Technically Speaking

The list is out and for the first time in a long while, Earth Simulator isn’t at the top of it. I posted a synopsis of the first ten on the list below. To see them all checkout the full listing.

Highlights from the Top 10:

• The list shows a major shake-up of the TOP10
• The new #1 is DOE’s IBM BlueGene/L beta system
  currently assembled and tested at the IBM Rochester site with a Linpack
  performance of 70.72 TFlop/s. This system ,once completed, will be
  moved to the DOE’s Lawrence Livermore National Laboratory.
• The Columbia system at NASA/Ames built by SGI gained the #2 spot, with an equally impressive 51.87 Tflop/s.
• The Earth Simulator, built by NEC and which held the #1 spot for 5 lists, is now #3.
• The #4 spot was captured by the new MareNostrum system at the Barcelona Supercomputer Center. It is an IBM BladeCenter JS20 based system with a Myrinet connection network and achieved 20.53 Tflop/s.
• The “SuperMac” is
  also back. The upgraded X-System at Virginia Tech built using Apple’s
  XServe boxes is at #7 now, with 12.25 Tflop/s performance.
• The entry level for the TOP10 approaches 10 Tflop/s - only
  one system with less than 10-TFlop/s Linpack performance is in the
  TOP10.

Quote Source

The still-incomplete Blue Gene/L system, designed for the analyzing our nation’s stockpile at the Lawrence Livermore National Laboratory, achieved a sustained performance of 70.72 trillion calculations per second. This speed is just shy of double the performance of Japan’s Earth Simulator. This is also the second time Blue Gene/L has claimed the world’s top spot in supercomputing power within the last couple of months.

In September, IBM announced that the Blue Gene/L prototype had sustained speeds of 36 trillion calculations per second. Last week, NASA announced that a system built by Silicon Graphics Inc. had topped that by sustaining 42 trillion calculations per second.

Both Blue Gene and the NASA computers are still unfinished, and the performance of both is expected to improve as more microprocessors are added.

Blue Gene, for instance, is just a quarter of its final planned size. When finished, it will exceed Earth Simulator’s performance by a factor of nine but require just a fraction of the electricity used by the Japanese machine.

Quote Source

The Cray XD1 is a highly modular Linux-based supercomputer, with great ease of scalability. The smallest complete base unit is a chassis. (This is where the name “mini” comes in.) It is a 12 processor Opteron machine. Up to 12 chassis can be installed in a rack. Multirack configurations integrate hundreds of processors into a single system. The system is a great marriage of Cray, Linux and AMD technologies.

One of the greatest features of the XD1 is its ability to self-heal whenever a problem is detected. It uses an elaborate fault detection system that monitors over 200 critical hardware functions as well as the Linux OS, predicting eminent failures and automatically recovering from them.

The Cray XD1 system provides extensive fault detection, isolation, and prediction capabilities, coupled with automated proactive and reactive self-healing intelligence.

Fault Detection: In each chassis, a dedicated management processor with its own super-visory network continuously monitors over 200 critical hardware functions, including temperatures, voltages, in-rush currents, parity errors, and component diagnostics. The management system also monitors the sanity and operation of the Linux operating system and key internal services such as DNS, NIS, and LDAP.

Proactive Management: Sophisticated proactive controls adjust a broad range of operating parameters to maintain peak performance and optimal operating conditions. The periodic refresh of system software in an SMP helps avoid problems with corrupted software. These proactive measures improve the mean time between failures (MTBF) and ensure system resiliency and job completion.

Recovery from Failures: The Cray XD1’s self-healing intelligence facilitates a quick and automated recovery in the event of a hardware failure, reducing outages from hours to minutes.

Redundancy features include “N+1 sparing” and the ability to reallocate resources in the event of a failure, enabling a replace-ment SMP to assume the persona of a failed SMP and restore full capacity to the affected partition. Jobs are automatically rescheduled from the last checkpoint.

Quote Source

Before you even look at the special technology, Cray makes sure that the system is adequately equipped.

The Cray XD1 features the direct connect processor (DCP) architecture, which removes PCI bottlenecks and memory contention to deliver superior sustained performance. According to the HPC Challenge benchmarks, the Cray XD1 has the lowest latency of any HPC system, with MPI latency of 1.8 microseconds and random ring latency of 1.3 microseconds. Tests conducted by the Ohio Supercomputer Center show that the Cray XD1 ships messages with four times lower MPI latency than common cluster interconnects such as Infiniband, Quadrics or Myrinet, and 30 times lower than Gigabit Ethernet employed in lowest-cost clusters. The Cray XD1’s interconnect delivers twice the bandwidth of 4X Infiniband for messages up to 1 KB and 60 percent higher throughput for very large messages.

