Tag: supercomputer

  • India’s supercomputing capabilities fall behind its friends

    From having two of the highest 100 supercomputers on the earth in 2020, the nation now has none in that bracket, and solely three within the prime 500, in line with a worldwide rating service for the quickest supercomputers on the earth, referred to as Top 500.

    Supercomputers can speed up calculations by offering exponentially bigger computing energy than common machines. In May 2020, researchers at US Oak Ridge National Lab ran 1000’s of simulations to seek out drug compounds that might cease the covid-19 virus from infecting host cells. They accomplished the simulations in simply 2-3 days utilizing IBM’s Summit supercomputer.

    “Supercomputers will be utilised for numerous issues in sure areas like chemistry formulations, protein folding, biomedicine. They will also be utilized in house for satellite tv for pc placements, that are very, very troublesome to resolve with classical computer systems,” mentioned Nikhil Malhotra, chief innovation officer at Tech Mahindra.

    Countries around the globe, together with India, have been ramping up supercomputing capabilities over the previous few years. This yr, the Ministry of Electronics and Information Technology (MeitY) has put in 4 supercomputers at numerous establishments within the nation. Fifteen supercomputers, with an mixture compute capability of 24 petaflops, have been put in within the nation since 2015 beneath the National Supercomputing Mission (NSM).

    According to consultants, constructing native supercomputing capabilities will strengthen high-quality analysis in biomedicine, house tech, and local weather, which require excessive compute energy. However, in addition they identified that the out there computing energy of Indian supercomputers is significantly decrease than their world counterparts.

    India’s first supercomputer, PARAM 8000, was arrange in 1990. Param Siddhi, which affords a peak efficiency of 5.27 PFlops, had the excellence of being ranked because the 63rd quickest supercomputer on the earth when it was launched in November 2020, in line with the Top 500 world rating service. It has since slipped to the 111th spot.

    Similarly, the Indian Institute of Tropical Meteorology’s (IITM) Pratyush supercomputer, which was once within the prime 100, is now ranked 132. Overall, India has three supercomputers within the prime 500’s total record. China and the US account for almost two-thirds of the highest 500 supercomputers on the earth, with 173 and 128 of the five hundred strongest supercomputers on the earth, respectively.

    The computing energy of a supercomputer is measured in floating-point operations per second, or FLOPS. One PetaFlops is the same as 1,000,000,000,000,000 (one quadrillion) FLOPS, or one thousand TeraFlops. The world’s quickest supercomputer, Frontier, positioned on the Oak Ridge National Laboratory, affords a peak efficiency of 1,685 PFlops.

    “The baseline behind each supercomputer is quicker efficiency. The world is shifting to ExaFlops and we’re nonetheless on PetaFlops,” said Preeti Syal, a tech policy expert, who used to work with Niti Aayog. “We are always two years behind. If we continue to play catch up we will always remain behind.”

    Suryachandra Rao, undertaking director at IITM, famous that leaders like China and the US are taking a look at exascale degree computing and that’s the subsequent milestone that India additionally has to succeed in. However, he additionally identified that India faces a twin problem of increasing the aptitude and in addition making such sources for extra researchers and institutes. “It is a step in the precise path,” he said.

    “At this moment the idea is to enable more people to use such resources. Ten years back, we used to have 1 or 2 systems. Now we are setting up at different locations. This will build capacities and allow more users to access it,” he added.

    Rao mentioned earlier climate and local weather forecasts have been the one precedence. Now there are numerous different purposes for supercomputing.

    But along with having supercomputers, having the newest {hardware} is equally essential. Syal famous that how India is making these supercomputers or the parts utilized in them can be essential to take their efficiency of supercomputers to the following degree.

    Building extra instances is the following problem for the NSM. Syal mentioned at present the use instances are restricted. “Globally, many of the utilization comes from trade. What we’re doing as an alternative is letting analysis establishments drive it. They won’t ever have the ability to assume when it comes to commercialization. These analysis establishments must also be linked on to large trade gamers.”

