Nations Need Supercomputing Leadership

Supercomputing is a reflection of a nation’s technological prowess. A country’s supercomputing capacity directly correlates with its self-reliance in critical areas such as defense, health, agriculture, weather forecasting, materials science, and more. Additionally, supercomputing contributes to national GDP, scientific publications, and innovation.

The India’s journey towards supercomputing leadership is a remarkable endeavor, which began in 1987, and as a young professor at the time, I was fortunate to be a part of it.

The Beginning

India’s journey began in the late 1980s when we faced significant challenges in acquiring supercomputers. In 1984, during Mrs. Indira Gandhi’s tenure, the Indian Institute of Science (IISc) expressed a collective desire to acquire a supercomputer. Although she approved the request, international restrictions and concerns about technology transfer led to denial.

One striking anecdote from that time involved an American engineer inspecting our facility. He arrived with a gun to test whether a bullet could penetrate the glass-enclosed area intended for the computer. Their concern was that foreign agents might steal the technology. Ironically, when the Cyber 1992 supercomputer was finally delivered, it arrived on a large transport aircraft, requiring 15 days to unload—hardly something anyone could carry off in their pocket! This experience underscored the need for self-reliance. The Government of India initiated an indigenous supercomputing program, with several institutions like C-DAC, BARC, NAL, and DRDO stepping up. Among these, C-DAC became the cornerstone for national supercomputing needs, developing India’s first series of indigenous supercomputers under the PARAM brand.

Growth of PARAM Systems

C-DAC’s PARAM series has come a long way:

PARAM 8000 (1991): With a performance of 1 GigaFLOP (less powerful than today’s PCs) and costing nearly ₹1 crore.
PARAM 10000: Achieving teraflop-level performance.
PARAM Yuva (2008): 38.1 TeraFLOPs.
PARAM Yuva II (2013): 360 TeraFLOPs.

Each iteration marked significant advancements, culminating in the National Supercomputing Mission (NSM), launched in 2015.
NSM was a joint initiative between the Department of Science and Technology (DST) and the Ministry of Electronics and Information Technology (MeitY), with IISc and C-DAC as implementing agencies. With an initial funding of ₹4,500 crores, NSM aimed to establish a robust supercomputing ecosystem in India, focusing on:

Indigenous technology development
Training manpower in high-performance computing (HPC)
Deploying cutting-edge supercomputers across academic and research institutions.
So far, NSM has delivered 24 petaflops of computational capacity, distributed across institutions like IISc, IITs, and NITs. Over 10,000 HPC users have been trained, more than 1,500 research publications have emerged, and critical national projects like flood forecasting and drug discovery have benefited from supercomputing resources.

Under NSM, India has achieved several milestones:
PARAM Siddhi-AI (C-DAC, Pune): Ranked among the world’s top 500 supercomputers with a performance of 5.26 petaflops.
Rudra Servers: Entirely designed in India, except for processors.
Trinetra Network: An indigenous interconnect technology.
HPC Software Stack and Storage Solutions: Developed by Indian teams.

The latest liquid cooling technology, asynchronous object storage, and advancements in optical interconnects further showcase India’s progress.

The Path Ahead: NSM 2.0
NSM 2.0, set to launch in December 2025, aims to achieve exascale computing—1,000 times the power of petaflops. Key initiatives include developing indigenous processors based on RISC-V and ARM architectures. Next, building AI-accelerated and GPU-based systems for specific applications and fostering grand challenge applications in disaster management, energy security, life sciences, and more.

The Role of IISc and Academic Research
IISc has been a pivotal player in HPC, hosting India’s largest academic supercomputer with 3.3 petaflops and running at 100% utilization. Over 104 HPC projects across academic institutions are ongoing, spanning basic scientific research and national applications.

In the end…