Learning Objectives

• Understand:
• Digital data representation,
• The binary nature of digital physics, and
• The relationship between these two ideas
• Understand how data and programs are represented to a computer and be able to identify a few of the coding systems used to accomplish this.
• Explain the functions of the hardware components commonly found inside the system unit.
• Explain how systems can be expanded in order to attach new peripheral devices or add new capabilities.
• Describe how the computer system’s CPU and memory components process program instructions and data.
• Name and evaluate several strategies that can be used today for speeding up the operations of computers, and some strategies that may be used in the future.

• Computers today are mostly digital computers—devices that can only understand two states.
• Digital vs. Analog representation:

• The binary nature of digital physics:

 
• Plus, not illustrated above, but important: electric charge or "static electricity", which is
• either positive or negative
• different from current along a wire, above left
• the basis of dynamic RAM (the most common form of RAM <Random Access Memory>)
• Digital data representation = translation of numbers, text, pictures, sound etc. into digital ("1's & 0's") format
• The two states of a digital computer are typically represented by 0s and 1s.
• But they are really (physically) for instance:
• A positive or negative electric charge in a memory location;
• A magnetic north  pole or south pole somewhere on a magnetic disk;
• The presence or absence of electrical voltage on a data bus line (below); etc.
• I.e., what we think of as "1's and 0's" are really just some physical condition somewhere inside the computer or its peripherals.
• Natural (spoken/written) language input is translated by the computer into binary form in storage.
• And are then translated from binary back to natural language for output.
• "HI" input > 01001000 01001001 binary (using ASCII encoding) > "HI" output
  

• What is a byte? A  group of 8 bits, or 8 binary digits

 

• Or, each is 1024 x its neighbor below
  
• 1MB = 1024 KB
  
• 1GB = 1024 MB
• 1TB = 1024 GB
• 1PB = 1024 TB
  
• Byte Terminology
  
• Byte - a string series of eight bits
• Bit - Binary digit

• What is a "number"? A concept that expresses quantity or size: how much, how big
  
• The same quantity or size is expressed differently in different number systems.
  
• Decimal numbers vs. binary (and other) numbers: the difference is the number base
  
• The decimal number system has 10 symbols: 0, 1, 2, 3, 4, 5, 6, 7, 8, 9
• The "base 10" system
• That we use every day
• Why did we wind up using base 10? (as opposed for instance to 2, 8, 12, or 16?)
  
• The binary number system represents all numbers using just two symbols: 0 and 1
• The "base 2" system
• Binary numbers are used in computers because of the binary nature of digital physics
  
• A favorite test question: The binary number system uses the digits 1 and 2. True/False
  

• A little mathematics of number systems and powers, or exponents:
• A "power of" a number is the number times itself a given number of times
• A superscript digit following the number says how many times, i.e.
  
• 4² = 4 x 4 = 16,
  
• 4³ = 4 x 4 x 4 = 64
  
• Any number in any number system to the 0 power = 1, e.g., 10⁰ = 16⁰ = 1⁰  (=0⁰) = 1 (Ask Dr. Math)
  
• Some powers of 10:
  
• 10⁰ = 1
  
• 10¹ = 10,
  
• 10² = 10 x 10 = 100;
  
• 10³ = 10 x 10 x10 =1000, etc.
  
• right?
• Decimal or "base 10" numbers
  
• are the ones we use every day
• with the digits 0-9 (i.e. 10 is not a decimal digit)
• The computer also shows and prints decimal numbers when we expect a "number"
  
• Example:
  

• Decimal or 7216 = 10³ x 7 + 10² x 2 + 10¹ x1 + 10⁰ x 6

• Binary or "base 2" numbers are the ones the computer uses
• They work the same way as decimal numbers but with a base of2 instated of 10
  
• Powers of 2:
  
• 2⁰ = 1
• 2¹ = 2
• 2² = 2 x 2 = 4,
  
• 2³ = 2 x 2 x 2 = 8,
• right?
• The rightmost digit indicates how many 2⁰'s (1's) there are in the number
• The next rightmost digit indicates how many 2¹'s (decimal 2's) there are in the number
• The next rightmost digit indicates how many 2²'s (decimal 4's) there are in the number
• The next rightmost digit indicates how many 2³'s (decimal 8's) there are in the number
• etc.
• I.e., the tenth from rightmost digit indicates how many 2⁹'s ( decimal 512's) there are in the number
• So, binary 1001 = 2³ x 1 + 2² x 0 + 2¹ x 0 + 2⁰
  

• The above applies only to numbers: the things we think of expressing size or how much of something.
  
• What about spoken and printed language, pictures,  music and video?
• The computer has only "1's and 0's" (again really just electric, magnetic, etc. impulses) to work with.
  
• Remember? "Digital Data Representation"
  
• We need ways to bend 1's and 0's toward representing these other kinds of things.
  

• Number riddle:
  
• Why do programmers get Halloween and Christmas mixed up?
• Because OCT 31 = DEC 25.
  

