Question

****New to C!****

I am running Dev-C++ 4.9.9.2 on Windows 7 (64 bit build)
My computer has 39GB of Physical Memory.

I am trying to create a large two dimensional array. I have already created code that tells me how many dimensions it has, and how many items are in each dimension.

As an example, let’s say the array is two dimensional: One million items long, and 6 wide:

[1,2,3,4,5,6],
[1,2,3,4,5,6],
[1,2,3,4,5,6],
...and on to one million items.

I have tried:

float MyArray[1000000][6];

but this crashes Dev-C. It seems to fail when I try to initialize an array larger than:

float Myarray[86486][6];

I imagine I am experiencing a “stack overflow” which amuses me since that is the name of this site.

I have been digging around and it seems I need to use malloc to help C understand how much memory to reserve. I have seen good examples of how to use this to set up a 1 dimensional array, but I would very much appreciate example code of how to set this up with a 2 dimensional array.

I have seen the example here: Initializing a large two dimensional array in C

But I’m afraid I am too much of a beginner in c to understand the brief explanation.
As background: I am coming from python where you can make an array of (almost) any size or dimension by just declaring MyArray=[] and then filling it with whatever you want.

Thank you!

Answer 1

Yeah, you’re running up against the limit of the size of an individual stack frame.

Here’s one approach:

#include <stdlib.h>
...
int main(void)
{
  /**
   * declare myArray as a *pointer* to a 6-element array of float
   */
  float (*myArray)[6];
  /**
   * dynamically allocate space for 1 million objects of type
   * "6-element array of float"
   */
  myArray = malloc(sizeof *myArray * 1000000);
  ...
  myArray[i][j] = ...;
}

Why this works:

The subscript operation a[i] is interpreted as *(a + i); that is, we compute the address of the i‘th element (not byte) after a (the base address of the array) and dereference it. Since myArray is a pointer to a 6-element array of float, myArray[i] gives us the address of the i‘th 6-element array of float after myArray.

The advantage of this approach is that the memory is allocated in a contiguous chunk, and you can subscript myArray like any 2-d array.