2d Array Troubles

[xanmas]

Hi,

I am trying to do a time evolution of a pressure wave in 2 dimensions for a computational physics class. Well, I have to define a 2d array for all the grid points I believe. Well, the best way I can think of to do this is by definining

double * grid;

outside any method with all my other variables and then

grid = new double[sizeX][sizeY];

in my init function where I initilize all my arrays and glut and whatnot.

Unfortunately, xcode is throwing me the following error for that declaration:
"error: cannot convert 'double (*)[128]' to 'double*' in assignment.

Later in the code with the code

			grid[x][y] = exp(-1 * ( pow((x-xcenter),2)/(2*sigmaXY*sigmaXY) + pow((y-ycenter),2)/(2*sigmaXY*sigmaXY)));
			P = grid[x][y];

I get the errors "error: invalid types 'double[int[' for arrays subscript' for both. I found hat if I use the declaration double ** grid; it gets rid of those two errors but not the first one.

I'm sure its just an error that I haven't picked up but any help would be very much appreciated.

Further info: I am on an intel mac using xcode and the intel c++ compiler version 11.0.

[Zaita]

// To Allocate
double **grid;
grid = new double*[sizeX];
for (int i = 0; i < sizeX; ++i)
 grid[i] = new double[sizeY];

// To Use
grid[x][y] = 0.0;

// To Clean Up
for (int i = 0; i < sizeX; ++i)
 delete [] grid[i];
delete [] grid;

Edit: You may find using a vector of vectors a better approach, less memory management you have to do yourself.

[xanmas]

Thanks for your help.

What would be the advantage to using a vector of vectors? What memory management is taken care of with that approach? Can my code still simply use grid[x][y]?

[Zhuge]

It will free you from having to allocate, resize, and free the dynamic memory. And yes, vectors do allow you to use the [] operators, but I recommend the at() member function since it will throw an exception if you access memory out of bounds.

[roodtree]

in c++ call:
double **grid;
grid=new double*grid[x];
for(int i=0;i<y;i++)grid=new double[y];
that is the vector of the vectors. We can call 3 4 d array by that way.
in c can call:
double grid[x][y];
grid=malloc(sizeof(x*y));
but in the momery it take X*Y+1 place on it. 1 for the rood which take the ip for the headot the 2d ray.
that right????????????????????

[Zaita]

@roodtree: incorrect.

Vector of Vectors.

#include <iostream>
#include <vector>

using namespace std;

int main() {

  int sizeX = 10;
  int sizeY = 10;
  vector<vector<double> > grid;

  // Resize our Grid
  grid.resize(sizeX);
  for (int i = 0; i < sizeX; ++i)
    grid[i].resize(sizeY);

  grid[0][0] = 3.0;
  grid[5][5] = 4.5;

  cout << "0/0 = " << grid[0][0] << endl;
  cout << "5/5 = " << grid[5][5] << endl;

  return 0;
}