Cannot multiply a 16-bit by an 8-bit to get a 32-bit ?

[John_Lintern]

I am using CCS PCW C Compiler in conjunction with MPLAB IDE to write a program for a PIC16C74B.

But I am having a problem multiplying 2 numbers together !

The result needs to be a 32-bit value and has to multiply a 16-bit value by an 8-bit value.

Take for example,

0x1234 * 0xA0 = 0x000B6080
(16bit) * (8bit) = (32bit)

But when I run the program the result gets truncated to a 16-bit value ??

So instead of getting 0x000B6080 I get 0x00006080

BUT..... if I declare ALL the variables as a 32-bit value it works !

This is ineffecient though because it uses up more memory - surely this can't be right ?

You should be able to multiply a 16-bit value by an 8-bit value to get a 32-bit value - shouldn't you ?

Below are the relevant pieces of code for my function 'Multiply':

unsigned int32 result=0x00000000; // 32-bit variable called result
unsigned int32 Multiply(unsigned int16,unsigned int8); //Multiply function
unsigned long A=0x1234; // 16-bit variable
unsigned int B=0xA0; // 8-bit variable

void(main)
{
result=Multiply(A,B);
while (TRUE)
{
}
}

// Multiply function
unsigned int32 Multiply(unsigned int16 X,unsigned int8 Y)
{
return(X * Y);
}

[kypec]

try this inside your function:

[Kieran]

I would cast both first then multiply

return (int32) x * (int32) y;

[Mark]

[John_Lintern]

[Mark]

It actually casts the 'x' to a 32 bit and then multiplies it. That is why it works. (int32)(x*y) would cast the result. However, you would still get the truncated result by just casting the result.