multiplication of two 8bit number

Hi, i want to compare the two multiplier output and count the no. of error output. The multiplier width is 8 bit.One is exact multiplier other one is approximate (inaccurate) multiplier. I created the function for approximate multiplier,under the name MULTIPLIER_APPROXIMATE, i checked its functionality, it works as designed. Next, i have to count the no. of error outputs out of (2^8 x 2^8) outputs. I wrote the matlab code which is given below, am not pretty sure whether it works fine. Can someone suggest is this written script is correct one? I am new to this MATLAB platform, am tried my level best. Thanks in advance.

a=0;
b=0;
i=0;
j=0;
num_correct=0;
num_wrong=0;
check=0;
for a=1:256
    for b=1:256      
[c]=MULTIPLIER_APPROIMATE(a,b);
    end
end
for i=1:256
    a=i;
    for j=1:256
         b=j;
        check=a*b;
        if c==check
           num_correct = num_correct + 1;
            else
           num_wrong = num_wrong + 1;
                                %dis = e-check;
                end
    end
end

Answer 1

[Guillaume]

Well, if you'd tested your code with the debugger you probably would have quickly seen that it does not work.

You have a first set of loop that does nothing but repeatedly overwrite c. At the end c will only contain the value of the last iteration. So your second set of loop after that is not going to do what you want.

I don't particularly understand why you decided to go with two sets of loops. Why can't you do your check at the exact same time you calculate c?

[SHOBA MOHAN]

Thank you Guillaume. I got it. This is the required code for executing my requiremnts.

for i=1:255

      for j=1:255

c=i.*j;

check=MULTIPLIER_ACCURATE(i,j);
if c==check
          num_correct = num_correct + 1;
            else
           num_wrong = num_wrong + 1;
                                %dis = e-check;
end
                disp(c);
                disp(check);
                disp(num_correct);
                disp(num_wrong);
    end

end

[Guillaume]

Yes, this is much better (and a lot simpler)

[Reetika Banerjee]

can anyone tell how to precompute all possible multiplication of two 4 bit binary numbers?

[Walter Roberson]

[a, b] = ndgrid(0:15)
c = a.*b;

A = randi([0,15]);
B = randi([0,15]);
c(A+1,B+1);   %lookup answer