Tutorial EXL1 Lecture

  • Upload
    satyam6

  • View
    220

  • Download
    0

Embed Size (px)

Citation preview

  • 8/10/2019 Tutorial EXL1 Lecture

    1/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top

    OVERVIEW

    The system unit of a computer is sometimes thought of as a mysterious "black box" that makes the computerwork, and often the user doesn't have much understanding of what happens inside it. In this chapter, wedemystify the system unit by looking inside the box and closely examining the functions of the partsinside. In doing so, the chapter gives you a feel for how the CPU, memory, and other devices commonly foundwithin the system unit work together to process data into meaningful information.

    To start, we discuss how a computer system represents data and program instructions. Here we talk aboutthe codes that computers use to translate data back and forth from symbols that the computer can manipulateto symbols that people are accustomed to using. These topics lead into a discussion of how the CPU andmemory are arranged with other processing and storage components inside the system unit, and then how aCPU is organized and how it interacts with memory to carry out processing tasks. Finally, we look atstrategies that are used today to speed up a computer, plus some strategies that may be used to createfaster computers in the future.

    While most of you reading this chapter will apply its contents to conventional personal computer systems --such as desktop and notebook PCs -- keep in mind that the principles and procedures discussed in thischapter cover a broad range of computer products. These products include microprocessors embedded intoys, household appliances and other devices, as well as powerful servers, mainframes, and supercomputers.

    DATA AND PROGRAM REPRESENTATION

    In order to be understood by a computer, data and programs need to be represented appropriately. There arecoding systems (also called coding schemes) that are used to represent numeric, text, and multimedia data,as well as to represent programs themselves. These concepts are discussed in the next few sections.

    Digital Data Representation

    Most computers today are digital computers; that is, computers that understand only two states -- offand on (typically represented by the digits 0 and 1). The mobile devices, microcomputers, midrangeservers, mainframes, and supercomputers discussed in Chapter 1 are all digital computers. Digitalcomputers do their processing by converting data and programs into strings of 0s and 1s, which theymanipulate at almost unimaginable speed, and then converting the processed results back to a form wecan understand. Converting data to digital form so that a digital computer can use it is called digitaldata representation.

    The two possible states recognized by a digital computer can be represented in a variety of ways. Forexample, electronic current is either present or absent, a circuit is either open or closed, a magneticspot or depression on a storage medium is either present or absent, and so on. This two-state, or

    binary, nature of electronic devices is illustrated in Figure 2-1 . Regardless of the form they take, forconvenience it is common to think of these binary states in terms ofOs and Is. In binaryrepresentation, these Os and Is are called bits, which is short for binary digits. A bit is the smallestunit of data that a digital computer can recognize. Because computers do all processing andcommunications by representing programs and data in bit form -- in other words, by using just 0s and1s -- binary can be thought of as the computer's "native language."

    People, of course, don't speak binary language. You're not likely to go up to a friend and say,

    0100100001001001

    which translates into the word "HI" using one binary coding system. People communicate with one

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-01.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-01.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    2/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 2

    another in their natural languages, such as English, Chinese, Spanish, and French. For example, thisbook is written in English, which uses a 26-character alphabet. In addition, most countries use anumbering system with 10 possible symbols -- 0 through 9. As already mentioned, however,computers understand only two symbols -- 0 and 1. For us to interact with a computer, a translationprocess from our natural languages to 0s and 1s and then back again is required. When we enter datainto a computer system, the computer translates the natural-language symbols we provide into binary0s and 1s. After it completes processing, the computer system translates the 0s and 1s thatrepresent the processed results back into the appropriate natural language. See Figure 2-2 below.

    Computers represent programs and data through a variety of binary-based coding schemes. Thecoding system used depends primarily on the type of data needing to be represented. Coding systemsfor numerical, text-based, and a few other types of data are discussed in the next few sections.

    Representing Numerical Data: The Binary Numbering System

    A numbering system is a way of representing numbers. The system we most commonly use is calledthe decimal (or base 10) numbering system because it uses 10 symbols -- the digits 0, 1,2, 3, 4, 5, 6,7, 8, and 9 -- to represent all possible numbers. Numbers greater than nine, such as 21 and 683, arerepresented by combinations of these 10 symbols.

    Because we are so familiar with the decimal numbering system, it may never have occurred to us thatwe could represent numbers any other way. However, nothing says that a numbering system has tohave 10 possible symbols. The binary numbering system(sometimes called the base 2 numberingsystem or the true binary numbering system), for example, uses just two symbols (0 and 1) instead of10 symbols. Because this matches the binary nature of computers, the binary numbering system isused extensively with computers to represent and process numbers.

  • 8/10/2019 Tutorial EXL1 Lecture

    3/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 3

    Here's a quick look at how these numbering systems work. In both the decimal numbering systemand the binary system, the position of each digit determines the power, or exponent, to which the basenumber (10 for decimal or 2 for binary) is raised. In the decimal numbering system, going from right toleft, the first position or column (ones column) represents 10or 1; the second position (tens column)represents 10 to the first power, or 10; the third position (hundreds column) represents 10 to the secondpower, or 100; and so forth. Therefore, as Figure 2-3 shows, the decimal number 7,216 is understoodas 7 x 10 to the third power + 2 X 10 to the second power + 1 X 10 to the first power + 6 X 10 to thezero power or 7,000 + 200 + 10 + 6 or 7,216. In binary, the concept is the same but the columns havedifferent place values. For example, the first column is the ones column (for 2 to the zero power), the

    second column is the twos column (2 to the first power), the third column is the fours column (2 to thesecond power), and so on. Therefore, as illustrated in the bottom half of Figure 2-3 , although 1001represents "one thousand one" in decimal notation, in the binary numbering system 1001 equals "nine"(1 x 2to the third power + 0 x 2 to the second power + 0 x 2 to the first power + 1 x 2 to the zero poweror 8 + 1 or 9).

    Coding Systems for Text-Based Data

    While numeric data is represented by the binary numbering system, text-based data is represented byfixed-length binary coding systems specifically developed for text-based data -- namely, ASCII,EBCDIC, and Unicode. Such codes represent all characters on the keyboard that can appear in text

    data, such as numeric characters, alphabetic characters, and special characters such as the dollarsign ($) and period (.).

    ASCII and EBCDIC

    Among the most widely used coding systems for text-based data are ASCII(American StandardCode for Information Interchange) and EBCDIC(Extended Binary-Coded Decimal InterchangeCode). Virtually all PCs use ASCII, developed largely through the efforts of the American

    National Standards Institute (ANSI). ASCII is also widely adopted as the standard forcommunications systems. EBCDIC, developed by IBM, is used primarily on IBM mainframes.

    Both ASCII and EBCDIC represent each character as a unique combination of 8 bits (see Figure2-3 below), although the original version of ASCII was a 7-digit code. A group of 8 bits allows256 (2 to the eighth power) different combinations, so an 8-bit code can represent up to 256characters.

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-03.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-03.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    4/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 4

    This coding scheme leaves more than enough combinations to account for the 26 uppercase and26 lowercase characters used in the English alphabet, the 10 decimal digits, and the othercharacters usually found on a keyboard. Because of this, additional special characters notincluded on a keyboard -- such as some mathematical symbols, drawing characters,punctuation marks, and dingbats (small graphical symbols) -- were also included in thesecodes. Many computer systems can accept data in either ASCII or EBCDIC.

    The 8 bits that represent a character in ASCII or EBCDIC are collectively referred to as abyte. It is important to be familiar with this concept because "byte" terminology is frequentlyused in a variety of computer contexts. For example, document size and storage capacity aremeasured in bytes, based on the amount of data that is contained in the document or that canbe stored on the storage medium. Since one byte holds only a very small amount of data (justone character), prefixes are commonly used with the term byte. A kilobyte (KB)is equal to1,024 bytes, but is usually thought of as approximately 1,000 bytes. A megabyte (MB)isabout 1 million bytes; a gigabyte (GB)is about 1 billion bytes; a terabyte (TB)is about 1trillion bytes; a petabyte (PB)is about 1,000 terabytes (2 to the 50th power bytes); an exabyte(EB)is about 1,000 petabytes (2 to the 60th power bytes); a zettabyte (ZB)is about 1,000exabytes (2 to the 70th power bytes); and a yottabyte (YB)is about 1,000 zettabytes (2 to the

    80th power bytes). Therefore, 5 KB is about 5,000 bytes, 10 MB is approximately 10 millionbytes, and 2 TB is about 2 trillion bytes.