Coding Systems for Text-Based Data

• What is a letter?
• A letter is a picture (or or visual pattern or "glyph") that represents a certain speech sound.
• Does the same pattern represent our "D" sound the world around? No; for instance
  
• The Greek letter delta: Δ
• The Cyrillic letter de:    Д
• In The Land of Ones and Zeros (cyberspace)
  
• There are no pictures or patterns or glyphs of letters.
  
• Just 1's & 0's
• Patterns of 1's and 0's have to represent everything in cyberspace.
  
• So the Roman alphabet letter D is represented by the (ASCII) pattern 01000100 (see fig. 2.4, p. 60)
  
• So how does the computer know if 01000010 represents a B or the binary number equal to decimal  68?
  
• By the context.
• 01000100 might also represent a bit of music, graphic, animation, or program
  

• The three main encoding schemes for letters, numbers, punctuation, etc. are:
• ASCII    (American Standard Code for  Information Interchange) Used on all IBM PC's & all other computers
• EBCDIC (Extended Binary Coded decimal Interchange Code) Used on IBM Mini's and Mainframes
• ASCII and EBCDIC are fixed-length codes that can represent any single character of data as a string series of eight bits
• "string" in computer science usually means string or series of characters, like a word or paragraph.
  

• Unicode: A 32-bit code (i.e., up to 4 bytes per character) that can be used to represent text-based data in virtually any written language.
• UTF-8 (Unicode Transformation Format) encoding allows the ASCII characters to be represented in 8 bits within Unicode.
• (Δ = Unicode 0000001110010100 <at www.unicode.org/charts>)
  
• Think of them as different alphabets, like English, Greek and КИРИЛЛИЦА (Cyrillic)

• Self quiz:
  
• ASCII 01000100 = what character? 
  
• The character "F" in EBCDIC = binary what?  (Figure 2-4, p. 60)

• Graphics data

• Pixel: abbreviation for picture element
  
• More familiar to us in the age of digital cameras
  
• Monochrome v. color graphics
• Simplest monochrome: a pattern or pure black and pure white pixels
• Grayscale: pixels in shades of gray varying from pure white to grays to pure black
  
• Two general methods of binary encoding for graphics
• Bitmaps (e.g. .bmp .jpg .gif .tiff .png
  
• monochrome bitmaps use one bit of storage per pixel
• 0 = dark, 1= bright (or vice versa)
• color bitmaps use several bits per pixel (below)
  
• Vector graphics (e.g. .ai, .eps, .wmf)
• represent an image using a set of directions for drawing and coloring it
• not pixel by pixel as in a bitmap
• Color depths: 16, 256 color images, and "true" color.
• color bitmaps use >1 bit per pixel
• 4 bits per pixel for 16 color
• 8 bits per pixel for 256 color
  
• 24 bits per pixel for "true" color (16,777,216 colors: why 16,777,216?)
• Dithering
  

• Audio data
• CD style storage (.wav)

     110 110 110 111 111 101 100 011 001...

• MP3 (MPEG-1 Audio Layer 3) audio encoding/compression

• Video data
• Frames & Frame Rates
• Video DVD's: 2 hours of video= ~66g compressed to 4.7g

• Lossy (e.g. .jpg .mp3) and lossless (e.g., .zip .jar)  compression
  

• A few Windows XP(?) Items you may or may not find useful
• Viewing file extensions, change dates, sizes in Windows Explorer:
  
• Click View | Details  (or use the Views Toolbar View Button)
  
• To apply the current view to all folders: Tools | Folder Options | View and click Apply to All Folders 
• (Earnestly hoped for XP enhancement: Allow Apply Only to this folder and Its Subfolders)
• Sorting the list: In the Detail View Right Pane,
  
• click on column heading to sort by: name, date, size, extension, etc.
• click again to sort in the opposite order
• most windows application column displays work similarly, i.e., Outlook Express
  
• sorting in reverse date order puts your most recently changes files at the top.
  
• Unhiding "known file extensions" in Windows Explorer: Tools | Folder Options | View and uncheck Hide extensions for known file types
• While we're at it, Snap To: Start | Control Panel | Mouse | Pointer Options and check Snap To
• Automatically moves your mouse cursor to the "default" command button: OK, Send, Save, etc.
• "Default" is the assumed command or value unless you say otherwise.
• Usually the one you are most likely to choose.
  
• Also: the Search and Folders Tool Buttons
• Also: the Views tool
  

• Machine language is the binary code used to represent program instructions.
• The basic set of machine-language instructions that a CPU can understand is that CPU’s instruction set.
• Most programmers rely on language translators to translate their programs into machine language for them.

