UPSC Prelims Practice Questions — India’s Rising Supercomputing Ecosystem

Q1. With reference to how the computing performance of supercomputers is expressed, consider the following statements: Which of the statements given above is/are correct?

  1. A capacity quoted in 'AI petaflops' at mixed precision can be added directly to standard double-precision petaflops to work out a system's position on a global ranking.
  2. A system rated at one exaflop performs about a thousand times as many floating-point operations per second as a system rated at one petaflop.
  3. Lower-precision arithmetic uses fewer bits for each number, so the same hardware can report a much higher FLOPS figure at mixed precision than at double precision.
  • A. 2 and 3 only
  • B. 1 and 3 only
  • C. 1, 2 and 3
  • D. 1 and 2 only

Q2. Consider the following facts: (i) As of September 2026, the 40 systems deployed under the National Supercomputing Mission have a combined capacity of about 68 petaflops. (ii) The world's fastest single supercomputer runs at about 2.19 exaflops. (iii) AIRAWAT is rated at 410 AI-petaflops in mixed precision and 8.5 petaflops sustained in double precision. Which of the following inferences can validly be drawn from the above facts?

  1. On these figures, the combined capacity of all NSM systems is less than one-thirtieth of the capacity of the world's fastest single supercomputer.
  2. On a ranking based on sustained double-precision performance, AIRAWAT would be placed above a system that sustains 100 petaflops in double precision.
  3. If ten more systems of the present average NSM size are added, the combined NSM capacity will cross one exaflop.
  • A. 1 and 2 only
  • B. 1 only
  • C. 2 and 3 only
  • D. 1, 2 and 3

Q3. With reference to the architecture of a supercomputer, consider the following statements: Which of the statements given above is/are correct?

  1. A supercomputer gets its speed mainly by splitting one large problem among many processors or nodes that compute at the same time.
  2. Because nodes exchange intermediate results during a computation, the bandwidth and latency of the network linking them can limit the speed of the whole machine.
  3. The interconnect is the component that performs the floating-point calculations, while the compute servers only route data between nodes.
  • A. 1 and 3 only
  • B. 1, 2 and 3
  • C. 2 and 3 only
  • D. 1 and 2 only

Q4. Consider the following statements: Which of the following is correct in respect of the above statements?

  1. The increase in the Ministry of Earth Sciences' HPC capacity to about 22 petaflops shortened weather-model run times enough to make high-resolution forecasts usable in real time.
  2. Greater computing power makes the atmosphere inherently more predictable at long lead times because it reduces the chaotic behaviour of atmospheric processes.
  • A. Both Statement-I and Statement-II are correct and Statement-II explains Statement-I
  • B. Both Statement-I and Statement-II are correct, but Statement-II does not explain Statement-I
  • C. Statement-I is correct, but Statement-II is incorrect
  • D. Statement-I is incorrect, but Statement-II is correct

Q5. A weather service halves the horizontal grid spacing of its numerical weather prediction model and keeps its computing capacity unchanged. Consider the following likely consequences: Which of the statements given above is/are correct?

  1. Each forecast run of the model would take longer to complete.
  2. Other things being equal, the model could represent smaller-scale weather features, such as intense local convective storms, more realistically.
  3. Fewer ensemble members could be run within the same operational time window.
  • A. 1 and 3 only
  • B. 2 and 3 only
  • C. 1, 2 and 3
  • D. 1 and 2 only

Q6. With reference to the 'build approach' followed under India's National Supercomputing Mission, consider the following statements: Which of the statements given above is/are NOT correct?

  1. The approach began with assembling supercomputers in India from subsystems that were largely sourced from abroad.
  2. Indigenous design of key components such as compute servers was the starting point of the approach, and assembly in India was added in later phases.
  3. The approach raises domestic value addition step by step across phases instead of seeking full indigenisation at the outset.
  • A. 2 and 3 only
  • B. 2 only
  • C. 1 and 3 only
  • D. 1 only

Q7. Consider the following assertion: The PARAM Rudra supercomputers represent a deeper level of technological self-reliance than systems assembled in India from imported subsystems. Which of the following statements support/supports the above assertion?

  1. They use HPC servers that were designed and manufactured in India.
  2. They run on a system software stack that was developed in India.
  3. They have been commissioned at institutions in more than one Indian city.
  4. The high-speed interconnect deployed in the PARAM Rudra system at C-DAC Pune was developed in India.
  • A. 1 and 2 only
  • B. 2, 3 and 4 only
  • C. 1, 2, 3 and 4
  • D. 1, 2 and 4 only

Q8. Consider the following statements: Which of the following is correct in respect of the above statements?

  1. AIRAWAT's position in the TOP500 list announced at ISC 2023 was determined by its peak capacity of about 410 AI-petaflops in mixed precision.
  2. The TOP500 list ranks supercomputers by their measured performance on a benchmark that solves linear-algebra equations, expressed in floating-point operations per second.
  • A. Both Statement-I and Statement-II are correct and Statement-II explains Statement-I
  • B. Both Statement-I and Statement-II are correct, but Statement-II does not explain Statement-I
  • C. Statement-I is correct, but Statement-II is incorrect
  • D. Statement-I is incorrect, but Statement-II is correct