Recently, JUPITER in Germany became Europe’s first exascale supercomputer.

About Supercomputers

• A supercomputer is a highly advanced computing system designed to perform extremely large-scale, complex, and calculation-intensive tasks that ordinary computers cannot handle efficiently. 
• Core Principle- Parallel Computing: Instead of relying on a single processor, supercomputers use thousands to millions of processors (CPUs and GPUs) working simultaneously. 
  • This parallel computing architecture allows smaller portions of a problem to be solved independently, with results aggregated for faster and more efficient computation.
• Structure and Components:
  • Processors: CPUs handle general tasks, while GPUs efficiently process repetitive mathematical operations used in scientific simulations.
  • Nodes: Groups of processors with memory act as small computing units, with supercomputers containing thousands of nodes.
  • High-Speed Network: Nodes are interconnected via ultra-fast data links, enabling seamless data exchange.
  • Memory and Storage: Each node has local memory, and large-scale storage systems handle petabytes of data, ensuring orderly read/write operations across thousands of nodes.
  • Cooling Systems: High computation generates enormous heat, necessitating liquid immersion, water pipes, or refrigeration units.
  • Power Supply: Supercomputers consume megawatts of electricity, equivalent to a small town, requiring efficient power distribution.
• Software and Task Management:
  • Parallel Programming: Languages like Message Passing Interface (MPI) and OpenMP schedule tasks across processors.
  • Load Balancing: Algorithms ensure even distribution of computational load, avoiding idle processors.
  • Job Scheduling: Users submit job scripts remotely, specifying computation requirements. A scheduler assigns tasks to nodes, optimizing efficiency.
• Performance Metrics: Measured in flops (floating-point operations per second). 
  • Top Supercomputers: Exaflops, performing quintillions of operations per second, exceeding what all humans could calculate in a lifetime using a calculator.
  • Flops is a measure of a computer’s computational performance, indicating how many floating-point calculations it can perform in one second.
• User Interaction:
• • Remote Access: Supercomputers are accessed via secure networks, often using terminal interfaces.
  • Data Output: Computation results are stored in the supercomputer’s file system, later downloaded for analysis and visualization.

Difference Between Supercomputers and Normal Computers

India’s Journey in Supercomputing

• Origins: Initiated in the late 1980s after Western countries denied high-end exports, leading to the creation of the Centre for Development of Advanced Computing (C-DAC) in 1988 for indigenous development.
• PARAM Series: PARAM 8000 (1991) was India’s first homegrown supercomputer, pioneering parallel computing architectures.
• National Supercomputing Mission (NSM): Launched in 2015 by the Department of Science and Technology (DST) and the Ministry of Electronics and Information Technology (MeitY), aiming to establish 70+ High-Performance Computing (HPC) facilities with teraflops to petaflops capacities.
• Key Installations: 
  • AI Research Analytics and Knowledge Dissemination Platform – Powering Science and Artificial Intelligence (AIRAWAT-PSAI) at C-DAC Pune.
    • India’s AI supercomputer ‘AIRAWAT’ has been ranked among the world’s top 100 most powerful supercomputers.
  • Pratyush at the Indian Institute of Tropical Meteorology (IITM), Pune, and 
  • Mihir at the National Centre for Medium Range Weather Forecasting (NCMRWF), Noida — used for weather, climate, ocean, and scientific simulations.
• Technological Collaboration: Partnerships with the Indian Institute of Science (IISc), Indian Institutes of Technology (IITs), and private vendors develop Rudra servers and AUM HPC nodes for Artificial Intelligence (AI), defence, and scientific research.
  • AUM refers to the specific indigenous HPC architecture/name under India’s National Supercomputing Mission.

Significance of Supercomputers

• Scientific Advancement: Enable simulations of cosmology, molecular dynamics, and climate systems.
• Defence & Security: Critical for war-game simulations, nuclear research, and strategic planning.
• Technological Leadership: Showcase national innovation and technological prowess, influencing global competitiveness.

• Disaster Management: Aid in monsoon prediction, flood modelling, and earthquake simulations, saving lives.

Need for Supercomputers

• Data-Intensive Research: Modern science generates massive datasets requiring high-performance computation.
• AI and Machine Learning: Training large neural networks is only feasible on HPC platforms.
• Global Competitiveness: Nations with supercomputing capabilities dominate strategic, economic, and scientific domains.
• Climate and Weather Forecasting: Accurate predictions require high-speed simulations and real-time analysis.

Global Initiatives and World Best Practices

• Exascale Computing: Machines like JUPITER (Germany) achieve >1 exaflop using renewable energy.Neuromorphic Computing: Brain-inspired architectures integrating processing and memory on a single chip, improving speed and energy efficiency.
• Quantum Computing Integration: Certain simulations may shift to quantum systems, reducing hardware and energy requirements.
• International Collaborations: Partnerships like PRACE (EU) and INCITE (USA) enable global access and scientific collaboration.

DOWNLOAD PDF