All You Need is Nodes: Supercomputing at Auburn
Auburn University’s new $1 million supercomputer is winning accolades not only for its lightning-fast research capabilities but for the way it came together.
Auburn’s new “Hopper” supercomputer boasts 16 terabytes of memory.
Auburn University’s new $1 million supercomputer, turned on for the first time last month, is winning accolades not only for its lightning-fast research capabilities but for the way it came together.
The new computer is named “Hopper,” in honor of the late Rear Admiral Grace Hopper, whose work laid the foundation for today’s computing world. Hopper’s computer power will be put to work on everything from microscopic gene sequencing to research projects that require terabytes of data, according to John Liu, Auburn University associate provost and associate vice president for research.
A typical home computer has eight gigabytes of memory and disk space of between 500 gigs and one terabyte. Hopper, a collection of 120 computers called “nodes,” is tied together into a coherent system controlled by a “head node” with its speed ranking in the top 1 percent in the world in terms of access to the large pool of shared storage. The system has more than 16 terabytes of memory and 1.4 petabytes of disk space. One terabyte equals one million megabytes, or 1,000 gigabytes, while one petabyte is 1,000 terabytes.
A small army of Auburn technophiles assembled the computer network, which runs the length and height of a classroom wall.
Researchers can store their work on the machine, which is housed in the Office of Information Technology building. Computer technology is estimated to become obsolete every six years, so Auburn plans to replace one of its supercomputers every few years.
Funding came from the National Science Foundation, Auburn’s Office of the Vice President for Research and Economic Development, Office of Information Technology, Office of the Provost, College of Sciences and Mathematics and Samuel Ginn College of Engineering.