Section: Software

fpLLL: lattice reduction using floating-point arithmetic

Participants : Ivan Morel, Xavier Pujol, Gilles Villard.

FPLLL is a library which allows to perform several fundamental tasks on Euclidean lattices, most notably compute a LLL-reduced basis and compute a shortest non-zero vector in a lattice. The latter functionality has been added in July 2008 by X. Pujol at the end of his Master's degree internship with D. Stehlé. This led to the release of version 3.0 of fpLLL. The fpLLL library relies on floating-point arithmetic for all the computations involving the Gram-Schmidt orthogonalisation of the lattice bases under scope.

FPLLL is becoming the reference for lattice reduction. It is, or adaptations of it are, included in SAGE, PARI GP and MAGMA.

Status: stable / License: LGPL / Programming Language: C++ / URL: http://perso.ens-lyon.fr/damien.stehle