India Set To Triple Speed Of Its Fastest Supercomputers
Context:
The National Centre for Medium-Range Weather Forecast (NCMRWF), Noida, and the Indian Institute of Tropical Meteorology (IITM), Pune, respectively, set up the supercomputers Prathyush and Mihir.
Prathyush and Mihir
- Supercomputers Prathyush and Mihir are situated in the National Centre for Medium-Range Weather Forecast (NCMRWF) in Noida and the Indian Institute of Tropical Meteorology (IITM) in Pune, respectively.
- The fastest supercomputers in India as of January 2018 were Prathyush and Mihir, with a maximum speed of 6.8 PetaFlops (floating-point operations per second).
- Mihir was located at NCMRWF in Noida and had a capacity of 2.8 PetaFlops, whereas Prathyush had a capacity of 4.0 PetaFlops and was placed at IITM in Pune. They produced a total of 6.8 PetaFlops when combined.
- Prathyush and Mihir’s establishment cost a total of INR 438.9 Crore.
- On January 8, 2018, Dr Harsh Vardhan, the Union Minister for Science and Technology, inaugurated the systems.
- The supercomputing system at IITM in Pune is called Pratyush, which is a Hindi word that means “rising sun.”
- These “supercomputers” fall under the category of High-Performance Computing (HPC) facilities and are made up of many computers that cooperate to produce significant processing power.
- Prathyush and Mihir’s main areas of interest are climate monitoring and weather forecasting in India.
- India hopes to improve its forecasting powers by using these supercomputers for a variety of things, such as monsoons, fishing, air quality, and extreme events like tsunamis, cyclones, earthquakes, lightning, as well as natural calamities like floods and droughts.
- With the development of Prathyush and Mihir, India joins Japan, the US, and the UK as the fourth nation in the world to have an HPC facility specifically designed for weather and climate research.
Points to Ponder:
- Scale-up Strategy: By deploying an 18-petaflop system this year, India hopes to considerably boost its supercomputing power. A processing speed of 18,000 trillion floating point operations per second would be made possible.
- Current Supercomputers: Pratyush and Mihir, the two most potent civilian supercomputers in India, now have a combined performance of 6.8 petaflops. Mihir resides at the National Centre for Medium-Range Weather Forecasting (NCMRWF) in Noida, whereas Pratyush is housed at the Indian Institute of Tropical Meteorology in Pune. Both are associated with the Ministry of Earth Sciences (MoES) and became operational in 2018.
- Imported from ATOS: The new supercomputers, which have not yet been given a name, have been imported from the French company ATOS. Information technology consulting and service provider ATOS. The Indian government and France agreed to a contract in December 2018 to buy 4,500 crore worth of high-performance computers by 2025.
- Cost: It is projected that the Ministry of Earth Sciences (MoES) new supercomputers will run about 900 crores.
- Enhanced Resolution: The improved supercomputers will improve the mapping of the weather and climate in terms of resolution. The present methods enable mapping with a 12×12 km resolution. The resolution can be increased to 6×6 kilometres with the new technology, offering better clarity and more precise local forecasts.
- Future Goal: The ultimate aim for India is to portray territories using grids of one km squares. Achieving this resolution would make it possible to spot and warn about quickly changing weather systems like cloudbursts.
- Global Comparison: worldwide comparison With a peak speed of one exaflop (about 1,000 petaflops), the Frontier-Cray system at Oakridge National Laboratory in the United States is now the fastest high-performance computing system in the world. Despite a huge increase in supercomputing power, India’s new 18-petaflop machine will still fall short of the world’s most advanced computers.
- Top Systems: Based on speed, the top ten high-performance computing systems in the world range from about 400 petaflops to 60 petaflops. These systems are employed for a variety of computer activities, including simulations, scientific research, and weather forecasting.
