3.5.1.17.3 C++ Arbitrary Precision Integer Library

The C++ ap_[u]int type allows specifying signed and unsigned data types of any bitwidth. They can be used for arithmetic, concatenation, and bit level operations. You can use the ap_[u]int type by including the following header file.

#include "hls/ap_int.hpp"

The desired width of the ap_[u]int can be specified as a template parameter, ap_[u]int<W>, allowing for wider types than the existing C arbitrary bit-width library.

An example using the C++ library is shown below.

#include "hls/ap_int.hpp"

#include <iostream>

using namespace hls;

int main() {

    ap_uint<128> data("0123456789ABCDEF0123456789ABCDEF");
    ap_int<4> res(0);

    for (ap_uint<8> i = 0; i < data.length(); i += 4) {
        // If this four bit range of data is <= 7
        if (data(i + 3, i) <= 7) {
            res -= 1;
        } else {
            res += 1;
        }
    }
// iostream doesn't synthesize to hardware, so only include this
// line in software compilation. Any block surrounded by this ifdef
// will be ignored when compiling to hardware.
#ifndef __SYNTHESIS__
    std::cout << res << std::endl;
#endif

}

In the above code we iterate through a 128 bit unsigned integer in four bit segments, and track the difference between how many segments are above and below 7. All variables have been reduced to their specified minimum widths.