The National Supercomputing Mission (NSM) had deployed 40 supercomputers with a combined 68 petaFLOPS (PF) as of September 2026, against a target of 50 systems and over 123 PF.
NSM was launched in April 2015 with about Rs 4,500 crore; DST and MeitY steer it jointly, and C-DAC Pune and IISc Bengaluru implement it.
6,000 indigenous Rudra servers run inside PARAM Rudra supercomputers, with 1,500 more under manufacture.
NSM systems have served over 16,000 researchers in 400+ institutions and run more than 1.5 crore compute jobs.
ESTIC-2026, organised by CSIR, will be held on 27-29 October 2026 at Bharat Mandapam, New Delhi.
Self-reliance in supercomputing: design, manufacture, system software and applications for research and societal use
Key: Cabinet approval March 2015 for seven years; supercomputers placed in academic and research institutions; also trains users through the HPC Shiksha Portal, a SWAYAM-NPTEL course and AICTE faculty programmes
High-speed network joining universities, laboratories and research institutions
Key: Approved March 2010 by the Cabinet Committee on Infrastructure at Rs 5,990 crore; implemented by the National Informatics Centre; links over 1,800 institutions
Systems an examiner is likely to name
| System | Year | Where | Why it matters |
|---|---|---|---|
| PARAM 8000 | 1991 | C-DAC, Pune | India's first indigenous supercomputer; 1 gigaflop; led by Vijay Bhatkar |
| PARAM Yuva | 2008 | C-DAC, Pune | 54 teraflops peak; ranked 69th in the world in November 2008 |
| PARAM Shavak | 2014 | Colleges and universities | 'Supercomputing-in-a-box': a desktop HPC and AI system of about 2 teraflops |
| Pratyush and Mihir | 2018 | IITM Pune and NCMRWF Noida | Weather and climate systems; 6.8 PF together |
| PARAM Shivay | 2019 | IIT (BHU), Varanasi | First supercomputer built under NSM; 833 teraflops |
| PARAM Siddhi-AI | 2020 | C-DAC, Pune | HPC-AI system under NSM; ranked 63rd on TOP500 in November 2020 |
| PARAM Pravega | 2022 | IISc, Bengaluru | 3.3 PF, installed under NSM |
| AIRAWAT-PSAI | 2023 | C-DAC, Pune | AI supercomputer under the National Program on AI; 13,170 teraflops peak; ranked 75th on TOP500 in June 2023 |
| PARAM Rudra (three systems) | 2024 | Pune (GMRT), Delhi (IUAC), Kolkata (S.N. Bose Centre) | Built on indigenous Rudra servers; about Rs 130 crore; dedicated on 26 September 2024 |
| Arka and Arunika | 2024 | Weather and climate research | HPC system worth Rs 850 crore, inaugurated with PARAM Rudra |
R&D society of MeitY, set up in 1988; designs the PARAM series, Rudra servers and Trinetra; co-implements NSM
Co-implements NSM with C-DAC; hosts PARAM Pravega
Jointly steer and fund NSM
Implements the National Knowledge Network
Organiser of ESTIC-2026
A supercomputer's speed is counted in floating-point operations per second. TeraFLOPS is a trillion (10^12), petaFLOPS a quadrillion (10^15), exaFLOPS a quintillion (10^18). The speed comes from many processors, or nodes, working on one problem together over fast interconnects.
Simple Analogy: One cook versus a thousand cooks sharing one recipe, with waiters running between them.
GS Paper 3 > Science and Technology > Awareness in IT and Computers; Indigenisation of Technology
General Awareness > Computers and Science & Technology
Computer Awareness > Types of Computers
One quadrillion (10^15) floating-point operations per second
C-DAC's indigenously designed HPC server, made by Indian EMS partners and used in PARAM Rudra
C-DAC's indigenous high-speed interconnect between compute nodes (100 and 200 Gbps)
A table-top 'supercomputing-in-a-box' for colleges and universities