    Unicode

    A newer code for text-based data that is gaining prominence is Unicode. Unlike ASCII, whichis limited to just the alphabet used with the English language, Unicode is a universal codingstandard designed to represent text-based data written in any language, including those withdifferent alphabets, such as Chinese, Greek, and Russian. It is a longer code (32 bits percharacter is common, compared to the 8 bits per character used in the ASCII and EBCDIC

  • 8/10/2019 Tutorial EXL1 Lecture

    5/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 5

    codes) and can represent over one million characters -- more than enough unique combinationsto represent the standard characters in all the world's current written languages, as well asthousands of mathematical and technical symbols, punctuation marks, and dingbats. Thebiggest advantage of using Unicode is that it can be used worldwide with consistent andunambiguous results. Unicode is used widely for Web pages and in recent software programs,such as Windows XP, Mac OS X, Netscape Navigator, and Internet Explorer.

    Coding Systems for Other Types of Data

    So far, our discussion of data coding schemes has focused on numeric and text-based data, whichconsist of alphanumeric symbols and special symbols, such as the comma and semicolon. Graphics,audio, and video data must also be represented in binary form in order to be used with acomputer. Some common coding schemes used with these types of data are discussed next.

    Graphics Data

    Graphics data consists of still images, such as photographs or drawings. One of the mostcommon methods for storing graphics data is in the form of a bitmapa grid of hundreds ofthousands of dots, called pixels (short for picture elements), arranged to represent animage. The color to be displayed at each pixel is represented by some combination of 0s and1s.

    A monochrome graphic, which is the simplest type of bitmapped image, has only two possiblecolors. Suppose that these colors are black and white and that the color white is representedby a 1, while a 0 represents black. Using this scheme, the graphic would be represented to thePC as a black-and-white bitmap, such as the one shown in the top part of Figure 2-5 , and the

    binary representation of that image would require one bit per pixel of storage space.

    More realism can be achieved using a grayscale image. In grayscale images, each pixel canbe not only pure black or pure white but also any of the 254 shades of gray in between. Thus,each dot or pixel could appear in any of 256 possible states. Since it takes a single byte (8bits) to represent 256 (2s) states, grayscale images require one byte of storage per pixel. So,11111111 could represent a pure white pixel, 00000000 a pure black one, and any byte patternin between -- such as 11001000 and 00001010 -- could represent a particular shade of gray (seethe middle part of Figure 2-5 ).

    Data representation of color images works similarly to grayscale images, in that the color ofeach pixel is represented by a specific pattern of bits. Computers often represent color graphics

    with 16 colors, 256 colors, or 16,777,216 colors. In a 16-color image, one-half byte (4 bits, suchas 0000, 1111, or some combination in between) is assigned to each pixel to represent the coloto be displayed in that pixel. In a 256-color image, 1 byte (8 bits) is assigned to each pixel torepresent its color. Finally, in a 16.8-million-color (called photographic quality or true color)image, 3 bytes (24 bits) are used to store the color data for each pixel in the image (see thebottom part ofFigure 2-5 ).

    Theoretically, the higher number of bits used in color coding, the higher the image quality. Inpractice, however, it is sometimes difficult for the human eye to tell much of a quality differencebetween low-end and high-end color images, unless the images have been enlarged. Sixteen-color images, such as those you often see on Web pages, can actually be quite respectable

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-05.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-05.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-05.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    6/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 6

    looking. Because using fewer colors results in a smaller file size and a faster downloadtime. Web developers often reduce both true-color images and 256-color images to 16-colorimages for use on Web pages. This is typically accomplished by a process calleddithering. Dithering produces colors not available on a limited color palette by coloringneighboring pixels with two available colors that appear to blend together to make a differentcolor. For example, your eye will see a lime green color on the screen when several yellow andgreen pixels are placed adjacent to one another.

    When images need to be resized, however, bitmapped images are not the best choice. When a

    bitmapped image is resized, the existing pixels are just made larger or smaller (no new pixelsare added), so the image usually looks distorted or blurry. A better choice in this situation isvector-based images, which use mathematical formulas to represent images instead of a map ofpixels; vector-based images can be resized without losing quality.

    Audio Data

    Audio datasuch as the sound of someone speaking or a song on a CD -- must be in a digitalformat in order to be stored on a storage medium or processed by a PC. One common audioformat is waveform audio. Waveform audio captures several thousand digital snapshots, calledsamples, of the sound sequence every second. When the samples are played in order, they

    recreate the sound of the voice or music.

    For example, audio CDs record sound using 2-byte samples sampled at a rate of 44,100 timesper second. As you can imagine, when these sampled sounds are played back at a rate of44,100 samples per second, the human ear cannot distinguish them, and they collectively sounlike continuous voice or music. With so many samples per second, however, you can alsounderstand why sound files take up a great deal of storage spaceabout 32 MB for a 3-minutesong (44,100 times x 16 bits x 180 seconds x 2 channels - 8 bits per byte).

    Because of its potentially huge storage size, when audio data is transmitted over the Internet itis often compressed to shorten the download time. For example, files that are MPS-encoded --that is, compressed with the MP3 compression algorithm -- are about 10 times smaller than

    their waveform counterparts, so they download 10 times faster and take up much less storagespace. Technology used to encode and decode audio or video data into binary form is referredto as a codec. For a look at how MP3 compression works, see the "How It Works" box in yourtextbook on page 94.

    Video Data

    Video data -- such as home movies, feature films, and television shows -- is displayed using acollection of frames; each frame contains a still graphical image. When the frames are projectedone after the other -- typically at a rate of 30 frames per second -- the illusion of movement is

    created.

    With so many frames, the amount of data involved in showing a two-hour feature film can bequite substantial. For instance, just a single 256-color image shown on a 640-by-480-pixeldisplay requires 307,200 bytes. When you multiply that figure by 30 times per second, 60seconds per minute, and 120 minutes you get more than 66 gigabytes of information for a typicatwo-hour movie. Fortunately, like audio data, video data can be compressed to reduce it to amanageable size. For example, a two-hour movie can be compressed to fit on a 4.7-gigabyteDVD disc.

  • 8/10/2019 Tutorial EXL1 Lecture

    7/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 7

    Representing Programs: Machine Language

    Just like data, which must be represented by Os and 1 s, programs also need to be represented andprocessed in binary form. Before a computer can execute any program instruction, such as requestinginput from the user, moving a block of data from one place to another, or opening a new window on thescreen, it must convert the instruction into a binary code known as machine language. An exampleof a typical machine-language instruction is as follows:

    01011000011100000000000100000010

    A machine-language instruction may look like a meaningless string of 0s and 1 s, but it actuallyrepresents specific operations and storage locations. The 32-bit instruction shown here, for instance,moves data between two specific memory locations on one type of computer system. Similarinstructions transfer data from memory to other locations, add or subtract values in memory, and soon. The basic set of machine-language instructions that a CPU can understand is known as thatCPU's instruction set. Because the instruction sets used with PC-compatible CPUs are different thanthe ones used with Macintosh CPUs, software programs are typically designated as either for WindowsPCs or for Macs.

    Early computers required programs to be written in machine language. Today, however, hardly anyonewrites machine-language code. Instead, programmers rely on language translators -- programs thatautomatically convert instructions written in a more natural-language -- like programming language intomachine language. The translation is so transparent that most people aren't even aware that it is takingplace. Programming languages and language translators are discussed in more detail in Chapter 13.

    INSIDE THE SYSTEM UNIT

    The system unit is the main case of a computer that houses the processing hardware for that computer, aswell as a few other devices, such as disk drives used for storage, memory, the power supply, and cooling fans The system unit for a desktop PC often looks like a rectangular box, although other shapes and sizes areavailable, such as in the all-in-one PC illustrated in Figure 1-12 and the modified PCs shown in the Inside theIndustry box in Chapter 1.

  • 8/10/2019 Tutorial EXL1 Lecture

    8/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 8

    The inside of a system unit for a typical desktop PC system is shown in Figure 2-6 and the variouscomponents are discussed in the next few sections. Portable PCs have similar components, but many of thecomponents are smaller and the system unit is typically combined with the computer screen to form a singlepiece of hardware.

