In 2020, one machine gave India a seat among the world’s 100 fastest computers. Four years later, it’s still getting upgraded.
SciTech0131
PARAM Siddhi-AI
India’s fastest AI-focused supercomputer.
210
AI Petaflops, PARAM Siddhi-AI’s peak AI computing power
2020
Commissioned by C-DAC and Atos under the National Supercomputing Mission
63rd
Global rank on the November 2020 TOP500 list
42
NVIDIA DGX A100 systems powering the machine
2024
C-DAC announces an 8-GPU upgrade to NVIDIA H100 chips
Still current, four years on: Most launch-day rankings age quickly. PARAM Siddhi’s own operators kept investing in it, upgrading its GPUs years after its debut rather than retiring it.
Science and Technologymcqquestion.com
Must Know
- PARAM Siddhi-AI is India’s fastest AI-focused supercomputer.
- It was developed by C-DAC, in collaboration with Atos, under the National Supercomputing Mission.
- It was commissioned in November 2020.
- Its architecture uses 42 NVIDIA DGX A100 systems, connected through NVIDIA Mellanox HDR InfiniBand networking.
- It reached a peak AI computing capacity of 210 AI Petaflops.
Good to Know
- PARAM Siddhi ranked 63rd on the TOP500 list of the world’s most powerful non-distributed computer systems, released in November 2020.
- Its Rpeak was 5.267 Petaflops, with a sustained Rmax of 4.6 Petaflops.
- It supports application development across computational chemistry, astrophysics, and drug design. It also powers flood forecasting for flood-prone cities.
- The National Supercomputing Mission is jointly run by the Department of Science and Technology and the Ministry of Electronics and Information Technology.
- In September 2024, C-DAC floated a tender to add 8 NVIDIA H100 GPUs to the system, a major planned upgrade.
Test Yourself
Great to Know
- Don’t confuse PARAM Siddhi with India’s original PARAM programme. PARAM 8000, built by C-DAC in 1991 under Vijay Bhatkar, was India’s first supercomputer. It’s a completely different, much earlier machine.
- The H100 upgrade matters for a real reason. H100 GPUs train and run AI models dramatically faster than the older A100 chips. That gap matters most for large language models.
- PARAM Siddhi is officially part of a combined system called PARAM Siddhi-AI (PSAI). That name reflects its specific AI-computing focus, distinct from India’s other general-purpose HPC supercomputers.
Beyond the answer
Browse all science-and-technology topics →
📚 Keep reading
SciTech0094 — Vijay Bhatkar: The Engineer Behind India's First SupercomputerIndSoc0090 — India's Higher Education Sector: A Decade of GrowthEthics0080 — Anna HazareLord Dufferin: The Founding of the Indian National CongressSciTech0136 — Large Language Models (LLMs)❓ Practice this topic
What does India's National Viral Hepatitis Control Programme aim to achieve by 2030?According to WHO, how does viral hepatitis's global death toll compare to HIV/AIDS?What can now cure most Hepatitis C infections, even without a vaccine?Since when has a Hepatitis B vaccine existed?🎲 Take a Science And Technology Quiz
Leave a Reply