    Budget is one other problem. Rao mentioned his institute is engaged on rising the capability of Pratyush. Though, he mentioned, their requirement is at the very least 150 PFlops they are going to have the ability to get someplace round 10-20 PFlops attributable to finances constraints.

    Centre for Development of Advanced Computing (C-DAC) didn’t reply to Mint’s e-mail question until the time of writing. A division of MeitY, C-DAC is liable for designing, creating, and commissioning supercomputers beneath NSM.

    Subscribe to Mint Newsletters

    * Enter a legitimate e-mail

    * Thank you for subscribing to our e-newsletter.

    First article

  • C-DAC, IIT Gandhinagar unveil Param Ananta, India’s newest supercomputer

    The Centre for Development of Advanced Computing (C-DAC) and the Indian Institute of Technology, Gandhinagar have collectively unveiled a brand new supercomputer in part two of the central authorities’s National Supercomputing Mission (NSM). Called Param Ananta, the supercomputer is able to providing peak efficiency of 838 teraflops.

    Sanctioned in October 2020, Param Ananta will provide IIT Gandhinagar with elevated capability for analysis initiatives in numerous fields, together with machine studying, information science, computational fluid dynamics, bioengineering and extra.

    According to a press release by the Ministry of Electronics and Information Technology (Meity), the supercomputer has a blended set of CPU and GPU nodes, together with excessive throughput storage and excessive bandwidth reminiscence modules. It can be stated to supply direct contact liquid cooling expertise to extend efficiency effectivity, though it isn’t clear as to what efficiency per watt does it provide.

    The supercomputer will rank behind C-DAC’s Param Siddhi-AI, which as of November 2021 was the 102nd strongest supercomputer on the earth — with peak efficiency functionality of three.3 petaflops.

    With Param Ananta, India now has 15 supercomputers disclosed to the general public, with mixed efficiency functionality of 24 petaflops.

    The announcement curiously comes on the identical day when USA’s Frontier supercomputer, run by the Oak Ridge Leadership Computing Facility, was formally ranked as essentially the most highly effective supercomputer on the earth. The latter outperformed Fugaku, now the second strongest supercomputer and operated by Japan’s Riken Centre for Computational Science, to take the highest spot.

    Frontier has been labelled because the world’s first ‘true’ exascale supercomputer, able to producing total computing energy in a number of exaflops. In direct comparability, C-DAC and IIT Gandhinagar’s newest Param Ananta supercomputer gives a fraction of the online efficiency functionality that the world’s strongest publicly recognized supercomputer can produce at present.

    The official checklist of the world’s high 500 supercomputers shall be printed in Hamburg, Germany on June 1.

    Subscribe to Mint Newsletters

    * Enter a sound e-mail

    * Thank you for subscribing to our e-newsletter.

  • US’ Frontier overtakes Japan’s Fugaku as world’s strongest supercomputer

    The world’s first exascale supercomputer is not only the world’s strongest by a margin however additionally it is essentially the most environment friendly. Frontier, developed by US’s designated supercomputer facility, the Oak Ridge Leadership Computing Facility, is the world’s first supercomputer to cross the efficiency margin of 1 exaflops. With this, the supercomputer has surpassed Fugaku, belonging to Japan’s Riken Centre for Computational Science, because the world’s strongest supercomputer proper now.

    Frontier’s nearest competitor, Fugaku, has a efficiency rating of 442 petaflops on the Linpack benchmark – a typical for formally rating world, publicly disclosed supercomputers. While Fugaku was based mostly on Arm’s core designs, US’s Frontier is powered by AMD.

    Interestingly, Frontier can be mentioned to be essentially the most energy environment friendly supercomputer constructed until date. According to stories, Frontier has an influence consumption ranking of 52.2 gigaflops per watt. This rating outdoes essentially the most environment friendly supercomputer until date, the Preferred Networks MN-3, which supplied 39.4 gigaflops per watt in energy ranking.