    The Motherboard

    A circuit board is a thin board containing chips -- very small pieces of silicon or other semiconductingmaterial onto which integrated circuits are embedded -- and other electronic components. The main

    circuit board inside the system unit is called the motherboardor system board. As shown in Figure2-6 , the motherboard has a variety of chips and boards attached to it; in fact, all devices used with acomputer need to be connected in one way or another to the motherboard. Typically, external devices(such as monitors and printers) connect to the motherboard by plugging into a special connector (calledapart) exposed through the exterior of the system unit case. The port is either connected directly tothe motherboard, or is connected through an expansion card plugged into an expansion slot on themotherboard. Ports and system expansion are discussed in more detail later in this chapter.

    The CPU

    The central processing unit (CPU), also sometimes called the processor, consists of a variety of

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-06.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-06.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    9/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 9

    circuitry and components packaged together and plugged directly into the motherboard. The CPU --often referred to as a microprocessorwhen talking about PCs -- does the vast majority of theprocessing for a computer.

    Most PCs today use CPUs manufactured by Intel or Advanced Micro Devices (AMD); some examplesof their processors are shown in Figure 2-7 . CPUs commonly used with desktop PCs include theIntel Pentium 4 and the AMD Athlon XP, although lower-end home PCs may use a Celerpn CPUinstead. A new Athlon CPU named the Athlon 64 and targeted at the high-end desktop market wasreleased in late 2003. Typically , portable computers use either desktop PC CPUs or similar

    processors designed for portable PC use, such as the Intel Pentium M and low-power TransmetaCrusoe', powerful workstations (powerful PCs designed for users running engineering and otherapplications requiring extra processing capabilities) and servers use more powerful processors, such asIntel's Xeon and Itanium 2, AMD's Opteron and Athlon MP, and Sun's UltraSPARC processors.

    Apple Macintosh computers use PowerPC processorsCPUs that were originally developed through acooperative effort by Apple, Motorola, and IBM. The most recent PowerPC CPU is the G5. For a lookat other uses for computer chips besides CPUs, see the Trend box on page 99 of your textbook.

    The type of CPU chip in a computer's system unit greatly affects what can be done with thatPC. Software is sometimes designed to work with a specific chip or chip family, and a program thatworks with one chip may not function on a PC using a different chip. For instance, a program

    designed for a'speedy Intel Pentium 4 chip may not work well, or even at all, on the earlier and far less-capable 80486 chip. Selected CPUs introduced since 2000 are summarized in Figure 2-8 ; thecharacteristics listed in Figure 2-8 are discussed next.

    Processing (Clock) Speed

    Virtually all CPUs come in a variety of processing speeds, also known as clock speeds. Clockspeedindicates how many ticks of the system clock -- the internal clock for a computer thatsynchronizes all its operations -- occur per second. Clock speed is rated in megahertz (MHz) o

    gigahertz. (GHz), which is millions or billions of ticks per second, respectively.

    A higher clock speed means that more instructions can be processed per second than the sameCPU with a lower clock speed. For instance, a Pentium 4 microprocessor running at 3.06 GHzwould be faster than a Pentium 4 running at 2.8 GHz, if all other components remained thesame. CPUs for the earliest PCs ran at less than 5 MHz; today's fastest CPUs run at morethan 3 GHz.

    Although processing speed is an important factor in computer performance, other factors (suchas the amount of memory, the speed of peripheral devices, bus width, and bus speed) greatlyaffect performance, as well.

    Word Size

    A computer word is the amount of data (measured in bits or bytes) that a CPU can manipulateat one time. Different CPUs may have different word sizes. Some newer CPUs are designed fo64-bit words, which means that data moves around within the CPU and from the CPU to memoryin 64-bit (8 byte) chunks. Usually, a larger word size allows faster processing in a computersystem.

    Cache Memory

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-08.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-08.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-07.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    10/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 10

    Cache memoryis a special group of fast memory chips located on or close to the CPU chip. Iis used to speed up processing by storing the most frequently and recently used data andinstructions in handy locations. It works, in theory, the same way you might work at a desk withfile folders or documents that you need most often placed within an arm's length. Other usefulmaterials might be placed somewhat farther away, but still within easy reach.

    The computer works in a similar manner. Although it can access materials in its main memory(RAM) relatively quickly, it can work much faster if it places the most urgently needed materialsinto areas that allow even faster access. These areas -- cache memory -- contain the most

    frequently and recently used data and instructions. To make cache memory more effective,when the cache is full and the CPU calls for a new instruction, the system overwrites the data incache memory that hasn't been used for the longest period of time. This way, the high-priorityinformation that's used continuously stays in cache, while the less frequently used information isoverwritten.

    Cache memory can be internal or external. Internal cache is built right into the CPU chip;external cache is located close to, but not inside, the CPU. Level numbers indicate the distancebetween a particular cache and the CPU: Level 1 is closest (virtually always internalcache). Level 2 (usually internal cache) is the next c losest, and Level 3 (appearing now in newecomputers as external cache) would be further away.

    The cache closest to the CPU is always faster but generally costs more and holds less datathan other levels of cache. Typically all levels of cache memory are faster than the PC'sRAM. As you might guess, the computer looks for data in the closest locations firstfirst inLevel 1 cache, then in Level 2 cache, then in Level 3 cache (if it exists), and then inRAM. Typically a larger cache memory results in faster processing.

    Bus Width and Speed

    A busis an electronic path within a computer over which data can travel. You can picture a busas a highway with several lanes; each wire in the bus acts as a separate lane, transmitting onebit at a time. The number of bits being transmitted is dependent on the bus widththe number

    of wires in the bus over which data can travel. Just as in a highway, the wider (or more lanes) abus has, the more data that can be transferred at one time.

    The bus that moves data around within the CPU is usually referred to as the PC's internal bus;the bus that moves data back and forth between the CPU and memory is called the system bus Today's computers typically have a specific system bus to connect the CPU to RAM called thefrontside bus (FSB). The backside bus (BSB) transfers data between the CPU and externalcache. Many CPUs today have 64-bit internal and system buses, although the speed of thesystem bus varies, as indicated in Figure 2-8 . Generally a faster system bus indicates a fastePC, and a faster frontside bus creates less of a bottleneck in the overall performance of thesystem.

    Memory

    When someone uses the term memory in reference to computers, they are usually referring to thecomputer's main memory -- random access memory or RAM. As discussed in Chapter 1, RAM isvolatile and is the location where data and programs are temporarily stored until they are no longerneeded. RAM is comprised of a group of chips attached to the motherboard and is used primarily bythe computer.

    This is in contrast to storage media (discussed in Chapter 4), which are typically disk-based and are

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-08.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    11/36

  • 8/10/2019 Tutorial EXL1 Lecture

    12/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 12

    Memory modules plug into the system board, as shown in Figure 2-6 . Most -- but not all --desktop PCs sold today come with empty memory slots so the user can add additional memorymodules, if needed.

    For example, in the motherboard shown in Figure 2-6 , there are two memory modules alreadyinstalled and room to add an additional two modules, if necessary.

    RAM capacity is measured in bytes. Because the cost of RAM has been decreasing, mostPCs today come with significantly more memory than just a year ago. Very low-end PCs maycome with 128 MB of RAM, but 256 or 512 MB is much more common in desktop PCs. Serversand workstations typically have between 1 and 4 GB of RAM. Today's 32-bit CPUs can onlyaddress up to 4 GB of RAM, so higher amounts of RAM are expected once 64-bit processors(which can address a virtually unlimited amount of RAM) become more widely available. It isimportant for a PC to have enough RAM to run the needed applications (minimum RAMrequirements are almost always stated on a software program's packaging), as well as to workefficiently (more RAM allows more programs to be opened at one time).

    Types of Ram

    Ordinary RAM is often referred to as DRAM (dynamic RAM) in technical literature. Thisname refers to its need for regular recharging as processing takes place. Duringrecharge, DRAM cannot be accessed by the CPU. A faster type of RAM -- called SRAM(static RAM) -- does not need recharging. Since SRAM chips are more expensive thanDRAM, computer makers use them sparingly.

    A newer type of DRAM is SDRAM (synchronous DRAM). SDRAM is much faster thanDRAM because it is synchronized to the system clock; RDRAM (Rambus DRAM) is eve

    faster. Double-data rate (DDR) SDRAM sends data twice per clock tick so it is twice asfast as the same speed SDRAM. DDR SDRAM was the standard for desktop PCs at thetime this book was published; double-data rate-II (DDR-II) SDRAM, which sends aboutfour times as much data per clock tick as SDRAM with half the power consumption, isexpected to be released shortly.

