Search All of the Math Forum:
Views expressed in these public forums are not endorsed by
Drexel University or The Math Forum.
|
|
Dan
Posts:
18
Registered:
8/2/07
|
|
Re: mex function for beginners
Posted:
Sep 1, 2007 10:32 AM
|
|
"basudha pradhan" <lucky_faith_911@hotmail.com> wrote in
message <fbb0lu$7b3$1@fred.mathworks.com>...
> I just started learning writing mex function. In C
function
> bitplan(), it take integer array as an input and it
convert
> each element of array from decimal to 8 bit binary. then
> MSB of every element is taken and written to a file,
after
> that next significant bit of every element is taken and
> written to a file. In this way integer array is separated
> into bit plane.
>
> I have tested the C function it works properly. But mex
> function that i wrote doesn't give the correct output.
> However it compiles with no error.
>
> Here is what i tried.
>
> #include <stdlib.h>
> #include <string.h>
> #include <stdio.h>
> #include <math.h>
> #include "mex.h"
>
> bitplan(int x[])
> {
> int i,j,k;
> FILE *fp;
> struct tag{
> int label;
> char bin[9];
> };
> struct tag data[100];
> char bvalue[9];
> int index;
> int temp_value, decimal_value;
> fp = fopen("qdata.txt","w");
> for(i=0;i<100;i++)
> {
> data[i].label=i;
> decimal_value=x[i];
> for(index=8;index>=0;index--)
> {
> // temp_value=decimal_value/pow(2,index)
> temp_value=decimal_value/(1<<index);
> if(temp_value>0)
> {
> bvalue[index]=(char)('0'+temp_value);
> // decimal_value=decimal_value%pow(2,index)
> decimal_value=decimal_value%(1<<index);
> }
> else
> {
> bvalue[index]='0';
> }
> data[i].bin[index]=bvalue[index];
> }
>
> for(j=0;j<8;j++)
> {
> for(k=0;k<100;k++)
> {
> fprintf(fp,"%c",data[k].bin[j]);
> }
> }
> fclose(fp);
>
> }
> }
>
> void mexFunction(int nlhs, mxArray *plhs[],
> int nrhs, const mxArray *prhs[])
> {
> /* Declare variables. */
> int elements, number_of_dims;
> int *pr, *pind;
>
>
> /* Check for proper number of input and output
arguments.
> */
> if (nrhs != 1) {
> mexErrMsgTxt("One input argument required.");
> }
>
>
> /* Get the number of elements in the input argument. */
> elements = mxGetNumberOfElements(prhs[0]);
>
> /* Get the data. */
> pr = (int *)mxGetPr(prhs[0]);
> bitplan(pr);
> }
>
>
> Any suggestion would be greatly appreciated.
>
> Thanks,
> Basudha.
How are you passing your variables into the mex function?
If I use your code and pass in a double matrix I get bad
data in pr. However if I cast the data as an int32 in
Matlab prior to passing the variable into the mex function
I get the proper data in pr.
It is probably best to use mxGetData instead of mxGetPr so
you get a void* back instead of a double*.
Dan
|
|
|
|
