C++ matrix boundary condition problems

Click For Summary
SUMMARY

The forum discussion addresses boundary condition issues in a C++ matrix implementation using a class called Lattice, which contains objects of type Dipole. The primary problem arises when attempting to access neighbors of dipoles located at the edges of the matrix, particularly when the index is negative. The user suggests that using the expression ((x + xSize) % xSize) for negative indices may resolve the issue. Additionally, debugging techniques are recommended to identify and resolve similar problems in the future.

PREREQUISITES
  • Understanding of C++ classes and object-oriented programming
  • Familiarity with matrix data structures and indexing
  • Knowledge of boundary condition handling in computational simulations
  • Experience with debugging techniques in C++
NEXT STEPS
  • Implement boundary condition checks in C++ matrix classes
  • Learn about modular arithmetic in programming, specifically for array indexing
  • Explore debugging tools in C++ such as GDB or integrated IDE debuggers
  • Study energy calculations in physical simulations involving dipoles
USEFUL FOR

C++ developers, computational physicists, and anyone working on simulations involving matrix data structures and boundary conditions.

rolotomassi
Messages
52
Reaction score
0
I have created a matrix with a class called Lattice. The lattice is filled with objects of type 'Dipole' which is created with another class. The problem I am having is with boundary conditions when I look for a neighbour. e.g If i pick a dipole on the top row, I want its 'above' neighbour to be the site on the bottom row etc...

Below is the .cpp and .h file which contains the functions I am having problems with. Note that the value passed to both functions can be negative. I thought the %xSize should take care of this but still running into problems. I believe the dipole functions are working correctly.

Code:
#ifndef lattice
#define lattice
#include "dipole.h"

class Lattice{

    private:
 
    Dipole* dipoleArray;
 
    double E;    // pointing in +ve y-axis V/m
 
    public:
        const int xSize;
        const int ySize;
        Lattice(const int, const int);
        ~Lattice();
        Dipole* GetDipole(int x, int y);  //  return a dipole
        double calcLocalEnergy( int x, int y);
 
 
};
#endif///////////////////////////////////////////////////////////#include "dipole.h"
#include "lattice.h"
#include <cmath>
#include <cstdlib>
#include<iostream>
Lattice::Lattice(const int xSize, const int ySize): xSize(xSize), ySize(ySize){
    dipoleArray =  new Dipole[xSize*ySize];    //store 1D array list
 
    for(int i=0; i<xSize; i++){
        for(int j=0; j<ySize; j++){
            //loop through the entire array and create dipoles
            dipoleArray[ i+j*xSize ] = Dipole();  // i + j*xSize to mimic 2D lattice form
         
        }
    }
}

Dipole* Lattice::GetDipole(int x, int y){

    return &dipoleArray[ (x%xSize) + (y%ySize)*xSize  ];    // modulo is for loop behaviour
}
double Lattice::calcLocalEnergy( int x, int y){
     
    double k = 1/(4*3.14158*8.85E-12);     double p1=0;     double p2 =0;        double r  = 1E-8;     double energy = 0;
    double E =1e4;    double pZero = 1e-27;    double pOne = 1e-29;
 
//    std::cout<<"0"<<std::endl;
    Dipole* central = GetDipole(x,y);  // define central dipole in the lattice at position [x,y] given by the argument
                                                    // take modulo because x,y can potentially  = -1 OR xSize for neighbour dipoles
                                                    // modulo ensures neighbours are found correctly
                                                    // for 0 <=  x,y < xSize modulo has no effect 
                                                     
    Dipole* neighbour[4];
    neighbour[0] = GetDipole(x,y-1);         //up = 0, right = 1, down = 2, left = 3
    neighbour[1] = GetDipole(x+1,y);
    neighbour[2] = GetDipole(x,y+1);
    neighbour[3] = GetDipole(x-1,y);
    int myDirection =  central->GetDirection() ;
    int type = central->GetType();
    if(type==0){p1 = pZero;    }    else{p1 = pOne;    }
 
    for(int i=0; i<4; i++){
     
        int neighbourDirection =  neighbour[i]->GetDirection();
        int Ntype = neighbour[i]->GetType();
        if(Ntype==1){    p2 = pOne;    }    else{    p2 = pZero;    }
    //    std::cout<<"1"<<std::endl;
        double    Energy[4][4]=
               {
                {-2*k*p1*p2/r/r/r - (p1+p2)*E, -p1*E, 2*k*p1*p2/r/r/r - (p1-p2)*E, -p1*E },    // ABOVE OR BELOW DIPOLES
                {-p2*E, 2*k*p1*p2/3/r/r/r, p2*E, -2*k*p1*p2/3/r/r/r    },
                {2*k*p1*p2/r/r/r + (p1-p2)*E, p1*E, -2*k*p1*p2/r/r/r + (p1+p2)*E, p1*E},
                {-p2*E, -2*k*p1*p2/3/r/r/r, p2*E, 2*k*p1*p2/3/r/r/r}
               };
        //       std::cout<<"2"<<std::endl;
       double    Energy2[4][4] =
               {
                { 2*k*p1*p2/3/r/r/r -(p1+p2)*E, -p1*E, -2*k*p1*p2/3/r/r/r -(p1-p2)*E, -p1*E },  // RIGHT OR LEFT ENERGIES
                { -p2*E, -2*k*p1*p2/r/r/r, p2*E, 2*k*p1*p2/r/r/r },
                { -2*k*p1*p2/3/r/r/r +(p1-p2)*E, p1*E, 2*k*p1*p2/3/r/r/r +(p1+p2)*E, p1*E },
                {-p2*E, 2*k*p1*p2/r/r/r, p2*E, -2*k*p1*p2/r/r/r }
               };
    //std::cout<<"3"<<std::endl;
    std::cout<<"i = "<< i << "\t" << type << "\t "<<Ntype <<"\t"<< neighbourDirection<<"\t"<<myDirection<<std::endl;
    std::cout<<"x = " << x << "y = "<< y <<std::endl;
            if(i==0 || i==2){
                energy += Energy[myDirection][neighbourDirection]; 
            }     
         
            else if(i==1 || i==3){          
                energy += Energy2[myDirection][neighbourDirection];         
            }     
    }//std::cout<<"4"<<std::endl;
    std::cout<<"4end clac local E"<<std::endl;
     return energy;
   
}
 
Lattice::~Lattice(){
 
    delete[] dipoleArray;
}
EDIT:: When it crashes it is always at an edge of the matrix. Usually the top row.
If my dipole is top row, it crashes when I find the 'above' neighbour.
When returning properties of this neighbour, "type" and "direction," defined in the dipole class to be 0,1 or 0,1,2,3 it returns large +ve or -ve values for the neighbour

PLEASE help as I can't see any problems :S
 
Technology news on Phys.org
I am not sure about the value of x%xsize if x = -1. Try using ((x+xsize)%xsize) instead. And, of course, the same for y%ysize.
 
I freaking love you! I just assumed negative modulo was fair game. Thanks a lot.
 
Learn how to step through your program line by line using your debugger, you'll be able to see things like that easily next time.
 

Similar threads

  • · Replies 5 ·
Replies
5
Views
3K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 15 ·
Replies
15
Views
4K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 36 ·
2
Replies
36
Views
6K
  • · Replies 23 ·
Replies
23
Views
3K
Replies
12
Views
2K
  • · Replies 6 ·
Replies
6
Views
12K
  • · Replies 35 ·
2
Replies
35
Views
4K