let's allocate two-dimensional array of dynamic space:
i)  Define 'rowptr', a pointer to pointer to int, as the name of the 2D array.
ii) To pointer 'rowptr', let us allocate an array of 'm' elements of type 'pointer to int', Here, 'm' is the number of rows we wish the 2D array to have.
iii) To each element rowptr[i] allocated above, let us allocate an array of 'n' elements of type 'int'.  Here, 'n' is the number of columns in each row of the array.
lets take an example of rowptr[4][4].

code:

int **  My2DAlloc ( int  rows ,  int  cols ) 
{ 
    int  ** rowptr  =  ( int ** ) malloc ( rows * sizeof ( int * )); 
    for ( int  i = 0 ;  i < rows ;  ++ i ) 
        rowptr[ i ]  =  ( int * ) malloc ( cols * sizeof ( int )); 
    return  arr ; 
}

Above method use malloc function (rows + 1) times.Too much use of malloc function will affect the efficiency of the program.How to avoid this ?

Allocate all space once and then move pointer to appropriate position.

lets calculate total no of space we required
Total space = no of rows + no of rows*no of rows  /*( as we can see in above diagram)*/

    
    int header = rows * sizeof(int*);   //no of rows
    int data = rows * cols * sizeof(int);  //no of rows*no of rows
    int** rowptr = (int**)malloc(header + data); // total space

now we have to skip no of rows and then store data
let take previous example of rawptr[4][4]
Total space = no of rows + no of rows*no of rows
Total space = 4 + 4*4=20

now we have to allocate first 4 space for row array

so we have to skip till no of rows. To do this we are using variable buf.

int* buf = (int*)(rowptr + rows);

now we have to allocate address for  row array.

    int header = rows * sizeof(int*);
    int data = rows * cols * sizeof(int);
    int** rowptr = (int**)malloc(header + data);
    int* buf = (int*)(rowptr + rows);
    int k;
    for (k = 0; k < rows; ++k) 
        rowptr[k] = buf + k*cols;

lets assume size of int is 4 and we have 4 rows and columns

header=4*4=16
data=4*4*4=64
total space allocated for rowptr is 64+16=80 and assume address of rowptr is 100
but  we have to skip space till no of rows, for that we are using variable buf.
buf=100+16=116

pass 0:
rowptr[k] = buf + k*cols=116+0*4=116                       //Here we Assume Size of int is 4.
pass 1:                                                                      //Therefore every address is increment by 4.
rowptr[k] = buf + k*cols=116+1*4=132
 pass 2:
 rowptr[k] = buf + k*cols=116+2*4=148
pass 3:
rowptr[k] = buf + k*cols=116+3*4=164

done!!
we did it in one malloc function call.

final code:

#include <stdio.h>

int ** rowptr,* buf,header,data,rows,cols,k,i, j;

void ** malloc2d(int rows, int cols)
{
     header = rows * sizeof(int*);
     data = rows * cols * sizeof(int);
     rowptr = (int**)malloc(header + data);
     buf = (int*)(rowptr + rows);
    for (k = 0; k < rows; ++k) 
        rowptr[k] = buf + k*cols;
}

void free2d(int** rowptr)
{
    free(rowptr);
}
int main( )
{
    malloc2d(3,4);
    for( i = 0; i < 3; i++)
        for(j = 0; j < 4; j++)
  {
            rowptr[i][j] =1;
            printf("%d",rowptr[i][j]);
        } 
    free2d(rowptr);
    return 0;
}

output:-

111111111111

Have questions about this code? Comments? Did you find a bug? Let us know!