Modification:

Code revision and optimization.

The (anti-)pre-kernel computation is now faster up to a factor 3 in serial as well as in parallel.

For instructions of a custom installation see v2.5.4 at the end.