NREL's new high performance computer data center will reside in the Energy Systems Integration Facility, which is currently under construction. The data center will expand the laboratory's capabilities in modeling and simulation necessary to advance renewable energy and energy efficiency technologies. | Photo courtesy of Dennis Schroeder, NREL.
Data centers are the energy hogs of the computing world, but a new public-private partnership proves that doesn’t have to be the case. Working with Hewlett-Packard (HP) and Intel Corporation, the Energy Department’s National Renewable Energy Laboratory (NREL) is creating a new energy-efficient, high-performance computer (HPC) system in Golden, Colorado, to help increase efficiency and lower costs for clean energy technology research.
NREL’s new HPC system will save and recycle energy. The data center, which was designed to achieve an annualized average power usage effectiveness rating of 1.06 or better, will be 94 percent more efficient than an average data center. The computer lab will reuse about 70 percent of the waste heat from the computer systems as its primary source of heat for the offices and lab space -- and excess heat will also be used to warm adjacent buildings on the NREL campus.
All together, the efficiency of the data center, the energy efficiency features of the HPC system, and the system’s ability to reuse heat will combine to lower overall energy use, deliver substantial energy savings, and reduce significant costs.
This $10 million HPC system is already under construction and will provide additional computing resources to be used for energy systems integration, renewable energy research, and energy efficiency technologies. The new center will increase NREL’s modeling and simulation capabilities, and better enable research on advanced materials and biological and chemical processes.
In addition to advancing the lab's research capabilities, NREL's new HPC system will be a showcase data center facility demonstrating best-in-class technologies for a holistic approach to energy efficient high-performance computing.