Commercial computer system development has followed development of hardware technology and is usually divided into four generations:
1. First generation (1945–1955)—vacuum tube technology.
2. Second generation (1955–1965)—transistor technology.
3. Third generation (1965–1980)—integrated circuit (IC) technology.
4. Fourth generation (1980–??)—Very Large Scale Integrated (VLSI) circuit technology.
We will not elaborate on the architectural details of the various machines developed during the three generations, except for the following brief evolution account. First-generation machines such as UNIVAC 1 and IBM 701, built out of vacuum tubes, were slow and bulky and accommodated a limited number of I=O devices. Magnetic tape was the predominant I=O medium. Data access time was measured in milliseconds. Second-generation machines (IBM 1401, 7090; RCA 501; CDC 6600; Burroughs 5500; DEC PDP-1) used random-access core memories, transistor technology, multifunctional units, and multiple processing units. Data access time was measured in microseconds. Assembler and high-level language were developed. The integrated-circuit technology used in third-generation machines such as the
IBM 360, UNIVAC 1108; ILLIAC-IV, and CDC STAR-100 contributed to nanosecond data access and processing times. Multiprogramming, array, and pipeline processing concepts came into being. Computer systems were viewed as general-purpose data processors until the
introduction in 1965 of DEC PDP-8, a minicomputer. Minicomputers were regarded s dedicated application machines with limited processing capability compared to that of large-scale machines. Since then, several new minicomputers have been introduced and this distinction between the mini and large-scale machines is becoming blurred due to advances in hardware and software technology. The development of microprocessors in the early 1970s allowed a significant contribution to the third class of computer systems: microcomputers. Microprocessors are essentially computers on an integrated-circuit (IC) chip that can be used as components to build a dedicated controller or processing system. Advances in IC technology leading to the current VLSI era have made microprocessors as powerful as minicomputers of the 1970s. VLSI-based systems are called fourth-generation systems since their performance is so much higher than that of third-generation systems. Modern computer system architecture exploits the advances in hardware and software technologies to the fullest extent. Due to advances in IC technology that make the hardware much less expensive, the architectural trend is to interconnect several processors to form a high-throughput system. Some claim that we are now witnessing the development of fifth-generation systems. There is no accepted definition of what a fifth-generation computer is. Fifth-generation development efforts in the United States involve building supercomputers with very high computational capability, large memory capacity, and flexible multiple-processor architectures, employing extensive parallelism Japanese fifth-generation activities aimed toward building artificial intelligence-based machines with very high numeric and symbolic processing capabilities, large memories, and user-friendly natural interfaces. Some attribute fifth generation to biology-inspired (neural networks, DNA) computers and optical computer systems. The current generation of computer systems exploits parallelism in algorithms and computations to provide high performance. The simplest example of parallel architecture is the Harvard architecture, which utilizes two buses operating simultaneously.
Parallel processing architectures utilize a multiplicity of processors and memories operating concurrently.
