Marc Stevens

The basic framework for the SIMD module is implemented in the 'simd' branch. It exposes optimized vector operations through the following class: ``` class SIMD_operations { public: std::string simd_double_version() const;...

The SIMD module is implemented in the 'simd' branch which exposes optimized vector operations. However, these are not actually being profited from in fplll in matrix and vector operations.

The recent bug in PR #39 was not automatically caught by TravisCI, while it could have easily done so.

So I don't forget: More simplified code for `gauss_sieve` in https://github.com/fplll/g6k/blob/master/kernel/sieving.cpp is sketched as follows: ``` if (queue_begin == 0) queue_begin=1; // invariant: there are no possible pair-wise reductions for...

Currently for every distinct field we need to compile a distinct binary as the fieldsize is a template parameter. By making the integer size the template parameter and the fieldsize...

Merge collisiondetection program into main.