The Linux/Opteron system runs x86 32/64 bit codes. Field programmable gate arrays (FPGAs) are available to accelerate applications, and the Active Manager subsystem provides single system command and control and high availability features. A 3VU (5.25″) chassis provides 12 compute processors, 58 peak gigaflops, 96 GB/second aggregate switching capacity, 1.8-microsecond MPI interprocessor latency, 84 GB maximum memory and 1.5 TB maximum disk storage. A 12-chassis rack provides 144 compute processors, 691 peak gigaflops, 1TB/second aggregate switching capacity, 2 microsecond MPI interprocessor latency, 922 GB/second aggregate memory bandwidth, 1 TB maximum memory and 18 TB maximum disk storage.

Quote Source

So if you or your company has the need for one of these systems you can pick one up in the neighborhood of just under $100,000 to about $2 million. This is really cheap as far as supercomputers are concerned.

On Tuesday, September 28, IBM announced that they have achieved a sustained performance from Blue Gene/L of 36.01 teraflops. This achievement puts Blue Gene/L in first place beyond Japan’s Earth Simulator which performs at 35.86 teraflops. One great difference between the two supercomputer is that Blue Gene/L is about 1% the size of Earth Simulator and power consumption differences are similarly minute at 3.6% of that of Earth Simulator.

This month, Japanese research laboratory AIST announced it had ordered a BlueGene/L supercomputer for use in its protein research.

A spokesman for NEC said he was not worried that IBM has batted the Earth simulator aside. He told The Register: “You have to remember that the Earth simulator has a peak performance of 40 teraflops, but that its actual operating performance is 90 per cent of that - just under 36 teraflops. The IBM machine is not as efficient.”

Supercomputers are rated in two categories: their theoretical peak performance, and their recorded maximum performance. The latter
determines their position on the Top 500 list of super computers, but the former sheds light on operational efficiency.

The NEC spokesman also reiterated statements made to the press this summer, that NEC is working on a new technology that will surpass its SX-6 supercomputer, but he could not say when it will launch.

Nonetheless, it is worth noting that the prototype on which IBM recorded its 36.01 teraflop score was built from 16,000 processors. A
bigger version, the size of half a tennis court, with 130,000 processors on board, will be built for the Lawrence Livermore National
Laboratory.

Quote Source

Met Office is presenting themselves a new supercomputer for their 150th anniversary. The supercomputer will be used to track weather systems around the world and provide more accurate forecasts for meterologists. The tracking capabilities will be on the order of “a massive dust storm to a single cloud.”

The Met Office, based in Berkshire, England, is a world leader in providing environmental and weather-related services, including public-access data, UK, World, and City forecasts, weather warnings, UV index, charts, marine information, and satellite imagery. Actually classed as a “Trading Fund” (the closest a U.K. government department can get to a commercial company) since 1986, the Met Office currently uses two T3E supercomputers built by Cray Inc. of Seattle.

The replacement from NEC boasts 6 times the power of the two Cray computers combined, with more nodes being added next year to take that to 12.5 times the power. Housed in the Met Office’s new headquarters in Exeter, the computer will initially consist of 30 NEC SX-6 nodes that include high-speed DRAM and Large-Scale Integrated Circuit (LSI), allowing for quicker data gathering from around the world and better atmosphere predictions. NEC will install 240 processors of the SX-6 series in parallel, each of which has a top-performance of 8 gigaflop/s (1 billion floating-point operations per second)–currently the fastest technical processor available.

NEC beat out several competitors for the contract, including IBM. Last month it also took the fastest supercomputer crown away from IBM with the Earth Simulator. IBM did hold the record with ASCI White, capable of 7 teraflops. NEC trounced this with the Earth Simulator’s 35 teraflops–Earth Simulator is technically very similar to new Met Office supercomputer.

Quote Source

The upgrade will increase the forecasting skill by 6 percent, which is considered significant. They look forward to the completed project in 2005.

Chip designers at Japan’s RIKEN have designed a processor that can perform 230 billion operations per second. The processor, named MDGrape 3, is the new hope for supercomputing reaching the petaflop, or a quadrillion operations per second range.

