Thread: 32 bit floating point multiplication

Hi,

I am writing a program for multiplication of 32-floating point values. In order to find the significand, I will be dividing the initial number entered by the user by two, can you check if it is programmed correctly.

Code:

	if (floatingnumber < 0) {sign[0]=1; floatingnumber *= -1;}  else {sign[0]=0;}
//	printf ("\changing to positive bit %f\n", floatingnumber);
	printf ("\nsign bit:         %i\n", sign[0]);

	i=0;
	while (floatingnumber !=0) {
		floatingnumber *= 0.5;
		tempnumber = (floatingnumber * 10);
		modulus = tempnumber%10;

		if (modulus > 0){temporaryarray[i] = 1;}
		else {temporaryarray[i] = 0;}

//		printf ("\nnew digit = %i\n", temporaryarray[i]);
		floatingnumber -= (modulus/10);
//		printf ("\nnew digit = %f\n", floatingnumber);
		i +=1;
	}

secondly, i will mutiply the two numbers entered by the user, can you please check if the shift and add algorithm is correct:

Code:

void multiplysignificands (int mantissa1[24], int mantissa2[24])
{
int finalmantissa[57], mantissa3[57];
int i=0, j=0, k=0;
int shift=57, carry=0;
for (i=0; i<=56; i++)  {finalmantissa[i] = 0; mantissa3[i] = 0;}

printf ("\n"); for(i=0;i<=23;i++){printf("%i" ,mantissa1[i]);}
printf ("\n"); for(i=0;i<=23;i++){printf("%i" ,mantissa2[i]);}

printf ("\nmultiplying\n");

int zero=0;
shift = 0;
for(i=23;i>=0;i--)
{
	
	for(int j=0;j<=23;j++)
			{	mantissa3[i+j] = (mantissa1[j] * mantissa2[i]);
			printf ("%i", mantissa3[i+j]);	}
	printf("\nmultippy:  ");
  k=0;

	for (k=0; k<=56; k++)
	{printf ("%i", mantissa3[k]);	 }
	printf ("\nafter adding array:");

	carry=0;
	for (k=56; k>=0; k--)
	{
		if ((mantissa3[k] + finalmantissa[k] + carry) == 0) 
			{finalmantissa[k] = 0; carry=0;}
		else if ((mantissa3[k] + finalmantissa[k] + carry) == 1)
			{finalmantissa[k] = 1; carry=0;}
		else if ((mantissa3[k] + finalmantissa[k] + carry) == 2) 
			{finalmantissa[k] = 0 ; carry=1;}
		else if ((mantissa3[k] + finalmantissa[k] + carry) == 3) 
			{finalmantissa[k] = 1  ; carry=1;}
		printf ("%i", finalmantissa[k]);
	}
	printf ("\n");
	shift +=1;
	zero +=1;
}
}

Note: This is not the complete program.

Ill do some really fast wicked 32-bit multiplication for you.

Code:

float value1=100.0f;
float value2=200.0f;

float result=value1*value2;