Gerhard Wesp wrote:
Whoever thinks about arbitrary precision arithmetic, please take a look at M.J. Kronenburg's proposal for TR2, which has been briefly discussed at the last WG21 meeting:
http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2004/n1692.pdf
Read it already. I also agree that it seems the most appropriate paper to influence the SoC work.
Please also look at NTL: http://www.shoup.net/ntl/ IMHO, NTL is one of the best and most complete library on the market, it would be very desirable (and maybe not even very difficult) to "boostify" it.
Certainly! Unfortunately, the license is incompatible for "in-place boostification". A reimplementation of that functional range is far beyond the scope of an SoC project. Anyway, you might find it interesting that my current SoC proposal http://tinyurl.com/qtyhh (still incomplete) would provide a solid fundament for algorithms like implemented by NTL (without requiring a procedural interface for maximum efficiency). Thank you for your comments! Regards, Tobias