Samples of the chip, which was designed for life sciences research, can now perform 230 gigaflops, or 230bn operations per second, while running at 350MHz, better than standard general-purpose chips. In a worst-case scenario, the chip performs 160 gigaflops at 250MHz, said Makoto Tanji, a researcher with RIKEN’s high-performance computing group. Tanji spoke at the Hot Chips conference taking place at Stanford University.

The computational power comes, he said, because the chip is specialised for workloads that involve numerous, similar calculations on a comparatively small set of data. This sort of workload is common in the life sciences and bio-nanotechnology field, where researchers need to examine, for example, how a single protein interacts with thousands of different molecules. Consequently, the chip and the computers based on it can be directly compared with general purpose supercomputers only in a limited field, but the processor excels there.

…

Architecturally, the MDGrape 3 differs substantially from most other chips. It comes with 20 pipelines for calculations, the equivalent of an assembly line for a processor. Commercial chips typically have one or two. The chip also features what RIKEN calls a broadcast memory architecture, where data is force-fed to the different pipelines simultaneously. Parallelisation, a design convention that aims to cut down on redundant or parallel calculations, is optimised in the chip’s design.

Despite the differences from other chips, the MDGrape 3 is built on the 130-nanometer process, a manufacturing convention that has been in place for the past few years.

The enhancements lead to huge advantages over general purpose processors. Tanji said the 350MHz Grape 3 can provide a gigaflop of computing power for $15, compared with $400 per gigaflop for a Pentium 4, $640 per gigaflop for the chips inside IBM’s Blue Gene/L and a whopping $4,000 per gigaflop from NEC’s Earth Simulator, currently the world’s most powerful supercomputer.

In terms of power consumption, the 350MHz MDGrape 3 consumers 14 watts of power, or 0.1 watts per gigaflop. A 3GHz Pentium 4 runs at 82 watts, or 14 watts per gigaflop, he said. The Blue Gene/L chip and Earth Simulator come in at 6 and 128 watts, he said.

Quote Source

This news comes just as the US gears up to create the world’s fastest supercomputers, Department of Energy’s Oak Ridge National Laboratory (ORNL) designing a computer with a peak capacity of more than 250 teraflops and IBM BlueGene/L supercomputer is expected to yield 40 teraflops. These projects are due to produce results of unmatched speed as soon as 2005. With this new MDGrape 3 processor designs are already in the making for the first petaflop machine.

RIKEN is also designing the computer that will house the MDGrape 3. Twelve chips will fit on a board, while two boards will fit into a 2U-high box (3.5 inches). The chips are all connected to each other through an 81-bit bus, and the boards are connected to the rest of the computer through PCI Express.

The petaflop computer will consist of 6,144 processors on 512 boards clustered together. In all, the system will fit into 32 boxes that will stand on 19-inch pedestals.

“It is very small,” Tanji said.

Quote Source

An Army contractor, the Huntsville, Alabama based Colsa, announced that they will be buying 1,566 Xserve G5 system costing $5.8 million. This system, called MACH 5, will be used to run simulations of hypersonic aerodynamics of flight. Colsa hopes to have the system running at 15 teraflops by this fall. The achievement will put the MACH 5 supercomputer very high on the Top500 list (I think #2), which will be another great score for Apple supercomputing.

Much of the credit to Apple’s successes thus far is due to the processor it uses–IBM’s PowerPC 970–Illuminata analyst Gordon Haff said.

“The Macintosh software and the nice management features of OS X are factors here, but certainly performance of the processor is an enormous factor,” Haff said. “PowerPC is a fast processor.”

Indeed, the chip’s abilities to perform a type of mathematical calculation called “floating-point operations” were compelling. “The floating-point units in that processor were particularly attractive,” Whitlock said.

Colsa will use Mac OS X as the primary operating system, though it will evaluate other options including Red Hat Linux and Yellow Dog Linux, Whitlock added.

MACH 5, which stands for Multiple Advanced Computers for Hypersonic, G5, will occupy 42 racks and 600 square feet of floor space, said Anthony DiRienzo, a Colsa executive vice president. Apple was the winning bidder among six companies, DiRienzo said, declining to name the competitors.

Quote Source

Apple is quickly making a name in the supercomputing market. I predict that many seekers of supercomputer power will opt to look Apples direction for the reliability and cost effectiveness. Scalability and serviceability can also be major factors in going with Apple. With their supercomputer setup, components can easily be swapped out or added when need be. Watch out IBM!

