Monday, September 17, 2012

India planning world's fastest supercomputer by 2017


As per a latest update, Indian government has drawn a blueprint for development of the next generation supercomputers, which could be 61 times faster than existing machines. 

It is reportedly said that the Telecom and IT Minister Kapil Sibal has shared a roadmap with Prime Minister Manmohan Singh to develop fastest supercomputer. According to the letter, the 'petaflop and exaflop range of supercomputers' will be developed at an estimated cost of Rs. 4,700 crore over 5 years in India. 

In addition, Sibal has proposed that Department of Electronics and Information Technology (DEITY) should be given chance to coordinate overall supercomputing activities in the country. 

As a matter of concern, it is found; petaflop is a measure for finding processing speed of a computer and can be expressed as a thousand trillion floating point operations per second. As far as exaflop is concerned, it is represents one quintillion computer operations per second. In simpler words, one exaflop is thousand times faster than one petaflop.

Presently, the fastest supercomputer of the world is Sequoia, which has registered a top computing speed of 16.32 petaflops which is equivalent to computing of power from over 7.8 lakhs high-end laptops put together, sources informed.

“EKA” is the fastest computer in India, developed by Computational Research Laboratories (CRL), wholly owned subsidiary of Tata Sons. On a global front there is nothing much to write about, as India has been falling rapidly over the years as far as the India’s prowess to build super-computer goes.

In 2007, India’s “EKA” was 4th fastest supercomputer in the world and number 1 in Asia, but since then, it has not made much progress. During the same time countries like China & Japan have built super computers that are more than 50 times faster. India’s fastest computer “Eka” currently ranks a lowly 58th on the global level.

SAGA-220 Recently unveiled supercomputer SAGA-220 built by ISRO, is capable of performing at 220000 GFLops (220 TFlops). It uses about 400 NVIDIA Tesla 2070 GPUs and 400 Intel Quad Core XeonCPUs.[6]


No comments:

Post a Comment