One of the new supercomputers is an expanded version of BlueBioU, a system IBM donated to Rice in 2010 to foster joint biomedical research within the Texas Medical Center. An NIH grant enabled Rice to add 30 linked "nodes" to BlueBioU's original 18. Each node is a powerful computer in its own right and contains several of IBM's latest-generation processors. BlueBioU is built with the same hardware that IBM used to create Watson, the supercomputer that bested two world champions on the television game show Jeopardy! in February.
The other new system is DAVinCI, which is short for Data Analysis and Visualization Cyber-Infrastructure for Computational Science and Engineering Applications. DAVinCI is a 192-node IBM iDataPlex system designed as a workhorse that will serve multiple scientific disciplines. The DAVinCI project will eventually incorporate a 3D visualization lab that will let researchers literally walk through their data, viewing images from different angles and tagging elements for later study.
"While computing remains a key enabler for advancing science and engineering, managing data is increasingly becoming the bottleneck we have to address", stated Jan Odegard, executive director of the Ken Kennedy Institute for Information Technology (K2I), a multi-disciplinary research institute that helped Rice faculty win funding for DAVinCI, the BlueBioU expansion and several other supercomputers over past decade. "The relative cost for generating data is approaching zero, and as a result many laboratories are practically drowning in data. Our ability to extract useful information and knowledge from data is increasingly limited by our ability to effectively manipulate, move, manage and visualize data."
To address the data deluge, Rice has added a scalable, high-performance storage facility that will be shared by BlueBioU and DAVinCI. The storage system will eliminate the need for users to move or maintain multiple copies of research data. The shared storage facility uses hardware from Data Direct Networks and runs IBM's General Parallel File System.
Jan Odegard said all of the new systems are undergoing a final shakedown in preparation for full-scale operation later this month. "Too often we focus on how many processors a new cluster has or how many trillions of calculations it can perform each second, but the important thing to keep in mind is the scientific knowledge that it will produce", stated K2I Director Moshe Vardi, Rice's Karen Ostrum George Professor in Computational Engineering.
The dozens of planned research projects on BlueBioU include analyses of horizontal gene transfer in bacterial genomes, testing of computer models that examine the neurological processes that govern memory and the design of a revolutionary rotary pump for an artificial heart.
"High-performance computers are an essential tool for 21st Century biomedical research, and the NIH's decision to place this powerful new system at Rice is another tangible sign of Rice's increasingly important role in the Texas Medical Center", stated Kamran Khan, Rice's vice provost for information technology. "This sizes up well with our HPC infrastructure planning process and with the investment Rice has made to enable future growth of the supercomputing environment based on cutting-edge research."
For example, Kamran Khan said extensive expertise is needed to manage, run, and maintain high-performance computers like BlueBioU and DAVinCI. Rice's commitment to establish and grow its Research Computing Support Group is one key to the university's continued success in winning support for the increasingly powerful supercomputers that researchers need.
DAVinCI, which was funded via an NSF grant, will be used by dozens of investigators across campus. Planned research includes geodynamic modeling, simulation and testing of earthquake-tolerant bridge designs, life sciences research and seismic imaging. Alan Levander, the principal investigator on the grant and Rice's Carey Croneis Chair of Earth Science, said teams across campus are racing to be the first to use the new system.
"I can't even imagine who's going to be the fastest to get something going on it", Alan Levander stated.
The NIH and NSF funding for the new supercomputers was made possible by the American Recovery and Reinvestment Act.