This Saturday past, hundreds of computer volunteers descended upon a gymnasium to collectively attempt to connect their laptops together to create a powerful supercomputer. “Flashmob I,” was the brain child of researchers at the University of San Francisco and was designed to measure the computing power of an Ad-Hoc supercomputer. The first attempt at this experiment turned out very impressive results but fell short of the researchers goal of reaching 500 Gigaflops and thus fell short of being listed as one of the world’s top 500 supercomputers.

Saturday’s event was a dry run designed to measure how much computing power could be generated, rather than tackle a specific task.

About 660 people took part, including programmers, self-described computer geeks, teens, college students and researchers.

Organizers had hoped to break into the ranks of the world’s top 500 supercomputers by generating more than 500 gigaflops of power. A powerful PC can generate about half a gigaflop. The top spot is held by a Japanese computer that generates about 35,000 gigaflops.

Saturday’s event generated 180 gigaflops, not enough to make the Top 500. Still, organizers said they were pleased.

“This proves that this kind of computing can be competitive with computers that cost tens of millions of dollars,” Witchel said.

Quote Source

The organizers hoped to have roughly 1200 volunteers, wielding relatively new laptops, assist in the experiment but only 669. Plagued with technical troubles, the crew could only get 256 computers connected together. So with nearly a quarter of the requested amount of computers Flashmob I was able to produce a vast amount of computing power. If there is another attempt at this I’m sure that the result will be even more impressive, probably world class.

Read more here and here.

On Saturday 3 April, over 1000 laptop owners will converge on the university gym in an attempt to build a “flash mob” supercomputer. The project’s organisers hope that FlashMob will run fast enough to beat supercomputers in the list of the world’s top 500 supercomputers. This feat is significant because it will be the first “Ad Hoc” supercomputer. Connecting so many computers together is not a new idea, “Big Mac” was a collection of 1,100 dual-processor Power Mac G5s (it is now using Xserve G5 servers) and it is ranked as the #3 Supercomputer in the world. (Wired Magazine claims that it is ranked #2) The software that will connect these laptops is similar to that which controls Big Mac but there is one other variable that had to be taken into account, not all the laptops will be the same model.

A supercomputer is an array of processors that work in parallel to solve a computationally challenging problem. Examples include modelling weather systems to improve forecasts and global warming predictions, protein folding to discover new drugs or oil flow through the Earth to locate oil fields.

The processors’ memories must be “tightly coupled” so that each one can update itself continuously according to its peers’ calculations. Ordinary LANs used in offices and homes allow computers to share saved files of about 3 megabytes in size. But a supercomputer requires processors to share hundreds of megabytes a second, so FlashMob will require some clever software.

Researchers at Virginia Tech wrote software to link a 1100-strong cluster of identical off-the-shelf Apple computers in October 2003, and “Big Mac” subsequently entered the top 500.

But Witchel and his colleagues have an additional problem: “You essentially don’t know anything about the computers until they show they up that day.” The team therefore had to write code that not only allows the computers to share lots of data quickly, but also determines each processor’s speed and memory as it goes. This allows the computational tasks to be allocated in the most efficient way possible.

Quote Source

The attempt will be an amazing one and it may also bring meaning to the relatively “new sport” of flash mobbing.

I knew it wouldn’t take long before someone came up with the angle that would sell Virginia State University’s previously owned Apple G5 computers. The computers are a piece of history for they made the world’s third fastest computer and they signify the start of Apple entering the supercomputer business.

The Virginia State University system helped put Apple on the supercomputing map in a big way. The machine proved the merits of Apple’s G5 processor - aka IBM PPC 970 - and the company’s server aspirations. Apple’s Xserve tends to end up in mostly Apple shops, but the university work proved it could be an option for any customer in the high performance computing field.

But if the “Big Mac” is a piece of history, it’s only because the system sticks out as a shining example of success in an otherwise dull story.

Apple has yet to prove it can capitalize on “Big Mac’s” success in a big way. It will take some time to see if the Mac maker can close deals with other schools or big business.

That bit aside, customers can pick up a refurbished system for $2,799. It ships with dual 2.0GHz processors, 1GB of memory and a 160GB hard drive. The computers start shipping next week.

Quote Source

The comparable computer new from Apple will run you about $3000.oo, so if you don’t mind a used machine, slightly used, then I think you would be getting a pretty good deal.

