Sébastien Paolacci

This is indeed an often requested feature. Just having `Sum128` enhanced seemed a little bit "asymmetrical", so I added a full support in 4ec5a0f56d4fc178129a8433576bf6f2fe672a9e. Please tell me if that fix...

Well, I doubt there's a bug in the original implementation ;), but I could use a reproducer (with the C++ version you're using). I would suspect an API misusage given...

Well, it's difficult to say without a proper _equivalent_ C++ benchmark, but I would say not so bad. - for "large" inputs (i.e > 1kB), the 128/64 bits version is...