KNNLSB: K Nearest Neighbors Linear Scan Baseline

Author: Georges Quénot (LIG-CNRS).
Last revision: 31-Mar-2010.

KNNLSB is a linear scan implementation for the k nearest neighbors search. It is especially optimized for multi-core architectures and it makes a very efficient use of processor caches. It is particularly optimized for the case in which the k nearest neighbors have to be computed for a set of query vectors.

KNNLSB is distributed under LGPL.

Download KNNLSB-1.00 (README).
Download KNNLSB-1.01 (README).
Download KNNLSB-1.02 (README).