Feeds

Brazil shells out for 244 teraflop Cray super

The price of better weather

The essential guide to IT transformation

Supercomputer maker Cray has pushed another XT6 massively parallel Opteron-Linux super out the door, this time to the Instituto Nacional de Pesquisas Espaciais in Brazil. You know, that South American country where they still have an economy that is roaring like an Asian Tiger.

The XT6 supers, which were announced last November at the SC09 supercomputing conference, sport Advanced Micro Devices' twelve-core "Magny-Cours" Opteron 6100s, which themselves started shipping at the end of March.

INPE says that the 244 teraflops XT6 cluster it is buying from Cray has more than 50 times the computing power of the gear it currently uses to forecast the weather. All that extra flopping will be installed in Cachoeira Paulista, in the state of São Paulo, and will be used by a variety of scientists for weather forecasting and long-term climate modeling.

INPE says that the new system will allow Brazil to forecast extreme weather, such as heavy rain, droughts, frost, and heat waves with "good reliability" and predict air quality and other environmental factors of weather with a higher resolution of 15km with a six-day forecast window. The super will allow the Brazilian weather forecasting service to do longer-range forecasting, with the level of detail down to 5km in South America and 20km for the entire globe.

(Which begs the question: Why are all the different nations of the globe paying to create a high resolution local forecast as well as global forecasts in lower res instead of building one shared, hi-res Interweather map? Oh, right. Because people think nation-states still exist, and because weather forecasting is of strategic value to the military...)

The last time INPE was on the Top 500 supercomputer list, it had an NEC SX-3/12R, rated at 3.2 gigaflops. This was back in 1995, which is ancient history in the computer market. In 2003, INPE installed a twelve-node SX-6 vector super from NEC (with 96 processors) and rated at 768 gigaflops of peak performance. More recently, NEC and Sun Microsystems (now part of the Oracle collective) partnered to make a cluster of Sun Fire X2200 blades rated at 4.4 teraflops sustained performance using 1,100 Opteron cores.

Cray says the INPE XT6 cluster will bring in $20m to its coffers; the deal includes services over an unspecified number of years. INPE is the first customer to ever buy an XT-class parallel super in Brazil, and it expects to have the machine up and in production late this year. The XT6 machine uses the SeaStar2+ interconnect, and given that the "Baker" systems with the higher-bandwidth and lower-latency "Gemini" interconnect are due in the third quarter, you might be wondering why INPE didn't hold out for a Baker box. Good question.

Assuming there is nothing out of the ordinary in terms of services, $20m for a 244 teraflops box seems kinda pricey. Cray just inked a deal for a petaflops-class super based on the Baker XT6/Gemini combo for $45m for Los Alamos National Laboratory. It looks like the supercomputer lab operated by the US Department of Energy got a pretty good discount on its XT6 machine - and got the faster interconnect thrown in too. INPE is paying $82 per gigaflops, while the DOE is paying $45 per gigaflops. This is the kind of deal you can strike when your budget basically makes a company's existence possible, as the US government, in many guises, certainly does for Cray. ®

Boost IT visibility and business value

More from The Register

next story
Pay to play: The hidden cost of software defined everything
Enter credit card details if you want that system you bought to actually be useful
HP busts out new ProLiant Gen9 servers
Think those are cool? Wait till you get a load of our racks
Shoot-em-up: Sony Online Entertainment hit by 'large scale DDoS attack'
Games disrupted as firm struggles to control network
Silicon Valley jolted by magnitude 6.1 quake – its biggest in 25 years
Did the earth move for you at VMworld – oh, OK. It just did. A lot
VMware's high-wire balancing act: EVO might drag us ALL down
Get it right, EMC, or there'll be STORAGE CIVIL WAR. Mark my words
Forrester says it's time to give up on physical storage arrays
The physical/virtual storage tipping point may just have arrived
prev story

Whitepapers

Top 10 endpoint backup mistakes
Avoid the ten endpoint backup mistakes to ensure that your critical corporate data is protected and end user productivity is improved.
Implementing global e-invoicing with guaranteed legal certainty
Explaining the role local tax compliance plays in successful supply chain management and e-business and how leading global brands are addressing this.
Backing up distributed data
Eliminating the redundant use of bandwidth and storage capacity and application consolidation in the modern data center.
The essential guide to IT transformation
ServiceNow discusses three IT transformations that can help CIOs automate IT services to transform IT and the enterprise
Next gen security for virtualised datacentres
Legacy security solutions are inefficient due to the architectural differences between physical and virtual environments.