컴퓨터구조) 01 Computer Abstractions and Technology

Electronic Control Unit (ECU)

• ECU is an embedded system in automotive electronics that controls on or more of the electrical systems in a vehicle

 

What is Computer Architecture

 

Abstraction Analogies (내부적으로 알 필요없다)

Eight Great Ideas in Computer Architecture

• Design for Moore’s raw (integrated circuit resources)
• Use abstraction to simplify design
• Make the common case fast
• Performance via parallelism (병렬화)
• Performance via pipelining
• Performance via prediction
• Hierarchy of memories
• Dependability via redundancy

 

Below Your Program

• The layers of software are organized primarily in a hierarchical fashion
• Application software is written in high-level language
• System software is sitting between the hardware and application software
  • Operating system interfaces between a user’s program and the hardware
    • Handling basic input and output operation
    • Allocating storage and memory
    • Resource sharing and task scheduling
    • ex) Linux, windows, ios
  • Compiler translates of a program written in high-level languages into instructions that the hardware executes
• Hardware includes processor, memory, I/O controller, etc

 

High-Level Language to Language of Hardware

• High-level language is a level of abstraction close to problem domain
  • Providing for productivity and portablility
  • Compilers translate programs into instructions
• Assembly language is a sysbolic representation of machine instructions
  • Aseemblers translate a symbolic version of instructions into the binary version
• Machine language is a hardware representation of machine instructions
  • Represented by binary digits, ex) bits
  • Encoded instructions and data 

 

Organization of a Computer

• The five classic components of a computer are input, output, memory, datapath, and control
  • Processor: datapath + control
  • The processor gets instructions and data from memory
  • Input writes data to memory
  • Output reads data from memory
  • Control sends the signals that determine the operations of the data path, memory, input, and output
• Same components for all kinds of computer
  • ex) desktop, server, and embedded

 

Opening the Box (1)

• I/O of the five components dominate the Apple iPhone XS Max
  • I/O : a capacitive multitouch LCD, front-/rear-facing cameras, microphone, speakers, Wi-Fi and Bluetooth networks, etc
  • The datapath, control, and memory are a tiny portion
• There are many integrated circuits in the board
  • A12 processor contains two large and four little ARM processors (aka CPU) that operate with a clock rate of 2.5 GHz
    • Datapath performs arithmetic(연산) operations
    • Control sequences datapath, memory, and I/O
• iPhone XS max package includes a memory chip with 32Gb
  • Memory is the storage area where programs and data are kept when running
  • The memory is built from dynamic random access memory (DRAM) chips that contain the instructions and data of a program
• Cache memory consists of a small, fast memory that acts as a buffer for the DRAM
  • Cache is built using static random access memory (SRAM), which is faster but less dense, and hence more expensive, than DRAM
  • SRAM and DRAM are two layers of the memory hierarchy
• Abstraction helps us deal with complexity by hiding lower-level details
  • Abstract of interface between the hardware and the lowest-level software is instruction set architecture (ISA), or simply architecture
  • Application binary interface (ABI) is the ISA + OS interface
  • Implementation is the hardware that obeys architecture abstraction

 

A Safe Place for Data

• Volatile(휘발성) memory loses instructions and data when power off (전원을 끌때)
  • Main memory or primary memory
  • Usually small size and fast
  • ex) DRAM and SRAM
• Nonvolatile memory keeps programs and data when power off
  • Secondary memory
  • Larger size and slower than the main memory
  • ex) magnetic hard disk, flash memory, optical disk (CD, DVD)

 

Communicating with other Computers

• Networks interconnect whole computers and networked computers have several major advantages
  • Communication : Information is exchanged between computers at high speeds
  • Resource sharing: Computers on the network can share I/O devices
  • Nonlocal access: Users need NOT be near the computer they are using
• Local area network (LAN) is desinged to carry data within a geographically confined area, typically within a single building: Ethernet
• Wide area network (WAN) is a network extended over hundreds of kilometers that can span a continent: Internet
• Wireless network: Wi-Fi, Bluetooth

 

Chip Manufacturing Process

• The manufacture of a chip begins with silicon, a substance found in sand
  • Silicon is called semiconductor because it does NOT conduct electricity well
  • With a special chemical process, it can be 1) excellent conductors, 2) excellent insulator, 3) areas that can conduct or insulate under specific conditions (switch)

 

Cost of Integrated Circuit

 

Defining Performance

• Performance considerations for computer
  • Processor characteristics (ex) clock frequency, # of cores)
  • System bus speed
  • Cache/ Memory size
  • Disk storage (SSD or HDD)
  • Graphic cards (GPU)
  • Power consumption
  • Price, weight, form factor, etc

 

Response Time

• Individual computer user is interested in reducing response time (반응시간이 짧은지)
  • The time between the start and completion of a task, called execution time
• Datacenter manager cares about increasing throughput (bandwidth) (많은 것을 처리하는지)
  • The total amount of work done in a given time
• How are response time and throughput affected by
  • Replacing the processor with a faster one → both response time and throughput 증가
  • Adding additional processors → only throughput 증가

 

Measuring Performance

• Time is the measure of computer performance
  • Program execution time is measured in seconds per program
• The most straightforward definition of time is called wall clock time, response time, or elapsed time
  • These terms mean the total time to complete a task, including disk/memory accesses, input/ouput (I/O) activities, operating system overheads - everything
• CPU execution time (or CPU time) is the time the CPU spends computing for the task
  • This does NOT include the time consumed by I/O or other programs
  • CPU time can be further divided into user CPU time and system CPU time
• Different applications are sensitive to different aspects of the performance of a computer system
  • To improve the performance of a program, one must have a clear definition of what performance metric matters
  • Then, proceed to find performance bottlenecks by measuring program execution

 

CPU Clocking

• Almost all computers are constructed using a clock that determines when events take place in the hardware
  • These discrete time intervals are called clock cycles (or ticks , clock ticks, clock periods, clocks , cycles)

• Designers refer to the length of a clock period both as the time for a complete clock cycle and as the clock rate (frequency)
  • Clock period: duration of a clock cycle (ex) 250ps)
  • Clock rate (or clock frequency- clock cycle time의 역수) : cycles per sound (ex)4GHz)

 

CPU Performance

• The bottom-line performance measure is CPU execution time

• The hardware designer can improve performance by
  • Reducing the number of clock cycles required for a program
  • Reducing the length of the clock cycle (i.e., increasing the clock rate)
• The designer often faces a trade-off between the number of clock cycles and the clock rate

 

Instruction Performance

• The execution time depends on the number of instructions in a program

• Instruction count for a program is determined by program, ISA, and compiler
• CPI (Clock cycle per instruction) is the average number of clock cycles each instruction take to execute for a program
  • Determined by CPU hardware
  • Different instructions may take different clock cycles → CPI is an average

 

Performance Summary

 

Power Wall

 

Amdahl’s Law

 

MIPS as a Performance Metric

• MIPS (million instructions per second) is given by

• Faster computers have a higher MIPS rating
• CPI varies between programs on a given CPU
• MIPS does NOT account for
  • differences in ISAs between computers
  • differences in complexity between instructions