    The newest listing of the world’s high 500 supercomputers is about to be unveiled in Germany within the coming days, and is taken into account to be the ‘official’ listing of the world’s strongest supercomputers.

    To be certain, Frontier has been constructed as the primary ‘true’ exascale supercomputer. The latter is a reference to traditional supercomputers (i.e. computing techniques not constructed on quantum computing ideas) that supply efficiency functionality of greater than 1 exaflop. While Japan’s Fugaku was billed as the primary exascale supercomputer, Frontier can seemingly apply its peak, 1.1-exaflop computing energy in real-world duties and never simply in benchmarks.

    In its earlier itemizing as of November 2021, India’s highest rating within the supercomputer listing got here in at 102, with the Param Siddhi-AI. The latter belongs to India’s Centre for Development of Advanced Computing (C-DAC) – that gives peak efficiency of three.3 petaflops on the Linpack benchmark scale.

    Interestingly, standard supercomputing stands at some extent the place it’s geared to develop even additional in its efficiency skills – making exaflop-scale efficiency extra common amongst supercomputers, and approaching zettascale computing. It is, nonetheless, additionally tipped to make method for the early levels of quantum computing, which may take excessive efficiency computing to the subsequent degree.

    Subscribe to Mint Newsletters

    * Enter a sound e mail

    * Thank you for subscribing to our publication.

  • Want to deliver HPC for the lots: Vivek Mahajan

    Fujitsu is credited for growing Fugaku, the world’s strongest supercomputer. Earlier this month, the Japanese IT companies firm started providing high-performance computing as a service (HPCaaS), which is able to enable corporations to make the most of the 442 petaflops of computing energy that Fugaku supplies. Petaflop is a measure of a supercomputer’s efficiency, and what number of calculations it could actually carry out per second. In easy phrases, a human being performing one calculation per second for 31,688,765 years would match what a pc system working at 1PFLOP can do in a second. In an interview, Fujitsu’s chief expertise officer Vivek Mahajan mentioned HPC as a service has a giant future in India, as do much more highly effective future applied sciences reminiscent of quantum computing. Edited excerpts:

    What is the state of superior computing in India at this time?

    There’s a whole lot of want for computing in India, however we consider that in future, HPCs are what each finish person, such as you or me, shall be utilizing. They have that want at this time, however they will’t use it since you or I, nor many corporations, should purchase a supercomputer.

    The approach we see that taking place is by computing as a service. HPC shall be provided by means of the cloud, and also you as a person ought to solely should care about what your drawback is, not whether or not it wants quantum or non-quantum, and so forth.

    Do you already see Indian companies turning to HPCs as a service? What will or not it’s used for?

    Well, it’s (Fugaku) not accessible in India proper now. But I see a whole lot of utilization for this. For occasion, (it’s helpful) if you wish to put your cash in retail, shares, and so forth., and also you need to mannequin that to seek out the perfect worth.

    It’s a quite simple drawback, and you’ll resolve to run two cycles (supercomputers are priced primarily based on time spent to run a simulation on them) in a supercomputer as an finish client.

    We simulated covid on Fugaku to calculate the correct angle for sporting a masks to minimise the affect of the virus and the space folks needs to be at from one another.

    We have additionally used it for tsunami predictions, which saves lives. We can do the identical in India for cyclones or floods. In addition, visitors, logistics and so forth. are actual issues at this time.

    But there are corporations in India coping with such issues with out HPCs; why would we want HPCs?

    I wouldn’t say that they’re fixing it (in the perfect method). How troublesome is it to unravel the covid drawback? I don’t assume that modelling is straightforward to do on regular computing platforms.

    They’re in all probability utilizing on-premise HPCs for climate forecasting, however (with HPC as a service) every state can do their very own flood predictions, and so forth., and so they don’t must have an HPC.

    Subscribe to Mint Newsletters

    * Enter a legitimate electronic mail

    * Thank you for subscribing to our e-newsletter.