Total Donates High-Speed Computer Cluster to UH


Total Donates High-Speed Computer Cluster to UH
Gift Will Expand University’s Computing Capabilities for Research and Education

The University of Houston’s Center for Advanced Computing & Data Systems (CACDS) is expanding its high performance computing capabilities, thanks to a computer hardware donation from Total, a leading multinational energy company.

Total Cluster
Cluster in actionThe cluster, with an estimated value between $500,000 and $1 million, includes 292 interconnected nodes that will occupy ten racks of space.

"Modern scientific research increasingly relies on large-scale computing and the processing of massive amounts of data,” said Ramanan Krishnamoorti, interim vice chancellor/vice president for research and technology transfer at UH. “This system will increase the competitiveness of researchers at the University of Houston. I look forward to continuing our collaboration with industry partners in the Houston area, in particular with TOTAL E & P USA, Inc.”

Total selected UH to receive the donation from among several companies and institutions in the United States. Criteria for awarding the donation included:

  • Ability to strengthen and sustain a research alliance/collaboration
  • Background history of collaboration between Total and the potential recipient
  • Potential recipient’s interest and needs for the cluster
  • Readiness of the recipient to integrate the cluster within its information technology data center facility.

“Total is proud to support research efforts in the places where we live and work around the world. The University of Houston is a Tier One public research university whose programs are an integral part of our community and economy here in Houston,” said Herve Coutrix, CEO and President of TOTAL E&P Research & Technology USA (TEP RT). “After thorough evaluation and discussion with the potential recipients, we decided to make the award to UH, which scored high in all categories under consideration.”

The systems are interconnected with InfiniBand technology, which provides a communication speed up to 40 gigabits per second. This high speed bandwidth will allow UH researchers to work on projects involving large model simulation and big data analysis.

Computer clusters of this size solve the largest computationally challenging problems, far surpassing the capabilities of individual PCs.

Total Cluster
From left, Jing Wen, Matt Bettinger and Harry Zong of TOTAL E&P Research & Technology USABarbara Chapman, director of CACDS, said the cluster will impact faculty and doctoral research projects in a diverse range of disciplines and will also be used by students in various course activities. Demand for the resources provided by the cluster comes from on-going research projects in chemical and biomolecular engineering, atomic physics, three-dimensional scans and analysis of archeological objects and computer science.

“This generous gift from Total will significantly increase the computing resources that CACDS makes available to the UH research community, which means faster and more impactful research,” Chapman noted. “We have several projects, including models of reactive dissolutions to stimulate oil and gas wells, simulations of biological physics and soft condensate matter, high resolution molecular design, computational catalysis and interface chemistry and mapping brain tissue alterations, that will make good use of this powerful resource.”

She said students will also benefit, gaining experience that can help them in the job market.

“At CACDS, we have a very active high performance computing (HPC) training component, which provides a competitive advantage to students who learn how to utilize these clusters for experiments,” Chapman said. “Some students also learn about HPC networking and system administration. CACDS and Total are helping to build a regional and national workforce.”

The cluster performs 26 teraFLOPS. “To put that in perspective, think of one computer operation as one addition or multiplication operation. This cluster can perform 26 trillion computations in a single second,” said Jing Wen, deputy vice president of advanced computing for TEP RT, who coordinated the Total donation.