Benchmark Hypergraphs and Detailed Experimental Results of "k-way Hypergraph Partitioning via n-Level Recursive Bisection"

This dataset contains hypergraphs derived from three benchmark sets: The
ISPD98 VLSI Circuit Benchmark Suite [1], the University of Florida Sparse Matrix Collec-
tion [2] and the international SAT Competition 2014 [3]. From the latter, we randomly selected
100 instances from the application track and converted them into hypergraphs as follows:
Each boolean variable (and its complement) is mapped to one vertex and each clause constitutes
a net [41]. The Sparse Matrix Collection is organized into 172 groups and each group contains
matrices of different application areas. From each group, we choose one matrix for each appli-
cation area that has between 10 000 and 10 000 000 columns. In case multiple matrices fulfill
our criteria, we randomly select one. In total, we include 192 matrices, which are translated into
hypergraphs using the row-net model, i.e. each row is treated as a net and each column as
a vertex. Empty rows are discarded. Both vertices and nets have unit weight. Together with the
18 ISPD98 VLSI instances , a total of 310 hypergraphs constitute our benchmark set. 4 Each of
these hypergraphs is partitioned into k ∈ {2, 4, 8, 16, 32, 64, 128} blocks with ε = 0.03. For each
value of k, a k-way partition is considered to be one test instance, resulting in a total of 2170
instances.

See the README for further information on the different files contained in this dataset.

[1 ]C. J. Alpert. The ISPD98 Circuit Benchmark Suite. In Proc. of the 1998 Int. Symp. on Physical Design, ISPD ’98, pages 80–85, New York, 1998. ACM.
[2] T. A. Davis and Y. Hu. The University of Florida Sparse Matrix Collection. ACM Trans. Math. Softw.,38(1):1:1–1:25, 2011.
[3] A. Belov, D. Diepold, M. Heule, and M. Järvisalo. The SAT Competition 2014. http://www.satcompetition.org/2014/, 2014.