    Memory Speed

    Memory speed is typically measured in megahertz (MHz) like CPU speed, althoughsome memory manufacturers measure speed in nanoseconds (billionths of a second)instead. Common speeds in megahertz are 133 MHz for SDRAM, 133 MHz for SRAM,

    400 MHz for DDR SDRAM, 667 MHz for DDR-11 SDRAM, and 1,066 MHz (1.066 GHz)for RDRAM. Memory manufacturers are continually working on ways to make memoryfaster. Having components, such as RAM or buses, operate at a slower speed than theCPU has caused processing bottlenecks in the past, where the CPU sits idle while itwaits for the appropriate component to finish what it needs to do first. Increasing thespeed of slower components helps to avoid this problem.

    Some memory today is dual-channel memory, which can transfer twice as much data atone time as single-channel memory of the same speed. It is expected that somememory in the near future will be four-channel. Because of these newer types ofmemory, as well as double-data rate RAM, it is becoming more common and moreaccurate to express memory speed in terms of its bandwidth; that is, the actual amount

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-06.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-06.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    13/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 13

    of data transferred (in GB) per second. For example, single-channel DDR SDRAMrunning at 400 MHz transfers 3.2 GBps (gigabytes per second), while dual-channel DDRSDRAM running at 400 MHz transfers data at a rate of 6.4 GBps.

    Memory Addressing

    Once the CPU stores something in RAM, it must be able to find it again whenneeded. To accomplish this, each location in memory has an address. Whenever a

    block of data, instruction, program, or result of a calculation is stored in memory, it isstored into one or more consecutive addresses, depending on its size (see Figure 2-10). Computer systems automatically set up and maintain directory tables that provide theaddresses of the first character of all data stored in memory, along with the number ofaddresses used, to facilitate the retrieval of data.

    When the computer has finished processing a program or set of data, it frees up thatmemory space to hold other programs and data. Thus, the content of each memorylocation constantly changes. This process can be roughly compared with the handling ofthe mailboxes in your local post office: the number on each P.O. box (memory location)remains the same, but the mail (data) stored inside changes as patrons remove their maiand as new mail arrives.

    Registers

    Registers are another type of high-speed memory built into the CPU. They areused by the CPU to temporarily store each program instruction and piece of data

    just before it is processed. Registers are the fastest type of memory used by theCPU, even faster than Level 1 cache. Generally, the more data a register cancontain at one time, the faster the CPU performs. Register size usually matchesword and internal bus size and so is usually 32 bits, although the newer 64-bit

    chips such as the Itanium 2, Opteron, and Athlon 64 use 64-bit registers and wordsizes (refer again to Figure 2-8 ). Most CPUs contain several registers; registersare discussed in more detail later in this chapter.

    ROM

    ROM (read-only memory)consists of non-volatile chips into which data orprograms have been permanently stored; the data or programs are retrieved by thecomputer when they are needed. Like RAM, these chips are attached to themotherboard inside the system unit. An important difference, however, is that youcan neither write over the data or programs in ROM chips (which is the reason

    ROM chips are called read-only), nor destroy their contents when you shut off thecomputer's power.

    Important pieces of system software are often stored in ROM chips. Forinstance, one of the computer's first activities when you turn on the power is toperform a power-on self-test or POST. Traditionally, the instructions for the POSThave been stored in ROM. POST takes an inventory of system components,checks each component for proper functioning, and initializes system settings,which produces the beeps you hear as your PC boots up.

    Flash Memory

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-08.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-10.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    14/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 14

    Flash memory(sometimes called flash RAM) is a type of non-volatile memorythat can be erased and reprogrammed in units of memory called blocks. Flashmemory gets its name because a block is erased in a single action or flash. Thistype of organization makes flash memory faster than RAM, although it is moreexpensive than RAM and has to store data in block-size, not byte-size, pieces.

    Flash memory chips have begun to replace ROM for system information, such asto hold a PC's BIOS or basic input/output system -- the sequence of instructions

    the PC follows during the boot process, including the POST discussed earlier. Bystoring this information in flash memory instead of ROM, the BIOS information canbe updated as needed. Flash memory chips are also used to store programs inportable PCs and mobile devices.

    While some flash memory chips within a PC are used only by the computer, otherflash memory chips are used by the user. Because flash memory chips are smaand hold a significant amount of data, they are commonly used as storage mediafor portable PCs, as well as for cell phones, digital cameras, MP3 players, andother small devices. Flash memory chips used for user storage can be builtdirectly into the PC or other device; they are also built into a variety of removableflash memory media, such as flash memory cards, sticks, and drives. Flash

    memory media is discussed in more detail in Chapter 4.

    Buses

    As already discussed, a bus is an electronic path within a computer over which datatravels. In addition to moving data within the CPU, buses are used to tie the CPU tomemory and peripheral devices. An example of how the buses in a PC might be set upis shown in Figure 2-13 .

    Buses that connect the CPU to peripheral (typically input and output) devices are usuallyreferred to as expansion buses. Expansion buses, typically 16 to 64 bits wide, areetched onto the motherboard, and either connect the CPU to ports on the system unitcase or to expansion slots on the motherboard. There are a variety of expansion busstandards. As illustrated in Figure 2-13 , most CPUs today use a chipset of one or morechips that help bridge or connect the various buses to the CPU. Some of the mostcommon types of expansion buses are discussed next; expansion slots and ports arediscussed shortly.

    ISA Bus

    The ISA (Industry Standard Architecture) bus has been around since1984. Although it is still included on some new PCs (typically connected to thePCI bus, as shown in Figure 2-13 ), it is beginning to be phased out and replacedby the newer bus standards discussed next. The ISA bus is slow, transmitting 8or 16 bits at one time. Consequently, when an ISA bus is used, it is for slowerdevicessuch as the system's mouse or sound card.

    PCI Bus

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-13.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-13.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-13.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    15/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 15

    The PCI (Peripheral Component Interconnect) bus is one of the most commontypes of expansion buses. It can transmit up to 64 bits at a time (although 32 bitsis more common) and is significantly faster than the ISA bus, delivering abandwidth of more than IWMBps (megabytes per second). A newer, faster versionof PCI introduced in 2000 is called PCI-X. PCI-X supports a 64-bit bus width and abandwidth of more than 1,000 MBps (1 GBps). In addition to connecting the CPUto the PCI expansion slots, the PCI bus is used to connect other buses to theCPU, such as the ISA and USB buses, as shown in Figure 2-13 . Eventually,the PCI bus may be replaced by a faster architecture, such as a newer version of

    PCI called PCI Express, expected to be available by the end of 2004 with an initialbandwidth of 16 GBps, and the emerging InfiniBand architecture, expected to beused primarily with servers and support a bandwidth exceeding 2.5 GBps.

    AGP Bus

    The AGP (Accelerated Graphics Port) bus is a relatively new bus standard thatwas developed in response to the trend toward greater performance requirementsfor graphics display. With the increasing demands of 3D graphics and full-motionvideo playback, both the processor and the video card need to process moreinformation much more quickly than before. Although the standard PCI bus has

    been adequate for video displays in the past, AGP provides improved performancewith a fast, 32-bit, direct interface between the video card and RAM, operating atover 2,000 MBps (2 GBps).

    USB Bus

    One of the more versatile new bus architectures in recent years is the UniversalSerial Bus (USB). The USB standard enables up to 127 devices to be connectedto a computer's PCI bus through a single port on the computer's system unit. At1.5 MBps, the original USB standard transmitted data at a much slower rate thanboth PCI and AGP. The newer USB 2 standard supports data transfer rates of 60

    MBpsstill slower than PCI and AGP, but the convenience and universal supportof USB have made it one of the most widely used standards for peripherals today.

    FireWire / IEEE 1394 Bus

    FireWire (also known as IEEE 1394) is a newer external bus standard developedby Apple. Like USB, FireWire can connect multiple external devices to the PCIbus via a single port. FireWire is fairly fast, supporting data transfer rates of up to40 MBps, and is commonly used with video cameras and other multimediaperipherals for both Apple and IBM-compatible computers. The even newerFireWire 2 (IEEE i394b) standard supports data transfer rates up to 100 MBps. It

    is expected that USB and/or FireWire will eventually replace the ISA and othertypes of older buses.

    Expansion Slots and Cards

    Most PCs have a way to add additional capabilities to the PC when new needsarrive. For instance, a user may need to add a modem for Internet connectivity, a bettersound card, a new storage device or printer, or a network interface. The overall waysthese devices are added depends on the type of PC (desktop, portable, or handheld)

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-13.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    16/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 16

    being used; the specific hardware that can be added to a PC also varies from computer tocomputer.

