lll.spad line 1 [edit on github]
This package implements LLL
reduction. Before use user must set appropriate floating point precision.Too low precision may cause nontermination.Higher precisions require more compute time.
LLL(m)
computes LLL
reduction of m
LLL
(m
) is the same as extendedLLL!(copy m
, ncols m
).
extendedLLL(m)
returns the LLL
reductions of m
together with the transformation matrix, i.e. if [l
, t
] is returned then it holds l
= t*m
.
extendedLLL!(m, n)
with 1 <=
n
<=
ncols(m
) performs the LLL
reduction of m
where only the first n
columns are considered significant. All matrix transformations, however, will be executed on the full matrix. m
is destructively modified and will be lost after this call.
find_relation(lf, k)
finds short integer relation between elements of lf
rounded to k
decimal places.