Large integers multiplications

Hi All,

I would like to ask question regards multiplying large integers. I would like to multiply two integers of size 32 bits and then multiply the result by 2. The output size should be of size three words each of size 32 bits.

e.g.

uint32_t a, b; // a and b is of size 32 bits;
uint_t output; // here is my question, what size should i use for output? and how?

output = (a * b) * 2;


Then the size of output should be 32+32+32 (one bit only will be used from this 32) = 96 bits. How to create such output size? and how do define it?

Hope its clear.

Please support.

Thanks a lot.

Either do everything yourself manually(i.e. have a struct with three partitions and do the multiplication manually), or the better option is to find some BigNum library that can support large integers.

There might be a compiler specific extension you can use. For example GCC has a __int128, but the documentation says "There is no support in GCC to express an integer constant of type __int128 for targets having long long integer with less then 128 bit width." so I guess that's out (__int128 - Using the GNU Compiler Collection (GCC))

It wouldn't be very difficult to write functionality to multiply two 96 or 128-bit types together. You could use partitions (e.g. 32 bit types) or just use an array of chars. It might take a bit of time though... so probably worth trying to find a library. The GNU MP Bignum Library will do it -- but it looks pretty big for what you want.

I've been asked to write an __int128 extension for gcc by someone already. Still haven't gotten around to starting it yet though.
In the mean time, I have my own big integer libraries on my website, which is miniscule compared to other libraries.