    Expansion Slots for Desktop PCs

    Most desktop PC motherboards have a variety of expansion slotsinto which

    expansion cards(also called add-in boards, interface cards, and adapter boards)can be inserted (refer again to Figure 2-6 ). As shown in Figure 2-13 , theexpansion slots (such as PCI, AGP, and ISP) are connected to the. appropriatebus on the motherboard. It is important to realize that expansion slots are notinterchangeable -- ISA slots are larger than PCI slots and AGP slots are a littlesmaller than PCI slots -- and each expansion slot is designed to be used with itscorresponding expansion card. Figure 2-12 shows a typical expansion card andlists a variety of common types of expansion cards. As shown in this figure, mosexpansion cards have a port that is exposed through the case of the system unit;this port is used to connect the external device (such as a monitor, printer,scanner, or networking cable) being added to the system. Ports are discussed inmore detail shortly.

    Instead of using expansion boards and tying up valuable expansion slots inside thesystem unit, some PCs integrate sound, video, or other capabilities into chips thatare attached directly to the motherboard. When an integrated feature is used, aspecial bus connects that chip to the appropriate port on the outside of the systemunit so the appropriate device can be attached to the PC.

    PC Cards: Exapnsion for Notebook and Tablet PCs

    Because space is so limited inside portable PCs, usually expansion cards anddevices are not installed inside the system unit. Instead, most notebook and

    tablet PCs come with a PC card slot(sometimes called a PCMCIA card slot)that conforms to the standards developed by the Personal Computer Memory CardInternational Association. PC card slots are typically located on the side of anotebook or tablet computer's case (see Figure 2-11a ); these slots connect tothe PC card bus inside the PC.

    PC cardscan be typically plugged in or unplugged while the computer is turned onand they come in three different thicknesses, from Type I (the thinnest) to Type III(the thickest). Some notebook PCs may have a slot that can hold more than oneof the thinner PC cards -- it is a good idea to double-check the number and sizesof PC cards your computer can use before buying a new PC card. In addition toPC card slots, notebook computers often have a universal drive bay which can be

    used to alternate between a second hard drive, CD or DVD drive, and floppy drive,as needed.

    Expansion for Handheld PCs and Mobile Devices

    Most handheld PCs and mobile devices have a limited amount of expandability,but most come with at least one built-in expansion slot designed for specialexpansion cards used to attach peripheral devices or add additionalcapabilities. Increasingly, the slot is designed for postage-stamp-size SDcardsflash memory cards and hardware adhering to the Secure DigitalInput/Output (SDIO) standard (see Figure 3-13 below).

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig03-12.jpghttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-12.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-13.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-06.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    17/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 17

    Although SD slots are often used with flash memory storage cards (as discussedin Chapter 4), SD slots will also accept Secure Digital Input/Output (SDIO)devicessmall devices, such as a digital camera or networking adapterthat areattached to an SD card. A limited number of SDIO devices are currently availablebut more are expected to become available in the near future. Some handheldPCs support additional types of flash memory cards; a few include a PC card slot.

    Ports

    Most system units feature exterior connectors or sockets called portsthat enable you toplug in external hardware devices. Each port is attached to the appropriate bus on themotherboard so that the devices plugged into each port can communicate with theCPU. Typical ports for a desktop PC are shown in Figure 2-14 . As shown in this figurethere are unique connectors for each type of port on a computer system.

    When connecting cables to the system unit, it is important to pay attention to the genderof the port, in addition to the shape, pin count, and pin configuration. Male connectorshave the pins extended and connect to female connectors with matching holes. If a portis of the proper type (such as serial), but is the wrong gender or has the wrong number ofpins, adapters or special cables can sometimes be used to convert the connector to thedesired configuration.

    To add a new device to your PC, if there is an available port for the device you want to addyou just need to plug it in (shut down the computer first, unless the device uses a USB oFireWire port). If there isn't an available port, you need to either insert the appropriateexpansion card to create one or use a USB or FireWire version of the device, if you haveone of those two ports available on your PC.

    Because a wide variety of hardware is available in USB format today, most recent PCscome with at least two USB ports. In fact, it is becoming common to see USB portslocated on the front of a system unit for easier access, such as to connect a digitalcamera or USB flash memory drive on a regular basis. See Figure 2-15 below. Some ofthe most common ports are discussed next.

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-14.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    18/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 18

    Serial portscan transmit data only a single bit at a time. However, they use veryinexpensive cables and they can reliably send data over long distances. Serialports can be used for such devices as keyboards, mice, and modems, althoughmost systems today come with dedicated ports to attach the mouse and

    keyboard, as shown in Figure 2-14 . Serial connectors typically have 9 or 25 pinsand are referred to as DB-9 or DB-25 connectors.

    Parallel portscan transmit data one byte (8 bits) at a time -- making datatransfers several times faster than those through serial ports -- but they requiremore expensive cables and cannot reliably send data across distances greaterthan 50 feet. Parallel ports typically connect nearby printers and scanners to aPC. Newer types of parallel ports include the Enhanced Parallel Port (EPP) andthe Extended Capabilities Port (ECP). These ports look like conventional parallelports and accept the same size and shape of plug as the conventional parallel port(a 25-pin connector), but are more than 10 times faster when used with anappropriate cable. When used with a regular parallel cable, the speed of these

    ports is similar to that of a conventional parallel port.

    SCSI (Small Computer System Interface) portsare highspeed parallel portsgenerally used to attach printers or scanners.

    USB portsare used to connect USB devices to the computer. Most new PCscome with two USB ports, but a USB hub -- a device that plugs into your PC'sUSB port to convert one port into several USB ports (see Figure 2-15) -- can beused to connect multiple USB devices to a single USB port, when necessary. Inaddition, USB devices are hot-swappable, meaning that they can be attached andremoved while the computer is turned on.

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-14.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    19/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 19

    FireWire (IEEE 1394) portsare used to connect FireWire devices to the

    computer. Similar to USB, a FireWire hub can be used to connect multipledevices to a single port and PireWire devices are hot-swappable.

    Network portsare used to connect a PC to a local area network. Most networkcards contain a port that accepts an RJ-45 connector, which looks similar to atelephone connector but is larger. Coaxial cable or fiber optic connectors can beused for network connections, as well. Networks are discussed in more detail inChapter 8.

    The keyboard portand mouse port typically use a PS/2 connector and are usedto connect the keyboard and mouse to the system unit.

    The monitor portcan be connected directly to the motherboard or can be locatedon an expansion card. In either case, it is used to connect the monitor to a PC.

    The modem portand phone port typically appear on a modem card. The modemport is used to connect the modem card to your phone outlet. The phone port canbe used to connect a telephone, if desired, so you can still have a telephoneconnected to the phone jack being used.

    A MIDI portis used to connect a MIDI (musical instrument digital interface) deviceto the computer. MIDI devices include musical keyboards and other instrumentsthat can be connected to the computer to compose music to be storedelectronically. A MIDI port usually looks similar to a keyboard port.

    An IrDA (Infrared Data Association) portreceives infrared transmissions fromsuch devices as wireless keyboards, wireless mice, and portable devices. Sincethe transmission is wireless, the port doesn't use a plug. With infraredtransmission, there cannot be anything blocking the infrared light waves, so newerwireless mice and keyboards tend to use radio wave transmission instead.However, IrDA ports are commonly used to "beam" data from a handheld PC orother portable device to another PC.

    A game portis used to connect a joystick, gamepad, steering wheel, or otherdevice commonly used with computer game programs.

  • 8/10/2019 Tutorial EXL1 Lecture

    20/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 20

    Notebook computers have ports similar to desktop PCs, but sometimes don't have asmany. One type of port not often found on desktop PCs is one used to connect a portreplicator -- a hardware device containing additional ports (such as serial, parallel, PS/2,USB, and networking ports) that can be used with the notebook whenever the portreplicator is connected (some port replicators connect via a USB or PC card slot instead)

    Some typical notebook ports are illustrated in Figure 2-16 . Additional ports not shown

    in this figure that are commonly found on notebook computers include IrDA ports forinfrared transmission and those used to connect the notebook to a docking station -- adevice with peripheral devices (such as a monitor, keyboard, mouse, and printer) alwaysattached; these devices can be used whenever the PC is attached to the docking station.

