May 29, 2012
A new HPC center will be launched by years end in Massachusetts. The Massachusetts Green High-Performance Computing Center (MGHPCC) will be outfitted with terascale hardware aims to deliver it in an environmentally friendly way. Beyond promises of power efficiency and reduced carbon footprint, the center is deviating from typical facility models. It will act as a shared resource between multiple universities, requiring users to develop new strategies of implementation.
Last week, IEEE Spectrum likened the facility to a Thanksgiving table of HPC for colleges. University members include MIT, Harvard University, Boston University, Northeastern University and the University of Massachusetts system. The $95 million center will be located in the town of Holyoke and provide the necessary infrastructure to house and remotely access compute resources. This includes power, network and cooling systems. The universities will provide their own hardware and migrate research computing to the center.
Francine Berman, professor of computer science at Rensselaer Polytechnic Institute (and former SDSC director), equated the design to building a city instead of a skyscraper. She expects to see social challenges between member parties. “As hard as the technical, computational, power, and software problems are, social engineering of the stakeholders is dramatically difficult,” said Berman. Potential conflicts may involve the center’s ability to produce groundbreaking research and papers versus expanding its user base, which typically receives less of a spotlight.
John Goodhue, MGHPCC’s executive director, wants the facility to be simple and highly accessible to its users. This provides somewhat of a challenge as the universities currently house compute resources locally. Transitioning to an external facility adds a number of layers between the users and their equipment. He seems confident that the model will work though, saying the center will provide “ping, power and pipe” for its members.
The challenge, he says, is making the physical hardware behave as a set of private local machines for the various users. But thanks to high bandwidth network pipes and machine virtualization, that should now be possible.
The facility is undertaking an ambitious mission. Assuming it can meet the technical needs of its users, the center will also need to deal with universities that share different priorities. If the project turns out to be a success, it could become the model for future collaborative efforts.
Full story at IEEE Spectrum
10/30/2013 | Cray, DDN, Mellanox, NetApp, ScaleMP, Supermicro, Xyratex | Creating data is easy… the challenge is getting it to the right place to make use of it. This paper discusses fresh solutions that can directly increase I/O efficiency, and the applications of these solutions to current, and new technology infrastructures.
10/01/2013 | IBM | A new trend is developing in the HPC space that is also affecting enterprise computing productivity with the arrival of “ultra-dense” hyper-scale servers.
Ken Claffey, SVP and General Manager at Xyratex, presents ClusterStor at the Vendor Showdown at ISC13 in Leipzig, Germany.
Join HPCwire Editor Nicole Hemsoth and Dr. David Bader from Georgia Tech as they take center stage on opening night at Atlanta's first Big Data Kick Off Week, filmed in front of a live audience. Nicole and David look at the evolution of HPC, today's big data challenges, discuss real world solutions, and reveal their predictions. Exactly what does the future holds for HPC?