Hi,

In my program I see a lot of time taken up by add/subtract_unsigned. I have a couple of thoughts on the code:

 

Carry propagation outside the range of the smallest number can be improved:

      for (; i < x && carry; ++i)

         carry = ::boost::multiprecision::detail::addcarry_limb(carry, pa[i], 0, pr + i);

 

Maybe this:

      for (; i < x && carry; ++i)

         carry = ::boost::multiprecision::detail::addcarry_limb(0, pa[i], 1, pr + i);

 

This generates just an inc on my machine but unfortunately it doesn’t eliminate the loads and stores. I think on average the carry chain must be small though.

You have the same situation on subtract.

 

 

add_unsigned and subtract_unsigned are inconsistent at the tail end after carry propagation:

 

add_unsigned:

 

 

      if (i == x && carry)

      {

         // We overflowed, need to add one more limb:

         result.resize(x + 1, x + 1);

         if (result.size() > x)

            result.limbs()[x] = static_cast<limb_type>(1u);

      }

      else

         std::copy(pa + i, pa + x, pr + i); ß- if I == x we do the copy. We also do the copy even if the output and input overlap

 

subtract_unsigned:

 

      // Any remaining digits are the same as those in pa:

      if ((x != i) && (pa != pr)) ß------ we do these checks here and avoid the copy

         std_constexpr::copy(pa + i, pa + x, pr + i);

 

I am seeing a lot of time being used by that general copy above in add_unsigned  when it can be avoided.

 

This stuff is totally awesome and works well in my program. So thanks for working on this stuff.

Neill.