    As already discussed, most handheld PCs and some smart phones have an external porthat accepts flash memory cards; some also have proprietary ports for additionaldevices. For instance, many Palm PCs have a Universal Connector to which a variety ofhardware devices can be attached (see Figure 2-18 ).

    Some portable PCs made by Handspring (which was recently acquired by Palm) useproprietary Springboard modules to add additional devices, such as a digital camera or

    GPS receiver. Many portable PCs and some smart phones also have a port to connect aspecial portable keyboard designed for the device. Common add-on devices for portablePCs include modems, networking adapters, keyboards, and MP3 players.

    HOW THE CPU WORKS

    Regardless of size, every computer's CPU is basically a collection of electronic circuits andcomponents. Electronic impulses from an input device pass through RAM and enter the CPUthrough a system bus. Within the CPU, these impulses move through the circuits and various

    components (as directed by the program) to create a series of new impulses. Eventually, a setof electronic impulses leaves the CPU headed for an output device.

    The key element of the microprocessor is the transistor -- a device made of semiconductormaterial that acts like a switch controlling the flow of electrons inside a chip. Today's CPUscontain tens of millions of transistors. A recent breakthrough reduced the size of the transistorby 30%, which will enable future microprocessors to contain a billion transistors. Thesemicroscopic transistors can turn on and off more than a trillion times per second andmicroprocessors created with these transistors could complete close to a billion calculations inthe blink of an eye -- a significant speed increase over current CPUs. The primary componentsof a CPU are discussed next.

    Typical CPU Components

    To begin to understand how a CPU works, you need to know how the CPU is organizedand what components it includes. This information will help you understand howelectronic impulses move from one part of the CPU to another to process data. Thearchitecture and components included in a CPU (referred to as microarchitecture) varyfrom microprocessor to microprocessor. A simplified example of the principalcomponents that may be included in a typical CPU is shown in Figure 2-19 anddiscussed next.

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-19.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-18.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-17.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    21/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 2

    Arithmetic / Logic Unit (ALU)

    The arithmetic/logic unit (ALU)is the section of the CPU that performsarithmetic (addition, subtraction, multiplication, and division) and logical operations(such as comparing two pieces of data to see if they are equal or determining if aspecific condition is true or false). In other words, it 's the part of the CPU thatcomputes. The fact that the CPU can perform only basic arithmetic and logicaloperations might not seem very impressive, but when combined in various ways at

    tremendous speeds, these operations enable the computer to perform immenselycomplex and data-intensive tasks, in a very short period of time -- a remarkablefeat.

    For example, editing a digital photograph in an image-editing program, running thespelling checker in a word processing program, and burning a music CD are allperformed by the ALU using only arithmetic and logical operations.

    Control Unit

    The control unitcoordinates and controls the computer's operations, such as

    retrieving instructions and passing them on to the ALU for execution. In otherwords, it directs the flow of electronic traffic within the CPU, much like a traffic copcontrols the flow of vehicles on a roadway. Essentially, the control unit tells the

    ALU what to do and makes sure that everything happens at the right time in orderfor the appropriate processing to take place.

    Registers

    As discussed earlier in this chapter, registersare groups of high-speed memorylocated within the CPU into which data is transferred just before processing. The

    ALU uses registers to temporarily store data, intermediary calculations, and the

    final results of processing. A CPU may have a variety of registers for differentpurposes, such as an instruction register to hold instructions, and an accumulatorregister to hold intermediary processing results. One of Intel's most recent chips -- the Itanium 2 -- has a total of 328 different registers.

    Prefetch Unit

    The prefetch unit, present in many recent microprocessors, orders data andinstructions from cache or RAM based on the task at hand. By attempting toretrieve the necessary instructions and data ahead of time, the prefetch unit helpsto avoid delays in processing. It also ensures that all instructions are lined up

    correctly to send off to the decode unit.

    Decode Unit

    The decode unittranslates instructions into a form that can be processed by theALU and stored in the registers. After decoding, the instructions go to the controunit for processing.

    Internal Cache

  • 8/10/2019 Tutorial EXL1 Lecture

    22/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 22

    As mentioned earlier, internal cache(such as Level 1 and Level 2 cache) is usedto store frequently used instructions and data. If the necessary items are not ininternal cache, they must be retrieved from external cache or RAM.

    Bus Interface Unit

    The bus interface unitis the place where instructions and data flow in and out ofthe CPU. It connects the CPU to the system bus to connect the CPU to external

    cache and RAM.

    The System Clock and the Machine Cycle

    As mentioned at the beginning of this chapter, every instruction that you issue to acomputereither by typing a command or clicking something with the mouse -- isconverted into machine language. In turn, each machine-language instruction in a CPU'sinstruction set (the collect ion of basic machine-language commands that the CPU canunderstand) is broken down into several smaller, machine-level instructions called

    microcode. Microcoded instructions, such as moving a single piece of data from one parof the computer system to another or adding the numbers located in two specificregisters, are built into the CPU to provide its basic instructions.

    As previously discussed, a PC's system clocksynchronizes the computer'soperations. This clock typically resides on the motherboard and sends out a signal on aregular basis to all other computer components, similar to a musician's metronome or aperson's heartbeat. Each signal is referred to as a cycle; the number of cycles persecond is measured in hertz (Hz). One megahertz (MHz) and one gigahertz (GHz) areequal to one million and one billion ticks of the system clock, respectively.

    During each tick or clock cycle, the CPU executes a certain number of pieces of

    microcode. Consequently, a PC with a higher system clock speed processes moreinstructions per second than the same PC using a lower system clock speed. Oldercomputers may take more than one clock tick for each instruction; today's fastersuperscaler computers are able to perform multiple instructions per clock tick. Differentcomputers can process a different number of instructions per clock t ick. Consequently,for a more universal measurement, processing speed can be measured in the number ofinstructions the CPU can process per second instead of the number of cycles persecond. In this case, the terms mips (millions of instructions per second), megaflops,gigaflops, and teraflops (millions, billions, and trillions of floating-point operations persecond, respectively) are used. Although originally used only with mainframe andsupercomputers, today's microcomputers have become so fast that their speeds are nowsometimes being quoted using this terminology. For example, Apple rates the speed of

    its PowerPC G4 CPU at over one gigaflop.

    Despite the number or fraction of clock ticks required per instruction, when the CPUprocesses a single piece of microcode, it is referred to as a machine cycle. Eachmachine cycle consists of the four operations that follow (see Figure 2-20 ):

    1. Fetch-- the program instruction or information about a needed piece of data isfetched.

    2. Decode- the instructions are decoded so the control unit and ALU can

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-20.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    23/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 23

    understand them.

    3. Execute-- the ALU executes the appropriate instruction, if any arithmeticcomputation or logical comparisons are required.

    4. Store- the original data or the result from the ALU's execution is stored either in aregister or memory, depending on the instruction.

    Because a machine cycle only processes a single instruction, many seemingly simplecommands (such as adding two numbers) may require more than one machine cycle,and a computer may need to go through thousands, millions, or even billions of machinecycles to complete a single program's processing. Consequently, the machine cycle isrepeated as many times as necessary to complete a user command or programinstruction. A simplified example of how a CPU might process the command 2 + 3 = isillustrated in Figure 2-21 . In this example, four machine cycles are used, as follows:

    1. The number 2 is fetched and temporarily stored in internal cache, then sent to thedecode unit where it is determined that the number needs to be stored, and thensent with the proper command to the control unit, which stores the number in aregister.

    2. The number 3 is fetched and temporarily stored in internal cache, then sent to thedecode unit where it is determined that the number needs to be stored, and thensent with the proper command to the control unit, which stores the number in aregister.

    3. The addition symbol is fetched and temporarily stored in internal cache as anaddition instruction, which is then sent to the decode unit where it is determined

    that the two numbers held in the registers need to be added. That instruction isthen sent to the control unit, which instructs the ALU to add the two numbers andstore the result in another register.

    4. The equal sign is fetched and temporarily stored in internal cache as an outputinstruction, which is then sent to the decode unit where it is determined that thesum located in a register needs to be output. That instruction is then sent to thecontrol unit, which outputs the sum located in a register.

    MAKING COMPUTERS FASTER NOW AND IN THE FUTURE

    Over the years, computer designers have developed a number of strategies to achieve fasterperformance in their machines. Researchers are also constantly working on ways to speed upthe computers of the future. This section discusses several of these methods. There are alsosome things that you can do to speed up your computer today, as discussed next.

