What Is A Computer


  • Definition:
    • electronic, digital, general purpose computing machine that automatically follows a step-by-step list of instructions for solving a problem.
    • This step-by-step list of instructions that a computer follows is also called a computer program.
  • Computer Part: case, motherboard, CPU, RAM, power supply, hard drive, graphic cards.

Turing Machine

  • In 1936, British mathematician Alan Turing developed a hypothetical device, the Turing machine, which is the abstract model of all computers
  • A Turing machine consists of

    • a tape divided into cells
    • a moving read/write head
    • a state register storing the state of the Turing machine
    • a finite table of instruction specifying what the machine does when reading the content of the current cell:
      • move right/left; erase/write a symbol; change the state
  • All things that can be computed can be computed by a Turing machine

Universal Turing Machine

  • Turing described a Turing machine that could simulate all other Turing machines.
    • inputs: data + a description of computation (Turing machine)
  • A computer is a Universal Turing Machine!

Historical Development

  • Generation Zero: Mechanical Calculating Machines (1642-1945) 机械计算机
  • The First Generation: Vacuum Tube Computers (1945-1953) 真空管计算机
  • The Second Generation: Transistor Computers (1954-1965) 晶体管计算机
  • The Third Generation: Integrated Circuit (IC) Computers (1965-1980) 集成电路计算机
  • The Fourth Generation: VLSI Computers (1980-) (very large scale integration computer)

The von Neumann Architecture

  • stored-program architecture
  • Both data and program are stored in the memory

  • A Central Processing Unit (CPU)
    • Control unit 控制单元
    • Arithmetic Logic Unit (ALU) 计算逻辑单元
    • Registers
    • IR: Current Instruction
    • PC: store the address of next instruction
  • Main memory
  • I/O- system

  • a single path between the main memory and CPU, called the von Neumann bottleneck The von Neumann Architecture

Von Neumann Execution Cycle

  • Fetch-decode-execute cycle
    1. the control unit fetch the next instruction from the memory 控制单元取指令
    2. the instruction is decoded into a language that the ALU understands 解码
    3. data operands are fetched from the memory into the registers inside CPU 从内存获取数据操作数到寄存器
    4. the ALU executes the instruction and places the result into the registers or memory 计算逻辑单元执行指令并把结果放到内存或寄存器

The von Neumann Bottleneck

  • CPU and memory are separate
  • All data and code are in the memory
  • CPU is usually faster than memory
  • CPU is forced to wait for needed data to be transferred to or from memory

The system bus model of the von Neumann Architecture


Levels of Abstraction

  • User level: applications such as qq.exe app
  • High level language: C, Java, C++ program language
  • Assembly language
  • Operating system
  • Machine Language: Instruction Set A
  • Control level: micro-code or hardwired
  • Digital logic: circuits, gates

Levels of Transformations

Levels-of-Transformations Levels-of-Transformations

  • Problem -> Algorithm -> Program -> Instr Set Architecture -> Microarch -> Circuits -> Device

The Machine Levels

  • Instruction Set Architecture (ISA): instructions that a CPU can execute
  • Microarchitecture: implementation of ISA
  • Circuits: Details of electrical circuits
  • Devices (transistors): Circuits are built by interconnecting transistors
  • Bits: Transistors operate on bits (“0” or “1”) that represent data and information

Hardware VS Software

Layers-of-Transformations Hardware&Software

  • Whatever can be done by hardware can also be done by software, and vice versa
  • Hardware implementations are faster but fixed
  • Software implementations are more flexible but slower