Each party agreed to invest $33 million in the project, and for each it was money well spent. Along with furthering the research capacity and reputation for Rensselaer and IBM, the center greatly reinforced the state’s efforts to grow Tech Valley into a global hub of high-tech research and industry. Plus, the entire region would benefit from the economic development aspect of creating up to 300 new high-paying jobs to staff and conduct research at CCNI.
In late 2006, IBM, Rensselaer’s primary industry partner on CCNI, installed a 2,000-processor Blue Gene supercomputer at Rensselaer through the company’s Shared University Research program. It was a “prelude” to CCNI, allowing Institute faculty to get their bearings with the machine in anticipation of the construction of the 32,000-processor supercomputing center.
Rensselaer held a CCNI “sneak peak” in May for the trustees and the media. The Institute held an official grand opening Sept. 7, featuring prominent speakers and demonstrations of some of the early research conducted at the center. President Jackson also hosted a colloquy about topics related to CCNI, titled “The Future of Computationally Enabled Discovery and Innovation”.
CCNI features a 4,300-square-foot machine room with 16 IBM Blue Gene supercomputers, each boasting 2,000 processors, three AMD Opteron machines with 154 dual processors with dual cores, an Opteron machine with 40 four-way dual core processors, an Opteron symmetric multiprocessing (SMP) machine with a pair of four-way dual core processors, an IBM P-Series SMP with 16 processors, and an Intel SMP with 32 dual core processors. It takes 2.5 megawatts to power and cool the computer room. That’s equivalent to the electricity needed to power about 1,200 residential homes.
The CCNI computers will work in tandem with several other machines on campus. On top of all that processing power, CCNI has 832 terabytes of storage. Vice President and Chief Information Officer John Kolb ’79 says it has enough memory to store four digital copies of each of the 30 million books in the Library of Congress. Plus, there is plenty of room in the Tech Park building to install new Blue Genes and other machines to boost the center’s processing capacity and memory and keep it at the forefront of supercomputing.
While 80 percent of CCNI’s processing power comes from Blue Gene machines, a number of different architectures are also incorporated into the center. This heterogeneous assortment is unique among supercomputer centers, Kolb says, and will allow researchers flexibility to run their simulations on different platforms, which have different benefits and features.
A key feature of CCNI is its accessibility. Faculty, researchers, and students working on CCNI projects can use their laptop computers to pull or input data, check the progress of a simulation, or share that information with collaborators at other universities.