    Speeding Up Your System Today

    Several strategies that may speed up your current computer are listed next.

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-21.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    24/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 24

    Add More Memory

    With today's graphic-intensive interfaces and applications, much more memory isrequired than was necessary even a couple of years ago. If you seem to have towait too long when opening programs or saving documents and your CPU isrelatively fast (over 1 GHz), you should consider adding more memory to your

    system to bring it up to a minimum of 512 MB. Be sure to check inside your PCfirst to determine if there is room for more memory modules, and then check withyour PC manufacturer to determine the appropriate type and speed of RAM thatyour PC uses. If all your RAM slots are full, you can usually remove some of theold modules and replace them with newer, higher capacity ones.

    Perform System Maintenance

    As you work and use your hard drive to store and retrieve data, and as you installand unin-stall programs, most PCs tend to become less efficient. Part of thereason for this is as large documents are stored, retrieved, and then stored again,

    they often become fragmented -- that is, not stored in contiguous (adjacent)storage areas. Because the different pieces of the document are physicallylocated in different places, it takes longer for the computer to retrieve or storeit. Another reason a computer may become inefficient is that when programs areuninstalled, pieces of the program are sometimes left behind or references tothese programs are left in operating system files. Yet another reason is that as ahard drive begins to get full, it takes longer to locate and manipulate the datastored there. All of these factors can result in a system performing more slowlythan it should.

    To alleviate some of these problems, regular system maintenance should beperformed. Here are some ideas:

    Uninstall any programs that you no longer want on your computer to free upspace on your hard drive. Be sure to use the designated removal procedurefor your operating system such as the Add/Remove Programs option in theWindows Control Panel or an "Uninstall" option for that program located onthe Start menu to remove the program for Windows PCs. See Figure 2-22

    If you have large files (such as digital photos or other graphical images)stored on your computer that you don't need on a regular basis, considermoving them to a removable storage medium, such as a Zip disk, CD-Rdisc, or DVD-R disc. Once copied onto the new medium, the files can be

    deleted from your hard drive to free up space. If the files are important, youmay want to make two copies, just to be safe. Be sure to open the filesfrom the storage medium to confirm that the transfer was successful beforedeleting the files from your hard drive and store multiple disks containingthe same files in different locations. Copying files, deleting files, andstorage media will be discussed in more detail in later chapters.

    Delete the temporary files stored by your browser (choose Internet Optionsfrom the Tools menu, if you are using Internet Explorer; select Preferencesfrom the Edit menu, and then choose Advanced and Cache in NetscapeNavigator). Browsers tend to store a huge amount of data -- it's notunusual to have over 1,000 temporary Internet files -- including your

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-22.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    25/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 25

    browsing history and copies of Web pages you've recently visited. If you nolonger need this information, delete the files to make room on your harddrive. Deleting temporary Internet files can also speed up your Internetbrowsing.

    Open the Recycle Bin (or similar location holding deleted files on your PC)and empty it. As long as you are sure that none of the files in the RecycleBin needs to be restored, those files are taking up room on your hard driveneedlessly.

    Scan for computer viruses and spyware regularly. A computer that sudenlyslows down might be the result of extra workload created by a computervirus, spyware program, or other threat. Anitivirus and antispywareprograms can help detect and correct these problems and are discusses inmore detail ion chapter 9. If you are using Windows XP, you can alsocheck the current CPU load any time by pressing the Ctrl+Alt+Delete keycombination one time. The resulting Windows Task Manager screen canbe used to view all open programs and processes, as shown in Figure 2-23below.

    If the CPU load stays very high for a long period of time, look for a processwith a high CPU percentage. Some processes, such as print spoolers,occasionally hang (stay running) after they should have finished; if thishappens, the process can be shut down manually using this screen. Butbe sure you know what a process does (use an Internet search site toresearch it, if needed) before shutting it down to avoid creating problemswith your PC.Use a utility program, such as Windows' Disk Cleanup, to deleteunnecessary files left over from installing programs, uninstalling programs,and browsing the Internet. There are also a variety of utility programs thatcan be purchased to clean up your hard drive (see Figure 3-21).

  • 8/10/2019 Tutorial EXL1 Lecture

    26/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 26

    Use a utility program, such as Windows' Disk Defragmenter, to arrange thefiles on your hard drive more efficiently. On large hard drives, this mayneed to be done during the night because of the length of time required. Defragmentation, utility programs, and other system maintenanceactivities are discussed more in Chapter 6.

    Buy a Larger or Second Hard Drive

    As already mentioned, hard drives get less efficient as they fill up. If your harddrive is almost full and you don't have any data or programs that you can remove,you should consider buying and installing a second hard drive, assuming that youhave an empty storage bay inside your computer. Some users can comfortablyinstall a second hard drive themselves; others prefer to have the new drive installedfor them.

    In either case, once the drive is installed, you can either transfer your data to the

    new drive and use it solely as a data drive, or uninstall a few of the larger programson your first drive and install them again on the new drive. Alternatively, you canreplace your existing hard drive with a larger one, although the data transferprocess will be a litt le more complicated. Before purchasing a second hard drive,be sure to check if there is room inside your PC for an additional drive; if not, youcan buy a USB hard drive, assuming you have a free USB port.

    Upgrade Your Internet Connection

    If your system seems slow primarily when you are on the Internet, the culprit mighbe your Internet connection. If you are using standard dial-up access, you can

  • 8/10/2019 Tutorial EXL1 Lecture

    27/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 27

    sometimes upgrade your service to a premium faster service for a little moremoney per month. If you use the Internet a great deal for work or multimediapurposes, consider switching to cable, satellite, DSL, or other type of broadbandInternet service. Although they are more expensive, they are significantlyfaster. The differences between these and other types of Internet connections aredescribed in Chapters 8 and 9.

    Upgrade Your Video Card

    If you're a gamer, computer artist, graphical engineer, or otherwise use 3D-graphic-intensive applications, consider upgrading your video card to one that bettersupports 3D graphics. If you don't use 3D applications, upgrading your video cardwon't usually improve your speed.

    Upgrade Your CPU

    Upgrading your CPU used to be a popular choice, since the price differencebetween a new system and a new CPU was relatively large. With today's lowcomputer prices and fast basic processor speeds, however, this upgrade is

    becoming less common. If you have a fairly slow Pentium chip (less than 600MHz), you might want to look into whether your motherboard would support afaster chip and the cost involved, but most of the time it would be better to put thatmoney into a new system.

    Strategies for Making Faster Computers

    There are several strategies that researchers and manufacturers are working on to buildfaster PCs. Some are described next.

    Moving Circuits Closer Together

    As complex as computers seem, all must follow one intuitive natural law: Circuitsthat are physically located closer together require less time to move bits betweenthem. During the past several decades, computer-chip makers have packedcircuitry closer and closer together. Today, they fit several million circuits on asingle, fingernail-size chip. This remarkable achievement doesn't mean that chipmakers can continue to shrink circuitry without constraint, however.

    When the circuits are in use, they generate heat, and circuits placed too closetogether can become hot enough to melt the chip. For this reason, virtually allCPUs today employ fans, heat sinks (small components typically made out ofaluminum with fins that help to dissipate heat), or some combination of the two tocool the CPU. Low-power CPUs that run cooler than traditional CPUs are beingintroduced and new cooling technologies are continually being tested, such asliquid-filled tubes that act as radiators to draw heat away from processors andmisters that spray a fine mist of liquid onto chips when they become too hot.

    At the time of this writing, computer manufacturer NEC had recently released whathey call the world's first water-cooled PC. This PC, called the Valuestar, uses

  • 8/10/2019 Tutorial EXL1 Lecture

    28/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 28

    water to cool the CPU. An added bonus is that operating noise is cut in half,because a CPU fan is not required.

    Faster and Wider Buses and Faster Memory

    As discussed earlier, wider buses lead to faster processing. Buses also vary inspeed; faster buses tend to result in faster processing. Buses today run as fastas 1,000 MHz. Faster memory also helps speed up processing. Memory today

    typically runs at over 100 MHz. One of the faster types of RAM available --Rambus RDRAM -- runs at over 1 GHz in hopes of more closely matching memoryspeed with CPU speed to reduce or eliminate bottlenecks.

