Computers perform an extremely wide range of tasks by following generalized sets of operations, called programs. They are used as control systems for various consumers and industrial devices, including smartphones and factory devices such as industrial robots.
The first mechanical computer was invented by Charles Babbage in the early 19th century. He also proposed the concept of a programmable computer, though fundamentals of the modern computer were presented by an English mathematician Alan Turing in 1936.
At present, computers can be classified in several ways. Below we have listed different types of digital computers based on their purpose and function.
Example: Smartwatches, industrial embedded systems, video game consoles
Microcomputers (or personal computers) are relatively inexpensive devices that include a microprocessor, memory, and input/output circuitry embedded on a single printed circuit board.
They became popular in the late 20th century with the arrival of increasingly powerful microprocessors. Their predecessors (minicomputers and mainframes) were comparatively more expensive and bulky.
The modern versions of microcomputers include desktop computers, game consoles, smartphones, tablets, laptops, programmable calculators, and in-vehicle computers (built into automobiles for navigation and entertainment).
Note: Embedded computers also fall under this category. They are designed for a particular task, and thus they execute a program stored in non-volatile memory. A washing machine, for example, is an embedded computer that contains only one microcontroller.
Example: PDP-8, LINC
Minicomputers were mid-size computers developed in the 1960s by IBM and its direct competitors. With the advent of single-chip CPU microprocessors, this type of computer came to mean a machine that lies between the microcomputers and the smallest mainframe computers.
They were capable of executing programs in a higher-level language, including BASIC and Fortran. Minicomputers had their own architectures, operating systems, and software. They were designed for instrumentation, communication switching.
By the end of the 1980s, minicomputers were replaced by microcomputers, file servers, and workstations. This happened due to the emergence of inexpensive microprocessor-based hardware and easily deployable local area network systems.
3. Mainframe Computers
The interior of IBM z13 mainframe
Example: IBM zEnterprise System, Hitachi zSeries
Mainframes are larger computers with more processing power than minis and microcomputers. They are mostly used by large organizations for bulk data processing, transaction processing, enterprise resource planning, and other critical applications.
These machines can handle and process extensive amounts of data in less time. Today’s mainframe computers, however, are characterized less by operating speed and more by the following factors:
- Reliable and secure internal engineering
- Hot-swapping of memory and processors
- Backward compatibility with older programs
- Extensive input/output facilities
- High computational utilization rates for impressive throughout
These computers are often used as servers. A server is a system that provides functionality for other devices or programs, known as ‘clients’. This client-server architecture offers a wide range of functionalities (services), for instance, carrying out computation for a client or sharing resources or data among multiple clients.
The most common types of servers are file servers, database servers, web and application servers, mail servers, file servers, and game servers.
All servers include hardware redundancy, such as RAID disk systems, dual power supplies, ECC memory, and extensive pre-boot memory testing and verification. Mission-critical servers use specialized hardware with low failure rates to minimize downtime, and are extremely fault tolerant.
Components of the world’s fastest supercomputer Summit | Image credit: Carlos Jones
Example: IBM Submit, Sunway TaihuLight, Tianhe-2
Supercomputers feature a high-level of performance compared to any other class of computers. Their performance is generally measured in floating-point operations per second (FLOPS).
They are used for computationally intensive tasks in a wide range of fields, including cryptanalysis, physical simulations, molecular modeling, weather forecasting, climate research, and quantum mechanics.
In terms of computational speed, memory size, bandwidth, latency, and input/output technology, supercomputers are the most powerful and very expensive (costs tens of millions) machines. It wouldn’t be cost-effective to use them for mediocre tasks, such as batch or transaction processing.
As of 2019, the fastest supercomputer in the world, named Summit, can reach a peak speed of 200 petaflops. Moreover, it is the first supercomputer to achieve a peak throughput of 1.88 exaops during a genomic analysis.
All of the fastest supercomputers in the world run Linux-based OS. As of 2018, the top 10 supercomputers combined exceeded the one exaFLOPS (1018) mark.
Several countries, including the United States, China, Japan, Taiwan, and the European Union, are currently conducting research to develop even faster and technologically superior supercomputers (that can perform quintillion calculations per second).
Difference Between Mainframes and Supercomputers
- Mainframes focus on batch or transaction processing, while supercomputers are used for complex scientific and engineering problems.
- The performance of mainframes is often measured in millions of instruction per second (MIPS), whereas supercomputer performance is measured in floating-point operations per second (FLOPS) or in traversed edges per second (TEPS).
Inside an HP Z820 workstation
Example: HP 9000, Sony NEWS, Sun Ultra 20
Workstations are specifically designed for scientific and technical applications. These machines are intended to serve one user at a time, and are usually connected to a local area network.
The term ‘workstation’ often refers to everything from a Personal computer to a mainframe terminal connected to a network.
Compared to personal computers, they contain more powerful CPU, graphics, memory, and multitasking capability. Some machines are optimized for manipulating multiple types of data, including complex engineering simulation, animation and rendering of high-quality pictures, and three-dimensional mechanical design.
Today’s workstations use x86-64 microprocessors and run different operating systems including Apple macOS, Microsoft Windows, Linux distributions, Oracle Solaris, and FreeBSD.
They are equipped with ECC memory, multiple processor sockets, powerful processors and graphics cards, multiple displays, and operating systems with advanced features that are not available on personal computers.