Some sixty experts from five continents, but mostly from European countries and Japan, attended the NEC User Group meeting (NUG-XVII) hosted by the UK Met Office at their new site in Exeter, UK.
This get-together enabled experts in computing, meteorology and other technical engineering fields to share their latest research results, crystallize future hardware and software needs and collectively leverage NEC to take these onboard in their development plans for new systems. These needs are not only for faster more powerful (future petaflop/s) systems for the scientific and technical market, but increasingly also in data management and storage handling of petabytes size file systems. In short, an infrastructure to deliver a timely total solution to the user application.
Attendance, and consequently, the talks were mainly from computer center directors explaining how they manage their own computing facilities to support their core business. NEC support staff described new hardware and software developments. Technology strategy overviews from NEC executives were given explaining how their deep technologies as microprocessor and telecommunications vendors is leveraged to develop HPC future products. NEC is committed to serving the specific needs of customers, with vector supercomputers, delivering an overwhelming sustained performance, high reliability, leveraging cutting edge technologies to deliver seamless connection with other servers.
In the climate environment field, short-term climate prediction requires, 20Tflop/s and 10TBs of memory and this is satisfied by the Earth Simulator at present. Long-term climate prediction requires around 200Tflop/s and 100TBs of memory, while forecasts of local hazards requires around 1Petaflop/s and 500TBs of memory. NEC described how it is putting in place a program of developments to deliver this sustained performance on its journey toward Petaflop/s computing, in line with the recent Japanese government announcement
Below are a couple of extracts from these talks to give a flavor of the event, starting with a brief introduction of the host site, the UK Met Office, which should be of interest to the supercomputer community.
In the early 1850s, the world experienced an increase in stormy climatic conditions. The UK Met Office was set up in 1854 to provide weather prediction for logistical support to the navy and merchant ships transporting goods from the colonies across the oceans. This is not only an example of necessity being the mother of invention, but also an example where technology advances enable new developments in solving pressing problems. The enabling technologies in this case (1840s/1850s) were of course the discovery of electricity and the invention of telegraphy. This synergy between enabling technologies and advances in weather prediction at the UK Met Office is evident throughout its history. The BBC radio broadcasting in 1922, North Sea floods in 1953 and television weather broadcasts in 1954, weather satellites in 1964, MeteoSat in 1977 and the advent of supercomputing enabling the development of Global Weather Numerical modeling from 1982, are but a few examples.
With the advent of new technology, one trend in high performance computing is the fusion of computation, simulation and data analysis. With advance satellite technology delivering massive data streams in the earth systems and climate area, the challenges and opportunities for fusing observational and/or experimental data with classical simulation have increased enormously.
In recent years, meteorology evolved from its esoteric weather prediction role and has become a high profile e-business, with enormous commercial potential. For example, the UK Met Office employs 1150 staff and has revenues close to 300MEuros a year. Its revenue comes from various sources, roughly 37% from the UK Ministry of Defense and 23% from Civil Aviation. The UK Met Office is one of two centers providing weather prediction for civil aviation throughout the world, the other is in the USA. The rest of its business is divided, 16% to other UK governmental civil departments, such as agriculture, health and safety, environment pollutants and so on, 7% goes to climate research, 13% to commercial customers, including supermarkets for just in time shelf stock distribution, to farmers, to the building industry, to agencies putting public events such as formula one racing, open air music events and many others. The UK Met Office provides (sells) several hundreds of different products a day tailored at differing levels of refinement, to a diverse set of clients.
Because of global warming, the frequency of extreme events such as flash flooding are on the increase, so the UK Met plans to refine their local weather prediction model from a 12Kms grid down to 1Km to improve the accuracy of their prediction. To achieve this refinement in a timely fashion requires the most powerful and productive supercomputers available. At present informed opinion concurs, that the parallel vector processors in the NEC SX series, are most suited for this task. In the meteorology field the Total Cost of Ownership (TCO) of the NEC SX system compares very favorably with PC clusters, its productivity is not only unsurpassed, but it is streets ahead compared to any other system. In addition, its high integration and fast memory/communication bandwidth provides the capability for higher resolution global and regional numerical models and at the same time enables the inclusion of new science and different assumptions in the models.
The UK Met Office moved from Bracknell near Reading to their new purpose built building in Exeter early 2004. At the same time they began phasing out their older Cray T3E computers and installing new NEC SX series, supercomputers. Their current configuration for weather numerical prediction (NWP) consists of three NEC SX series clusters. Two clusters totaling of 34 SX-6nodes and one cluster of 16nodes of the latest SX-8 system. The total peak performance is close to 4Tflop/s, but more importantly it provides about 1Tflop/s, sustained performance to the user applications all year round. This is more than 5 times the productivity of equivalent peak performance of PC based systems. The newest NEC SX-8 has a very fast CPU (16Gflop/s) and this means fewer of them are needed to deliver the performance. This has enormous benefits across the whole spectrum of computing, as it simplifies programming, reduces operational workload balancing complexity and minimizes communication overheads.