    Improved Materials

    Traditionally, CPU chips used aluminum circuitry etched onto a siliconbacking. As the limit of the number of aluminum circuits that can be packed ontoa silicon chip without heat damage or interference approached, chip makers beganto look for alternate materials. Copper was one of the next choices, since it is afar better electrical conductor and it can produce chips containing more circuitry at

    a lower price. A more recent development is SOI (silicon on insulator). SOI chipsuse a thin layer of insulating material over the silicon to reduce heat and powerconsumption. This results in being able to place the circuits closer together thanis possible without the insulating material.

    One possibility for the future is the use of superconductive materials. Superconductive materials resist heat buildup and can be used to pack a chip'scircuitry closer together, resulting in chips with faster speeds. Two technologiesbeing researched are ceramic superconductors and a new form of carbonconsisting of 60 carbon atoms surrounding a hollow cluster. When one or morealkali metals are inserted into the empty space between the carbon atoms, theresulting compoundcalled a fullerideoften becomes a superconductor.

    For integration into clothing and other flexible materials, a number of companiesare developing flexible electronic components, such as the flexible microprocessorshown in Figure 2-24 below.

  • 8/10/2019 Tutorial EXL1 Lecture

    29/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 29

    In addition to the ability to be bent without damaging the circuitry, flexiblemicroprocesors are thinner, lighter, generate little heat, and consume significantlyless energy than conventional microprocessors.

    An even more radical idea currently in the testing stages is replacing the CPUentirely with field-programmable gate arrays {FPGAs). An FPGA is a type of chipthat can be programmed and reprogrammed as needed. To replace a CPU, suchas in the system developed by Wincom Systems that does the work of 50 or more

    $5,000 servers but costs only $25,000, groups of these chips work together toprocess several tasks at the same time. In addition to being a possiblereplacement for CPUs in future PCs, FPGAs are currently being used in storagedevices, networking hardware, cell phones, and digital cameras.

    Pipelining

    In older PC systems, the CPU had to completely finish processing one instructionbefore starting another. More recent PCs, however, can process multipleinstructions at one time. One way this is accomplished is through pipelining. With pipelining, a new instruction begins executing as soon as the previous one

    reaches the next stage of the machine cycle. Figure 2-25 illustrates this processwith a 4-stage pipeline. Notice that while the pipelined CPU is executing oneinstruction, it is simultaneously fetching and getting the next instruction ready forexecution. Without a pipeline, the arithmetic part of the processor would be idlewhile an instruction is being fetched and decoded. CPUs today are commonlybuilt with multiple pipelines, which increases the number of instructions performedper machine cycle.

    Multiprocessing and Parallel Processing

    Despite the astounding evolution of computer technology over the past half

    century, the vast majority of desktop PCs are still driven by single CPUs. In therace to develop tomorrow's ever-faster computer systems, scientists havedeveloped multiprocessing and parallel processing. In contrast to techniques thatenable multiple operations to be performed simultaneously within a single CPU,multiprocessing and parallel processing use two or more coordinated CPUs toperform operations simultaneously.

    With multiprocessing, each CPU typically works on a different job. Becausemultiple jobs are being processed simultaneously, they are completed faster thanwith a single processor. With parallel processing, multiple processors worktogether to make one job finish sooner (see Figure 2-26 ). Two of the mostcommon designs are symmetric multiprocessing (SMP) and massively parallel

    processing (MPP). With SMP, a single copy of the operating system is in chargeof all the processors and the processors share memory. Typically, SMP systemsdo not exceed 64 processors. MPP systems, in contrast, can use hundreds orthousands of microprocessors and each processor has its own copy of theoperating system and its own memory. MPP systems are typically more difficultto program than SMP systems.

    Multiprocessing can increase astronomically the number of calculations performedin any given time period. For example, NEC's Earth Simulator (shown in Figure1-17 in Chapter 1) operates at approximately 40 teraflops. In other words, it canprocess about 40 trillion operations per second (most desktop PCs today operateslightly faster than 1 gigaflop).

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig01-17.jpghttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-26.bmphttp://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/Figures/Fig02-25.bmp
  • 8/10/2019 Tutorial EXL1 Lecture

    30/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 30

    When tremendous amounts of processing are required but access to asupercomputer is not feasible, one recent trend called grid computing uses thecomputing capabilities of a large collection of PCs to work together to process thenecessary data. For instance, large businesses can use their employees' desktoworkstations to process data overnight while the office is closed. Thesedistributed processing systems can be as powerful as a supercomputer.

    For example, one business that uses 5,000 engineering workstations after their

    employees go home at night has reached a combined computing speed of 6teraflops. Grid computing can also be performed over the Internet with homePCs. Usually a small program is downloaded and then, when the system is idle,the software puts the PC to work on a small piece of the current project. Often theprojects to which individuals donate their PC's processing time involve scientific ormedical research.

    Hyperthreading

    A technique being used with some CPUs today that is related to multiprocessingand parallel processing is hyperthreading. Hyperthreading, developed by Intel,

    enables software to treat a single processor as two processors. Since it util izesprocessing power in the chip that would otherwise go unused, this technology letsthe chip operate far more efficiently, resulting in faster processing, provided thesoftware being used supports hyperthreading.

    Improved Instruction Set Design

    An earlier discussion in this chapter mentioned that each computer has its owninstruction set. Traditionally, computers have been built under a design principlecalled complex instruction set computing (CISC). A computer chip with CISCdesign contains a relatively large instruction set. Starting in the mid-1970s,

    processors with instruction sets limited to only the most frequently usedinstructions became available. These devices follow a principle called reducedinstruction set computing (RISC). Experience has shown that computers usingthe RISC design work faster, unless the instruction is extraordinarily complex.

    With the introduction of 64-bit CPUs, such as Intel's Itanium chip and AMD'sOpteron, a new design optionexplicitly parallel instruction computing (EPIC)emerged. The primary philosophy behind EPIC is instruction-level parallelism(ILP). With ILP, individual operations are executed in parallel, significantlyspeeding up processing. ILP is made possible by increasing the number oftransistors on a chip to enable multiple operations to be performed at one time.

    Related technologies that help to expand the instruction set design for Intel CPUsare MMX (Multimedia Extensions), streaming SIMD extensions (SSEs), and thenewer streaming SIMD extensions 2 (SSE2). MMX is a set of 57 multimediainstructions for handling many common multimedia operations. SSE and SSE2help CPUs perform floating-point-intensive applications like video and audiohandling, 3-D modeling, and physical simulations much more quickly than before,provided the software being used supports the extensions being used.

    Future Trends

  • 8/10/2019 Tutorial EXL1 Lecture

    31/36

    17/7/2014 Tutorial EXL1 lecture

    http://www.mtsac.edu/~rpatters/CISB11/Chapters/Chapter_02/Chap02/LectureMain.htm#top 3

    While some of the strategies discussed in the prior sections are currently being used,some new ideas are further from being implemented. Selected trends we may see in thenear future are discussed next.

    Nonotechnology

    Today's process of forming integrated circuits (ICs) by imprinting patterns onsemiconductor materials -- called lithography -- allows for circuits smaller than onemicron (1,000 nanometers or one-millionth of a meter). A newer field of science --called nanotechnology-- involves working at the individual atomic and molecularlevels to create computer chips and other components that are thousands of timessmaller than conventional technologies permit (one nanometer is one-billionth of ameter).

    Researchers at several organizations -- such as IBM, Motorola, Dow Chemical,Xerox, and Hewlett-Packard -- are working to create nanometer-sized products andmaterials. Some prototypes include a miniscule device (about one-fiftieth the

    width of a human hair) that can measure small amounts of electric charge; a singleswitch that can be turned on and off like a transistor but is made out of a singleorganic molecule; and a miniature hard drive that holds 20 GB of data and is readby a device that is only a few nanometers thick.

    Using newly developed technologies, such as extreme ultraviolet (EUV) light,transistors as small as 15 nanometers have been created in lab settings. Currentnanotechnology consumer applications include stain-resistant fabrics and long-lifetennis balls. Possible future applications include improved military uniforms thatprotect against bullets and germ warfare, microscopic robots that can enter thebloodstream and perform tests or irradiate cancerous tumors, molecular circuitsthat can function similar to today's silicon-based circuits but that are much, much

    smaller and much more efficient, and memory and storage media significantlysmaller and denser than is currently possible.

    Nanotechnology is also expected to eventually lead to computers that are smallenough to be woven into the fibers of clothing or embedded into paint and othermaterials. In addition, nanotechnology may eventually solve much of the toxicwaste problem associated with e-trash (electronic waste, such as old computerequipment) by being able to rearrange dangerous components at the atomic levelto become inert substances. The federal government has had increased int