The recently completed supercomputer of Virginia Tech is about to make a major change. All of the 1100 PowerMac G5 computers used to make the supercomputer will be swapped for Xserve G5 servers, one for one.

The new system, which went online toward the end of last year and which Virginia Tech said was the most powerful supercomputer at any university in the world at the time, will be completed by May.

By moving to the thinner servers, the supercomputer will consume less power and generate less heat, said Srinidhi Varadarajan, assistant professor of computer science, college of engineering, at Virginia Tech.

“It cuts the system’s size down by a factor of three,” Varadarajan said. “The new system will take much less power and generate less heat and free up space.”

The current supercomputer that uses 1,100 PowerMac G5 desktop computers occupies 3,000 square feet.

The price of this change has not been disclosed. The university is looking to find “good homes” for the PowerMac G5 machine that they currently have. If you are looking for a G5 machine, I’d suggest contacting the university.

If you have need for a supercomputer, Apple gives you a relatively inexpensive way of obtaining one. Xgrid is the latest product from Apple that is designed to let you easily connect several Macs together to harness their computing power.

Xgrid turns a group of Macs into a supercomputer, so they can work on problems greater than each individually could solve. You can let Xgrid operate in screensaver mode, so whenever you aren�t working, your Mac can crunch away at some data set. Or if you have a group of Macs dedicated to the task, Xgrid makes it easy to set up a cluster that works around the clock, every day of the year.

XGrid utilizes the Rendezvous networking solution which allows anyone to network their desktops or Xservers. Apple lives up to their reputation of easy computing with this product.

Xgrid uses zero configuration Rendezvous to discover available resources, so you never have to enter IP addresses to set up a cluster. An easy-to-use System Preference panel lets you control how your machine gets used by the network, and also tells the cluster which computer can send problems to the group for number crunching. Xgrid takes the grunt work out of splitting jobs and collecting results. Many scientists who already use command line tools in their work will immediately be able to take advantage of Xgrid and have the power of a cluster without the hassle of setup.

Xserve G5 is credited with the ability to crunch data with 9 GigaFlops of processing power. If you are able to cluster a few of these servers together think of the power you would have. To get a copy of the XGrid software, you can download it from here. It appears that the licensing is free, so all you have to worry about is the hardware.

The Big Mac supercomputer of Virginia Tech is now ranked 3rd fastest of all supercomputers.

Big Mac — the first supercomputer made of Macs — trails only Japan’s Earth Simulator and Los Alamos National Laboratory’s ASCI Q in the Top500, a list of the world’s 500 fastest machines.

The ranking will be unveiled Sunday at the opening of the Supercomputer Conference in Phoenix.

The machine, strung together from 1,100 dual-processor Power Mac G5s, is the first supercomputer made from Macs. It is also one of the cheapest supercomputers ever made, costing a relatively modest $5.2 million. The Earth Simulator cost an estimated $350 million to $400 million.

The machine is capable of performing 10.28 trillion operations a second (teraflops). The Earth Simulator operates at about 35 teraflops.

A spokeswoman for Virginia Tech said the college’s next bargain supercomputer, due to begin construction in 2006, will be shooting for 50 teraflops.

Meanwhile, IBM is working on a monster supercomputer that will easily rank as the world’s fastest supercomputer when it comes online next year. Blue Gene/L will be capable of performing 360 trillion calculations per second, or 360 teraflops.

I find it amazing how these supercomputer keep leapfrogging into position shooting for the #1 spot. IBM is working on several computers that will be up in the world’s top 10 fastest supercomputers. Here are the ones to watch for, ASCI Purple(100 teraflops), Wired Magazine

IBM is expected to announce the type of performance they are predicting for their dishwasher sized supercomputer. Performing calculations at 1.4 trillion per second, IBM believes will make this machine the world’s 73rd fastest supercomputer.

Blue Gene/L is a somewhat exotic machine that’s the first phase of an IBM project to tackle an intractable computing problem in genetic research: using the laws of physics to predict how crucial biochemical components called proteins fold from a long chain of building blocks into a complicated structure.

Blue Gene/L’s specialized processors currently run at 700MHz, but next year will be 40 percent faster, Bill Pulleyblank, director of IBM’s Deep Computing group and the executive overseeing the project. With the full configuration of 64 racks and roughly 131,072 processors, IBM hopes to achieve a speed of 360 trillion calculations per second–360 gigateraflops in supercomputing argot.

