Visit additional Tabor Communication Publications
November 19, 2008
Because of the compute and power density of petascale systems, all new supercomputer facilities are being built with energy efficiency in mind. This includes that new supercomputer center at the University of California at San Diego and the facility under construction at University of Illinois at Urbana-Champaign. The latter is being built to house the multi-petaflop "Blue Waters" supercomputer in 2011. Both datacenters will employ chilled water to be routed directly into the into the computer housing -- a much more efficient cooling method than forced air. State-of-the art cooling for petascale machines is now a given, but even industrial HPC datacenters are going green.
This includes the CFD centers being built for Formula One racecar designs. While racecars aren't exactly known for their fuel efficiency, there are plenty of opportunities to save energy when developing them. Most serious F1 teams now use high performance computers to help design these cutting-edge autos, so choosing the right HPC system and housing it in a well-designed facility can go a long way in minimizing environmental impact.
At SC08 this week, Appro announced it had completed the final deployment of 38 teraflop Xtreme-X supercomputer for the ING Renault F1 Team. The new system embodies pretty much the latest generation of cluster technology, with AMD quad-core nodes lashed together with DDR InfiniBand. The Appro machine represents a new level of commitment to HPC by the F1 team at Renault. Its previous machine was a 1.8 teraflop cluster housed in a conventional forced-air computer room. The new system lives in a brand new Computational Aerodynamics Research Centre located in the English countryside, north of Oxford.
The facility was built green -- not just in terms of energy efficiency, but also in regards to overall environmental impact. According to Graeme Hackland, the CFD center's IT manager, they were committed to operating an environmentally responsible facility from the start. And lessons learned from their previous computing facility led them to develop a much more energy-efficient plan.
Since the facility was built in the countryside, they had to negotiate with local farmers to bring the electric cable across their fields, while also working with Scottish and Southern Energy to get the energy onsite. "The cost of upgrading energy on this site is going to be huge, so the more we can do to reduce waste, the better it is," explained Hackland.
The whole structure, which includes the offices and the computer room, was built underground. Undoubtedly, this was more expensive to build than an above-ground structure, but it was still just one-fourth the cost of building a new wind tunnel, even taking into account the cost of the computer hardware. The unconventional design also presented another immediate advantage. The underground nature of the building meant they had no planning restrictions. The request for the new structure passed on its initial application. In the UK, where land is especially precious, there are many more land use restrictions than the US, so getting past the local planning commission is a big deal.
The other nice attribute of an underground facility is an evenly cool temperature. Once you get into the subsoil, the temperature varies very little from season to season, since the soil acts as an enormous thermal buffer. In the middle of England, the temperature below ground is about 10 degrees Celsius (50 degrees Fahrenheit). While this may be a bit chilly for humans, its pretty much perfect for sweaty supercomputers.
Of course you can't rely on the ambient temperature of the room to cool a multi-teraflop cluster, even at 50 degrees Fahrenheit. The Appro machine is water cooled, using APC's InfraStruXure solution, which allows them just to cool the hot aisle instead of the whole room. No forced air is used at all, saving even more energy. Furthermore, the CFD center operators have plans to recycle some of the waste heat to be used in the rest of the facility.
Presently the CFD center is using about 40 percent of its allotted power, so they have some room for further expansion. They're also counting on increases in performance per watt as new processors and systems are rolled out. Since the size of the datacenter is static, computational density is also important. Here again, they're counting on Moore's Law and clever system engineers to keep shrinking computers.
So is the Appro cluster performing as expected in its new digs? It's probably too soon to tell. The Renault engineers have only had access to the machine for production work since late summer. They've already used the system for some design mods for two of the races for this year's R28 F1 racecar, but the 2008 circuit is coming to a close. Most of the CFD design work is now being applied toward next year's R29. The first physical iteration of that car is expected before Christmas.
Wayne Glanfield, the CFD Analysis project leader, says with the larger system, they're able to run more simulations concurrently, vastly improving turnaround time for design explorations. They're also able to run much more refined simulations than they could with the 1.8 teraflop machine. On the old system only 10 percent of the aerodynamic design was done on the cluster, the remainder was accomplished with physical modeling in the wind tunnel. With the new system they're aiming for a 50-50 split. "We're currently running about three times the size of the model we were previously running," said Glanfield. "Our option was to build a second wind tunnel, or to do this -- to go for CFD in a really big way."
Jun 18, 2013 |
The world's largest supercomputers, like Tianhe-2, are great at traditional, compute-intensive HPC workloads, such as simulating atomic decay or modeling tornados. But data-intensive applications--such as mining big data sets for connections--is a different sort of workload, and runs best on a different sort of computer.
Jun 18, 2013 |
Researchers are finding innovative uses for Gordon, the 285 teraflop supercomputer housed at the San Diego Supercomputer Center (SDSC) that has a unique Flash-based storage system. Since going online, researchers have put the incredibly fast I/O to use on a wide variety of workloads, ranging from chemistry to political science.
Jun 17, 2013 |
The advent of low-power mobile processors and cloud delivery models is changing the economics of computing. But just as an economy car is good at different things than a full size truck, an HPC workload still has certain computing demands that neither the fastest smartphone nor the most elastic cloud cluster can fulfill.
Jun 14, 2013 |
For all the progress we've made in IT over the last 50 years, there's one area of life that has steadfastly eluded the grasp of computers: understanding human language. Now, researchers at the Texas Advanced Computing Center (TACC) are utilizing a Hadoop cluster on its Longhorn supercomputer to move the state of the art of language processing a little bit further.
Jun 13, 2013 |
Titan, the Cray XK7 at the Oak Ridge National Lab that debuted last fall as the fastest supercomputer in the world with 17.59 petaflops of sustained computing power, will rely on its previous LINPACK test for the upcoming edition of the Top 500 list.
05/10/2013 | Cleversafe, Cray, DDN, NetApp, & Panasas | From Wall Street to Hollywood, drug discovery to homeland security, companies and organizations of all sizes and stripes are coming face to face with the challenges – and opportunities – afforded by Big Data. Before anyone can utilize these extraordinary data repositories, however, they must first harness and manage their data stores, and do so utilizing technologies that underscore affordability, security, and scalability.
04/15/2013 | Bull | “50% of HPC users say their largest jobs scale to 120 cores or less.” How about yours? Are your codes ready to take advantage of today’s and tomorrow’s ultra-parallel HPC systems? Download this White Paper by Analysts Intersect360 Research to see what Bull and Intel’s Center for Excellence in Parallel Programming can do for your codes.
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?
Join our webinar to learn how IT managers can migrate to a more resilient, flexible and scalable solution that grows with the data center. Mellanox VMS is future-proof, efficient and brings significant CAPEX and OPEX savings. The VMS is available today.