In this article I’ve tried to go through the period of this Supercomputing systems development, from around 1970 until now. It also involves a brief account of giant machines which Supercomputer makers are planning to be design in the coming years.
Cray-1
Cray developed the first Supercomputer under the supervision of legendary Seymour Cray. It’s called Cray-1. Based on Freon cooling technique with a 64-bit architecture and working at 80 MHz with 8 megabytes of RAM. If used with proper care and the vector instructions are vigilantly shared with the system then it could produce a topmost performance of 250 megaflops. Including the Freon-based cooling system, it weighed 5.5 tons and was transported to the Los Alamos National Laboratory in 1976.
By the way, the C –shaped was not created to show that it’s the cray1 but to make the cables shorter on the inside that helped it operate at higher frequency 0f 80MHz, considered very fast at that time.
Cray-2
After the Cray-1, the Cray Corporation developed another giant named Cray-2. It remained the world’s fastest Supercomputer between the years 1985 to 1989, capable of performing 1.9 gigaflops. Like the Cray-1, it was also 64-bit organization. The first Cray-2 was produced with 2 gigabytes of RAM which was thought to be enormous at the time. Its circuit sheets were stacked in eight close-fitting, interconnected levels that are one above the other, resulting into any kind of cooling almost unmanageable. Instead a special electrically protective fluid called Flourinert from the 3M Company was put in to cool the circuits. Thanks to the heat exchanger water tank that came with the liquid cooling system, as seen in the pictures underneath, the Cray-2 got the nickname “Bubbles”.
The Connection Machine 5
The CM-5 Supercomputer was one of the machines initially designed by Thinking Machines Corporation and was to be used for artificial intelligence applications and symbolic processing purposes, but found more useful and efficient in the domain of Computational Science. In 1991, a well-known CM-5 structure known as FROSTBURG was put into action at the US National Security Agency (NSA), effectively functioning until 1997. The FROSTBURG computer was initially delivered with 256 processing nodes, but was improved in 1993 to an overall of 512 processing nodes and 2 terabytes of RAM. It made use of SPARC RISC processors and had a topmost performance of 65.5 gigaflops. It had cool-looking light panels that showed processing node usage and could also be used for diagnostics purposes.
A CM-5 was also featured in the film Jurassic Park in the island control area.
Fujitsu Numerical Wind Tunnel
Fujitsu Corporation made the supercomputer known as the Numerical Wind Tunnel in collaboration with Japan’s National Aerospace Laboratory. As its name implies, it was used to simulate wind turbulence on aircraft and spacecraft, and also to forecast weather. The NWT assembly was capable of 124.5 gigaflops and was the most powerful in the world when it was introduced in 1993. It was the first computer to reach the 100 gigaflops benchmark.
ASCI Red- Accelerated Strategic Computing Initiative
As a result of partnership between Intel and Sandia Labs, The ASCI Red was manufactured, that was the world’s fastest supercomputer between the years 1997-2000. The original ASCI Red used Pentium Pro processors clocked at 200 MHz which was later improved to Pentium II Overdrive processors functional at 333 MHz. This Supercomputer is exceptional for the reason that it was the one which went past the teraflops benchmark. After the upgrade it received later on it also crossed the speed of 2 teraflops. The machine had 4,510 processor nodes equivalent to 10,000 Pentium Pro CPUs, 1,212 gigabytes of RAM, and 12.5 terabytes of disk storage. This upgraded machine of was kept in 104 cupboards, taking up 230 square meters of ground space. The purpose of its creation was a US government initiative to substitute live nuclear experiments with simulation using these high speed computation machines.
ASCI White- Accelerated Strategic Computing Initiative
ASCI White took charge of Supercomputing industry as the world’s fastest supercomputer after ASCI Red and maintained that honor between years 2000-2002. Although similar in name, the ASCI White was completely different from ASCI Red. Intel had built the ASCI Red, but the ASCI White was an IBM corporation production, involving the use of IBM’s RS/6000 SP computer system. It housed an overall of 8,192 POWER3 processors operating at 375 MHz, 6 terabytes of RAM and 160 terabytes of disk storage. The ASCI White was capable of performing more than 7 teraflops.
