I have written a Boost-compatible library for the representation and manipulation of integers of large or arbitrary size (including 128bit and above). I would like to make a preliminary submission of this library for your review and input, for potential inclusion into Boost v1.51.0 or later, as per the Boost Library Submission Process.
It would have been great if you could have joined the discussion a couple of weeks ago. Perhaps it is not too late for you to write a review of the proposed Boost.Multiprecision, since you clearly have some experience of the subject.
That would be good. This is such a rich and multifaceted topic---fast big integers, floats, etc. Remember, we're far from done with big integers and floats in boost and C++. It's tough nut to crack. But in my opinion, we seem to be making progress.
- The storage of these types is just the actual numeric data, so I should be able to do binary I/O on them etc., which is good. I would perhaps be a bit more confident of this if it were just an array of int32s or int64s, rather than the current recursive organisation.
In the past, I wrote both kinds: classes based on recursive templates as well as classes based on traditional arrays of limbs. Both the performance as well as the resulting code size of the recursive template versions were disappointing compared with traditional arrays of *limbs*. I really can't make a time commitment now. But in the fall, I'd be happy to give you any codes that I have or discuss performance or size issues within my modest range of experience. Best regards, Chris.