how to do fixed point multiply C code?

[boeysue]

Is the following correct??

x.data=a.data *b.data;
x.wordlen=a.wordlen+b.wordlen;
x.fractlen=a.fractlen + b.fractlen;

??


Bcz My answer seems wrong...so help....~~

 

[echo47]

Maybe yes, maybe no. Need more info.
Please post a small, complete program, and tell us what answer you expected and what answer you got.

 

[nand_gates]

Your code is correct if
type of x.data is correct
What I maen is "if type of a.data and b.data is same lets say unsigned int (32 bits)
then type of x.data must be long long unsigned int (64 bits)
Hope this help

 

[echo47]

What I maen is "if type of a.data and b.data is same lets say unsigned int (32 bits)
then type of x.data must be long long unsigned int (64 bits)

That wouldn't work. You have to cast at least one of the multiplicands to the larger integer type *before* multiplying them.

 

[boeysue]

mmm..
What I want is a general equation for all Q.format ....such as [16 18],[16 0],..[16 15],..........Is there any information about it?thanks for that.