The Earth Simulator
The Earth Simulator was assembled by NEC for the Japan Aerospace Exploration Agency, the Japan Atomic Energy Research Institute and the Japan Marina Science and Technology Center. It was the world’s fastest computer between years 2002-2004. The Earth Simulator is involved in studying global climate patterns and simulating them for both the atmosphere and the oceans. Based on NEC’s SX-6 architecture, it had an over-all of 5,120 processors, 10 terabytes of RAM and 700 terabytes of disk space plus 1.6 petabytes of mass storage based on tape drives. The initial Earth Simulator was capable of a calculated top performance of 40 teraflops. However, only 35.86 teraflops have been achieved when the LINPACK benchmark was run on it.
Blue Gene/L
IBM initially made the Blue Gene family of supercomputers to simulate biochemical processes relating proteins. The Blue Gene/L at the Lawrence Livermore National Laboratory (LLNL) was the world’s fastest computer between years November 2004 till 2008, when it lost the crown to another IBM project, the Roadrunner. In its existing arrangement, the Blue Gene/L at LLNL has 131,072 IBM PowerPC processors running at 700 MHz, a total of 49.1 terabytes of RAM, 1.89 petabytes of disk space and a calculated highest performance of 367 teraflops. A more polished and improved successor of this has attained a ultimate performance of 596 teraflops.
Jaguar
Now we’ll talk a little about the Jaguar Supercomputer that is located at Oak Bridge National Laboratory, This Cray giant is capable of performing 1.64 petaflops. This giant machine has been through many advancements and currently using Cray’s XT4 and XT5 architecture that is based on quad-core AMD Opteron processors.
It comprises of 181,000 CPUs, around 362 terabytes of RAM and 10 petabytes of disk space. The Jaguar inhabits 284 racks with the XT5 unit itself containing 200 cabinets, is a giant machine with size comparable to a basketball court. In year 2008.Performing at above 1 petaflops, it was ranked the second-fastest supercomputer in the world
IBM’s Roadrunner
The IBM Roadrunner is capable of performing 1.71 petaflops and has been the world’s fastest supercomputer since June 2008 and was the1st supercomputer which was able to maintain its performance to a 1 petaflops level.
It has 12,960 IBM PowerXCell 8i CPUs functioning at 3.2 GHz and 6,480 dual-core AMD Opteron CPUs operating at 1.8 GHz, resulting in a total of 130,464 CPU cores. It also contained about 100 terabytes of RAM. This supercomputer is kept in 296 frames and inhabits 6,000 square feet at the Los Alamos National Research Facility in New Mexico.
K computer
The K computer – termed for the Japanese word "kei" meaning 10 quadrillion – A supercomputer created by Fujitsu, currently set up at the RIKEN Advanced Institute for Computational Science campus in Kobe, Japan. The K computer is built on a system of distributed memory architecture i.e. each processor has its own share of memory allocated to it, with over 80,000 processor nodes.
Last year in June, the TOP500 classified K the world's fastest supercomputer, with a capability of performing over 8 petaflops, and after some improvements and up gradations in the architecture in November 2k11, K Supercomputer became the first computer to achieve a remarkable 10 petaflops. The K computer involves the use of 88,128 2.0GHz 8-core SPARC64 VIIIfx processors packed in 864 cabinets, for an overall of 705,024 cores, manufactured by Fujitsu with 45 nm CMOS technology. Each cabinet comprises of 96 computing nodes, in addition to 6 input/output nodes. Each computing node contains a single processor with 16 GB of memory each. K computer had achieved LINPACK top score with a performance of 8.162 petaflops. It is scheduled for completion in June 2012.
Pleiades
A Supercomputer built by SGI at NASA Ames Research Center in Mountain View, California. As of June 2011, it was the world's seventh fastest computer with a top performance of above 970 teraflops. Additional extensions are being done and it is anticipated in the supercomputing industry that it will touch the 10 petaflops mark in the coming year.
Also Read:
0 comments:
Post a Comment