July 2012

On the origin of CSC

In the early 1990s, before there was a field called combinatorial scientific computing (CSC), the members of what would become today’s vibrant CSC community had to find each other.

“There was a collection of different communities using combinatorial ideas, but they were scattered across the landscape of scientific computing, and they all felt somewhat orphaned and underappreciated,” says Bruce Hendrickson, senior manager for computational sciences and mathematics at Sandia National Laboratories.

These communities included researchers developing new parallel algorithms for scientific computing, using geometric algorithms for mesh generation and exploiting structure in sparse matrix problems. At heart, each of these fields is based on combinatorial problems – how to determine the key relationships in a complex data set with graph-like properties.

In the late 1990s, Alex Pothen from Old Dominion University (now at Purdue) spent a sabbatical at Sandia and the seeds that would sprout CSC were planted. In 2000, Hendrickson and Pothen were leaders of about 30 international researchers who began formally discussing how they could facilitate intellectual connections.

In 2002 they chose the name combinatorial scientific computing and two years later they held the first SIAM workshop on the subject. The field has prospered with various journal special issues and a sixth workshop scheduled for next year.