SciNet Runs Canada’s Fastest, Greenest Computers

November 19, 2012 in blog, blog-general, frontpage

SciNet-run computers tops in Canada on Green500, Graph500, and Top500.
 
 

The two new Blue Gene/Q systems that are part of the Southern Ontario Smarter Computing and Innovation Platform (SOSCIP) for joint commercial/academic applications are Canada’s biggest, fastest, and greenest big compute platforms – indeed, among the greenest and fastest in the world.  And they’re in just the right place.

“SciNet was the natural choice to host, run, and support these supercomputers,” said Dr. Chris Loken, CTO of SciNet.  “We’ve built a centre that has the concentration of expertise to support users looking to make use of this system for research and development; and we have one of the largest, most energy-efficient, research computing datacentres in Canada, and still have lots of room to grow.”

SciNet’s green datacentre, which makes use of Canada’s cold winters to help reduce the costs and energy needed to cool these behemoth computers, means that the computer’s rankings on the twice-annual Green 500 list, where the machines are tied for 6th and 24th in the world for energy efficiency,  actually understates the case for how energy efficiently they actually run.  Because of judicious use of “free cooling” provided by Ontario’s mild climate whenever possible, almost all the energy used by SciNet’s datacentre goes into compute equipment, not air conditioning infrastructure.  Partly as a result, SciNet uses less than one half of the possible four megawatts of power to which the facility has access.

But although the systems only sip energy compared to similarly large-scale systems, they can tear through “big data” or raw number-crunching computational problems with ease.   On the Graph 500 ranking of supercomputers, which ranks the worlds largest computing resources by how well they can handle the sort of big-data problems that business analytics or digital humanities problems need, the new systems rank the 13th and 35th fastest in the world.  And for raw, brute-force number crunching capability, the larger of the two systems ranks 68th.

While researchers look eagerly forward to use such computational engines for discovery and innovation, some need some help in scaling their research software up to effectively utilize such powerful computers. “And that’s where SciNet really shines,” says Dr. Daniel Gruner, CTO-Software of SciNet. “We’ve got an amazing team of expert analysts – second to none – who can help researchers and innovative companies retool in order to take advantage of the largest machines in the world, and realize their full potential.”

More information about the computers, SOSCIP, and SciNet’s role can be found at

SciNet is Canada’s largest supercomputer centre, providing Canadian researchers with computational resources and expertise necessary to perform their research on scales not previously possible in Canada. SciNet powers work from the biomedical sciences and aerospace engineering to astrophysics and climate science, and is funded by CFI, NSERC, the Ontario Government, the Federal Economic Development Agency for Southern Ontario, and the University of Toronto.  SciNet runs computers for, and provides computational expertise to users of, Compute Canada’s National Platform, the Southern Ontario Water Consortium, and the Southern Ontario Smarter Computing and Innovation Platform.