Computers are often built and optimised for a particular type of workload. Supercomputers are the most powerful category of computer, and their design priority is raw processing speed for extremely demanding calculations, rather than handling many small, everyday input-output tasks.
Supercomputers are used for mathematically intensive scientific applications such as weather forecasting, climate modelling, nuclear simulations, and complex physics calculations. These tasks require performing enormous numbers of floating-point calculations at extremely high speed, which is exactly what supercomputer architecture (many processors working in parallel) is built to deliver.
Data retrieval, input-output intensive processing, and simple record manipulation are workloads that are typically handled efficiently by mainframe computers or ordinary servers, which are optimised for handling many transactions and large volumes of routine data access rather than for extreme mathematical computation.
Examination reminder: associate supercomputers specifically with heavy scientific and mathematical computation, and mainframes with high-volume transaction and data processing; the two categories are often confused but serve different purposes.
