Combinatorial Scientific Computing and Petascale Simulations

Accelerating the development and deployment of fundamental enabling technologies in high performance computing

Alex Pothen (project webpage)
Purdue University

The coming era of petascale computing is filled with enormous potential for scientific simulation, but also with daunting challenges. Petascale machines of the near future are likely to have hundreds of thousands of processors, complex memory hierarchies and relatively poor interconnecting network performance. Thanks to the advances being made under the SciDAC program, the applications that will run on these machines will involve complex multiscale or multiphase physics, adaptive meshes and/or sophisticated numerical methods. A key challenge for scientific computing is obtaining high performance for these advanced applications on such complicated computers.

To address this challenge, the CSCAPES Institute (pronounced “seascapes”) will accelerate the development and deployment of fundamental enabling technologies in high performance computing. CSCAPES will focus on three broad areas:

  • providing advanced new capabilities in load balancing and parallelization toolkits for petascale computers
  • accelerating the development of new automatic differentiation capabilities for complex SciDAC applications
  • advancing the state of the art in sparse matrix software tools.

These seemingly disparate areas are unified by a common set of abstractions and algorithms based on combinatorics, graphs, and hypergraphs.

The Institute will train selected students in the methods of combinatorial scientific computing and multidisciplinary research and each participating student will have a DOE laboratory scientist serve on his or her thesis committee. These students will also work one summer at a DOE lab during their training period, helping them gain valuable experience in one domain science in addition to the computational sciences.

Training workshops will be offered in combinatorial scientific computing software every year of the project. SciDAC researchers and other interested researchers from academia, national labs and industrial partners will be invited to participate in these workshops, which will serve two purposes. First, participants will learn about the software tools the Institute provides for their applications; and second, the Institute will learn about applications and problems of users so that software tools can be made more responsive to their needs. Short courses on combinatorial scientific computing software tools will also be offered at SC and other parallel processing conferences. A bi-annual SIAM workshop on combinatorial scientific computing is the primary international meeting on this topic. CSCAPES researchers are actively involved in this workshop, and presentations there will extend the outreach of CSCAPES to an international community of researchers.

The CSCAPES Institute will work with other SciDAC research groups (from both computational sciences and the domain sciences) to integrate software tools into other codes and will reach out to other interested researchers. All software created under this project will be available for download from the CSCAPES website under an open-source public license.

Complementing these tangible software deliverables and their integration into application codes, this Institute is focused on outreach and educating the next generation of researchers in the application of combinatorial techniques to scientific computing. This inherently interdisciplinary area requires a diversity of skills and exposures that falls outside the scope of all but a select few academic programs. Therefore, CSCAPES will be a novel collaboration among groups of researchers in various areas of combinatorial scientific computing.

The SciDAC Institutes contribute to DOE missions by advancing computational sciences and broadening the community of practitioners capable of effectively utilizing the Department’s Leadership Class Facilities in areas that contribute to mission areas. Addressing these problems with algorithmic and software solutions and with education and outreach will enable new discoveries in applications that require the solution of discretized partial differential equations, numerical optimization, eigenvalue computations, and management of massive data sets, such as accelerator design, biological remediation, groundwater flow modeling, radiation transport and computational biology.

SciDAC Institute: Applied Mathematics

Project Title: Combinatorial Scientific Computing and Petascale Simulations (CSCAPES)
Project Website:

Principal Investigator: Alex Pothen
Affiliation: Purdue University

Project Webpage:

Participating Institutions and Co-Investigators:
Argonne National Laboratory - Paul Hovland, Boyana Norris, and Jean Utke
Ohio State University - Umit Catalyurek
Old Dominion University - Florin Dobrian
Purdue University - Alex Pothen (PI), and Assefaw Gebremedhin
Sandia National Laboratories, Erik Boman - Bruce Hendrickson and Karen Devine

Funding Partners: Office of ScienceOffice of Advanced Scientific Computing Research

Budget and Duration: Approximately $1.3 million per year for five years 1

Other Media:
Combinatorial Algorithms for Petascale Science, article in Issue 5 of SciDAC Review

Other SciDAC Institutes
Other SciDAC applied mathematics efforts

1Subject to acceptable progress review and the availability of appropriated funds


Home  |  ASCR  |  Contact Us  |  DOE disclaimer