It’s been two years since the National Program for AI was launched by NITI Aayog, which aims to uplift the research and development in AI for social good. The national strategy for AI, termed #AI4All, is focused on promoting inclusive growth using AI. In this endeavour, several strategic collaborations with research institutions, academia and industry partners have been forged – with the singular goal of developing technology, augmenting R&D and enhancing ethical standards in the use of AI for social good.

Today was a historic step in moving one more step closer to this goal. India’s largest HPC AI supercomputer PARAM SIDDHI AI, commissioned by C-DAC, was ranked #63 in the TOP500 Supercomputing list, making it India’s largest and fastest supercomputer. The TOP500 list is released twice a year, with inputs from high-performance computer experts, computational scientists, manufacturers, and the Internet community.

At the media briefing was Shri. Ajay Prakash Sawhney, Secretary, MeitY, Govt. of India, Prof. Ashutosh Sharma, Secretary, DST, Govt. of India, Dr. Hemant Darbari, Director General, C-DAC, Govt. of India and Vishal Dhupar, MD, NVIDIA South Asia – all of whom touched upon this historic milestone and discussed what makes PARAM SIDDHI AI such a powerful machine.

Ajay Sawhney, Secretary, MeitY remarked that the supercomputer was developed under the aegis of the three-phase National Supercomputing Mission that aims to provide HPC facilities to 75+ institutions, researchers and academicians – all working together through the National Knowledge Network (NKN). “These advanced compute capabilities will augment the efforts to bring largescale AI deployment for social development. PARAM SIDDHI AI will be able to expedite image surveillance, image processing, NLP and computer vision among other AI-based functions, to drive efficacy in healthcare, manufacturing, urban planning and transport projects.

PARAM SIDDHI AI, a machine with 210 AI Petaflops (6.5 Petaflops Peak DP), is based on the NVIDIA DGX SuperPOD reference architecture comprising of 42 NVIDIA DGX A100 systems, connected with NVIDIA Mellanox HDR InfiniBand networking along with indigenously developed HPC-AI engine, software frameworks, cloud platform by C-DAC. Atos is the integration partner, providing the Atos Codex AI Suite to deliver real-time analytics and support AI application development.

Dr Hemant Darbari, Director General of C-DAC added that PARAM SIDDHI AI’s superior compute capabilities, a mature HPC AI framework and AI libraries in SDK will jumpstart research efforts and support startups to execute largescale projects. “We have seen first-hand the efficiency and speed of supercomputing, and the monumental impact it can have on sectors like healthcare. When COVID19 began earlier this year, drug repurposing and related simulations through AI led researchers to discover the efficacy of Remdesivir to treat patients, and since then have been used to save countless lives. We believe PARAM SIDDHI AI will be capable of such largescale advancements in areas like healthcare, agritech, cybersecurity, robotics and more.”

Vishal Dhupar, MD of NVIDIA South Asia spoke of the future of a tech-first world, and how supercomputers are the answer to augmenting the efforts of software engineers. “With super compute capabilities, we can teach computers to write software. This is how we can make strides in largescale automation to address problems at scale. With PARAM SIDDHI AI being ranked in the top 100 supercomputers globally, India is marking its dominance as an AI garage of the world.”

Understandably, with the debate around ethical use of AI and the need for data integrity, PARAM SIDDHI AI has been developed with a three-layer data security framework. India, which recently joined the Global Partnership on AI (GPAI), is working on multiple policies to make data privacy a core tenet in hardware infrastructure as well. Dhupar added that in addition to policy and infrastructure fortifications, technology amendments are also in place that ensure data is encrypted and secure. “For this to happen, edge computing in AI is key. In areas like healthcare where patient data is sensitive and localised, edge AI will help in keeping data confidential and within the control of relevant players like doctors and patients.”

PARAM SIDDHI AI is currently at C-DAC’s facility in Pune.