
                 weight[z++]=0;     
            }       
        }
    }
    return 0;
}


                        
这篇关于为数组赋值时为什么会崩溃？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！