The message from the UK Met Office presenters was that the NEC SX system's high performance, ease of use, and reliability resilience, surpassed their most optimistic expectations. As Stuart Bell, head of Numerical Weather Modeling (NWP) said: “The porting of our NWP codes to the NEC SX series systems, was relatively straightforward. Our major codes were benchmarked during the procurement process, so they more or less, work out of the box. The few problems encountered were on peripheral codes, data reformatting, moving data from Cray T3E and of course tuning the codes to take advantage of the shared memory vector components of the NEC SX systems”.
Stuart Bell went on to say: “The transition of the operational NWP system, from the SX-6 to the SX-8, was more straightforward than any comparable upgrade over the past 25 years. All the more remarkable when one considers that the Met Office SX-8 was first off the NEC production line. We are pleased with the stability and reliability of our NEC SX systems”.
Climate modeling is an important element of the UK Met Office activities under the auspices of the Hadley Climate Research Center. They are actively involved in research for the International Panel on Climate Change (IPCC) validating scenario simulations for the IPCC AR4 conclusions expected in 2007. One of the scenarios shows that Britain would be a much hotter place by year 2080. This has serious strategic implications for long term strategic economic planning.
Both NWP and climate research generate enormous amounts of data, presently running at 1.7PetaBytes per year, requiring archiving. Thus, the UK Met computer facility is a heterogeneous complex, has a number of other systems, NEC Itanium2 based TX-7s are used to handle the ever increase data, using a unified file system, as well as some IBM machines to support and handle data tapes.
The UK Met Office forecast center operates around the clock, 24hours a day and 365days a year. Robust reliability, continuous availability and the capability for timely delivery of forecasts, are mission critical essentials to the whole UK Met Office business. Judging from Stuart Bell's statements and also statements by other UK Met presenters, their choice of an NEC SX system, for handling their business, was an excellent one.
Prof. Dr.-Ing. Michael Resch, director of the High Performance Computing Center – Hoechstleistungsrechenzentrum (HLRS) in Stuttgart gave a thought provoking keynote presentation titled: “Supercomputing in the 21st Century”. Stuttgart has a long history in using the latest supercomputers. It is now operating a 72Nodes (576CPUs) NEC SX-8 system (9.216Tfop/s peak) delivering 4Teraflop/s, sustained performance. This is equivalent to one quarter of the Earth Simulator. It also has a series of commodity-based systems as for example, 256AMD-Opteron processors Cray XD1, NEC TX-7 system and other systems, used to support their file system, visualization and other elements of their operational environment.
HLRS supports R&D users from Europe, in the use of leading edge supercomputer technology and its applications. The mission of HLRS is to provide its users with tools and expertise to achieve top international positions in their research field. Capabilities and economy of scale are possible, at the high-end, through a joint operation of supercomputer systems with T-Systems Solutions for Research GmbH and Porsche AG, who provide about 20% of the HLRS income. For Porsche the benefit of using supercomputers is quantifiable, namely, as the shortest time to deliver a solution for say crash analysis.
Michael Resch's message on supercomputing was: “We want to use the fastest system available to get insights that are not possible with slower systems. We are willing (able) to spend the necessary funds to receive a much higher level of performance and be about 10years ahead of standard systems”.
Professor Resch also lamented at the failure of qualitatively improving supercomputer systems, saying that in recent years, despite the hype about exceeding Moore's Law, we are merely making them quantitatively faster. He went on to describe the current state of supercomputer hardware the difficulties of increasing clock frequency, cooling multi-cores and the inherent difficulties of increasing programming complexity. A more comprehensive exposition of his views and vision of supercomputing is contained in the interview of Michael Resch, soon to be published.
The importance of meteorology was reflected in the NUG program where the whole of Monday was allocated to the Special Interest Group for Meteorology Applications, (SIG-MA). Speakers came from weather and climate centers worldwide from Australia, Japan, Brazil and most of those in European countries.
Technology strategy overviews from NEC executives were given explaining how their deep technologies as microprocessor and telecommunications vendors is leveraged, to develop HPC future products. In addition, there were many interesting user talks, illustrating fascinating results from supercomputing centers. All in all, this meeting, like its predecessor meetings, in Yokohoma (Earth Simulator, 2002), Brazil (2003), Germany, (in 2004) was very fruitful said, Dr. D. Maric retiring President of NUG. The user community not only shared their own research and operational experiences, but also had the opportunity to leverage technology from all of NEC's businesses for their own benefit. For more detail readings of presentations visit: www.nug.ch
(Brands and names are the property of their respective owners)
Copyright: Christopher Lazou, HiPerCom Consultants, Ltd., UK.
June 2005.