Algorithmic Optimisations #3
Description
- Instead of parallelising the enumeration (via subtrees) as we do, try to parallelise reduction+enumeration via re-randomizing the input basis and then running reduction+enum, thus having on each thread different bases. This should remind of an extreme pruning for SVP.
- Consider various types of pruning strategies optimising for the bounds given the Gram-Schmidt norms of a reduced basis.