“Blue Gene” is an ambitious project to expand the horizons of supercomputing, with the ultimate goal of creating a system that can perform one quadrillion calculations per second, or one petaflop. Today’s fastest machine, NEC’s Earth Simulator is comparatively slow–about one-thirtieth of a petaflop–but fast enough to worry the U.S. government that the country is losing its computing lead to Japan.

IBM plans to have Blue Gene/L by 2005 and Blue Gene/P in 2006. Blue Gene/P is expected to be the first supercomputer to surpass the petaflop barrier, making us the country with the world’s fastest supercomputer.

Source: ZDNet

The godfather of supercomputing, Cray, plans to release a version of its Red Storm scientific platform, which promises to be the fastest computer in the world, for the enterprise market, next year.

The original Red Storm is being built under the US Energy Department�s Accelerated Strategic Computing initiative, for tasks such as simulating nuclear explosions in the Department�s Sandia Labs. It will be finished in mid-2004 and is expected to outpace the current fastest supercomputer, NEC�s Earth Simulator, with peak processing of 40 Teraflops, the equivalent of 40 trillion calculations per second.

A commercial version would be based on the same fundamental design � a massively parallel architecture based on 10,000 AMD Opteron Model 246 processors, the Linux operating system and the HyperTransport interface technology for chip-to-chip communication.

Source:Rethink Research Associates

Flash –

A small chip-design firm will unveil a new processor Tuesday it says will transform ordinary desktop PCs and laptops into supercomputers.

At the Microprocessor Forum in San Jose, California, startup ClearSpeed Technologies will detail its CS301, a new high-performance, low-power floating-point processor.

According to the company, the chip has the potential to bring supercomputer performance to the desktop.

An ordinary desktop PC outfitted with six PCI cards, each containing four of the chips, would perform at about 600 gigaflops (or more than half a teraflop).

At this level of performance, the PC would qualify as one of the 500 most powerful supercomputers in the world.

Source: Wired.com

This is not too surprising for China is making leaps and bounds in the technology field. The biggest project, as of now, is trying to get a man into space. That’s another story… They’re driven to make a name in the science/technology arena and in a big way. A supercomputer would be a logical tool to help them reach that goal.

Chip giant Intel Wednesday teamed with China’s Ministry of Education to build a national computing grid–a network of computers harnessed to work together.

When the grid is completed, the MOE expects it to have performance of more than 15 teraflops, or trillions of calculations per second, making it one of the world’s most powerful high-performance computing grids, according to a statement from Intel.

Intel will work with server vendors starting this year to provide the schools with computers equipped with the chipmaker’s Itanium 2 processors.

Completion of the project is slated for some time in 2006. At that time, 100 universities will be connected. The grid will be used to help make advances in various fields.

The grid will be used for work in life sciences, the petroleum industry, earthquake research and commercial financial projects.

It will also be used to help power the “Digital Olympics” initiative, to support the 2008 Olympic Games in Beijing.

Building the most powerful supercomputer is like a game of leapfrog. As soon as your machine is up and running, the next one is nearing completion. I have recently seen remarks about the world being a covered by massive grid supercomputers. It looks like we are seeing the start.

Source: CNET News.com

This is a very interesting topic that I came across a couple months ago. I have wanted to write about it but I have been afraid that I wouldn’t be happy with the outcome. So I am going to highlight a few aspects and let you read the rest from the source.

The world’s most powerful particle accelerator is being constructed at CERN, the European Organization for Nuclear Research, on the border between France and Switzerland near Geneva. The Large Hadron Collider (LHC), housed 100 meters (110 yards) underground, is some 27 kilometers (17 miles) in circumference and feeds several particle detectors, one as large as a six-story building.

In or around 2007, nearly 10,000 scientists and researchers from over 30 countries, will be using the collider. The data generate through these scientists experiments will be in the arena of 5 to 8 petabytes [PB] a year. At least 100PB will be needed for the analysis of this data. This is where Oracle comes top the rescue.

To meet these requirements, CERN has been architecting and developing systems for more than 10 years, and Shiers claims that the necessary computing system will be the largest ever assembled. The resulting computing environment, according to Shiers, will be distributed among participating countries, with about two-thirds of the system installed in regional computing centers spread across Europe, America, and Asia.

The computational system to support the LHC will be implemented as a global grid, integrating large, geographically distributed computing “fabrics” (networked computer clusters or farms) in a virtual computing environment.

The solution for this challenge is Oracle’s Grid Technology.
(more…)