1 barvinok is a library for counting the number of integer points
2 in parametric and non-parametric polytopes as well as projections
4 For parametric polytopes the count is represented by either
5 an explicit function or a generating function.
6 The first comes in the shape of a piece-wise step-polynomial.
7 This is a generalization of both Ehrhart quasi-polynomials
8 and vector partition functions.
9 The second is a generalization of an Ehrhart series.
11 This library shares some functionality with LattE
12 (http://www.math.ucdavis.edu/~latte/)
13 and uses PolyLib (http://icps.u-strasbg.fr/polylib/ or
14 http://www.kotnet.org/~skimo/polylib/) and
15 NTL (http://shoup.net/ntl/).
17 Additionally, the following packages can optionally be used
18 - PIP (http://www.piplib.org/)
19 - Omega (http://www.cs.umd.edu/projects/omega/ or
20 http://www.liacs.nl/~sverdool/gitweb.cgi?p=omega.git;a=summary)
21 - GLPK (http://www.gnu.org/software/glpk/)
23 Both NTL and PolyLib need to have been compiled with GMP
25 For NTL this means you have to specify
28 Suppose you want to put everything in /opt,
29 Then you configure/compile PolyLib using
31 ./configure --with-libgmp=/path/to/gmp --prefix=/opt
38 ./configure NTL_GMP_LIP=on PREFIX=/opt GMP_PREFIX=/path/to/gmp
44 vi -c '/DESTDIR=/|s/=.*$/=\/opt/|wq' Makefile.config
45 make libomega.a libcode_gen.a
50 ./configure --with-gmp=/path/to/gmp --prefix=/opt
54 Configure this library using
56 ./configure --prefix=/opt --with-polylib=/opt --with-ntl=/opt
60 ./configure --prefix=/opt --with-polylib=/opt --with-ntl=/opt --disable-fractional
68 If you want to put any one of these packages in the default location
69 (typically /usr/local), then the prefix argument can be omitted.
70 For more information on other arguments,
74 Please use PolyLib version 5.21.0 or later.
76 The input format of barvinok_enumerate is the same as that
77 of testehrhart from the PolyLib distribution.
78 See the PolyLib manual, barvinok_enumerate.c or the
79 examples in tests/ehrhart/ .
81 Note that the fractional representation is subject to change.
83 The library supports three specialization algorithms.
84 The first performs a direct specialization using a random vector.
85 The second performs a "depth-first" incremental specialization.
86 The third performs a "breadth-first" incremental specialization.
88 The first is usually the fastest, but can fail if the random
89 vector happens to be chosen incorrectly.
90 The incremental algorithms should never fail, but can be slower,
91 in some cases by quite a bit. The breadth-first version should
92 be faster than the depth-first version, especially on polytopes
93 of moderate dimension with a lot of "structure", but can require
96 Sven Verdoolaege <skimo@kotnet.org>