I.e., few of us program in machine language these days

• The system unit of a PC is the case that houses processing hardware and other hardware.
• All of the hardware contained within the system unit is connected to the system board or motherboard

• Central Processing Unit = the CPU chip, the supporting "chip set" and RAM
• The microprocessor (CPU chip) contains a variety of circuitry and components and is connected to the motherboard; can contain many millions of transistors (electronic switches)

• Some styles of CPU chips
• Intel/AMD ("wintel")  (Fig 2-8 p. 67 shows generations of wintel chips)
• Intel Pentium, Core 2 Duo, Core 2 Quad, Celeron
  
• AMD Athlon, Athlon 64 X2
  
• Motorola/IBM/Apple PowerPC
• Transmeta Crusoe
• DEC Alpha
• Sun Sparc
• ARM (the most used CPU in the world, 10 billion sold)
  

• Single, dual, quad and more core CPU chips
• Multiple cores solve heat problems
  
• Processing Processor clock speed is measured in megahertz (MHz) or gigahertz (GHz).
• A computer word is a group of bits or bytes that may be manipulated and stored as a unit.
• Larger words support higher CPU throughput.
• Most microprocessors today have a 32 bit word; starting to see 64 bit
• Other factors that affect the speed of the computer include
  
• Amount and speed of RAM,
• Memory vs. storage
• Memory is directly accessible to the CPU: RAM, ROM, registers (Main or Primary Memory)
  
• Storage storage on a peripheral device: disk, DVD,  (Secondary or Auxillary Storage)
• Memory is volatile; storage is not.
  
• Amount and speed of cache memory,
• Width and speed of CPU and system buses affect their bandwidth (throughput)
• More below.
  

• Moore's Law: predicts CPU chip transistor density (and clock speed) growth (p. 81)

• 29,000 transistors on the 8086 (in the 1980 IBM PC) to 592,000,000 on the Itanium = a 20,000 times increase
• Moore's Law also applies to to CPU speeds: 4.77MHz in 1980; now in 2-4GHz range = a 1000 times increase
  

 

• RAM abbreviates Random Access Memory
• "RAM" is understood to mean Main Memory (= Primary Storage) even though there are other types of random access memory (below).
  
• RAM is volatile
• Two access methods
• Random access: the speed of access to the next byte is independent of location of the last byte accessed, e.g., memory & disk storage.
• Sequential Access: the speed of access to the next byte depends on the location last byte accessed, e.g., cassette or other magnetic tape.
• Types of RAM
• SDRAM: synchronous dynamic ram, the most common
  
• SRAM: faster, used on L1-L3 Cache
  
• SIMM's, DIMM's, RIMM's: Single, Double and Rambus Inline memory modules that contain SDRAM.
  

• Current Memory types, at How stuff Works
• Memory speeds (Hz or Ns)
• Memory Addressing (actually, all storage addressing):

• But... "all storage is linear..."
• i.e., addresses are not laid out in blocks or rows; can be thought of as one long row of bytes
  
• CPU word size determines (limits) size of memory address space
• A 32 bit word allows the CPU to address 2³² bytes of memory, or 4 gigabytes

Memory Classifications

• The Storage Pyramid
• RAM conventionally refers to main memory (above)
  
• Registers:
• the fastest memory
• on the CPU chip
• very limited storage capacity
• Cache memory
• L1 Cache: On the CPU chip
• L2 Cache: Separate chip on the motherboard
• Memory manager
• Goal: to have the next needed bytes in L1 cache
• By moving needed bytes form RAM to L2, and L2 to L1
  
• But if there's L3 cache, both L1 & L2 are in the CPU chip, and L3 is on the motherboard
  
• CMOS: holds setup options for the computer (e.g., boot device sequence)
• Complementary Metal Oxide Semiconductor
  
• ROM: holds the BIOS (Basic Input Output System), also referred to as Firmware
• Read Only Memory
  
• POST program
• subprograms for accessing that machine's hardware
• Flash memory

• For BIOS storage: can be updated with new BIOS programs
• Also used for secondary storage (Chapter 3).
• MRAM (Magnetoresistive RAM): nonvolatile RAM using magnetism rather than electric charge to store data
  
• PRAM (Phase change RAM): changes phase of a material to store data
  
• What is a "phase"? For instance, the 3 phases of water are ice, liquid and vapor.
  
• All of which are random access memory

• 2 fans cool the power supply and the CPU
• Heat Sink: a finned block of usually aluminum in contact with the CPU chip to help radiate excess heat
• Liquid cooling: circulates cool water around the CPU, like a car's cooling system
• Or how about liquid nitrogen CPU cooling?
  
• Ion Pump Cooling: uses an electronic effect to move cooling air across components

• Buses are electronic paths that data travels around on a computer system

• Internal buses move data around within the CPU.
• System Buses move data around on the motherboard and to/from peripheral devices
  

• Expansion cards for desktop PCs
• Fit in expansion slots (see system unit graphic above)
  
• ExpressCards: expansion for notebook & other portables

• Expansion for handhelds and mobile devices: USB port, MM cards, and Springboard modules, etc.

• Ports:

• Six of the twelve kinds of port mentioned on pp. 77-8
• Parallel (Printer port or Centronics port)
• Serial   (RS-232 port; usually was for external modems and mice)
• SCSI    (Small Computer systems Interface; printers, scanners, also external disk and CD)
• USB     (above)  and the USB hub
• Most PC's come with "enough" USB ports, front and back
• USB is replacing most parallel and serial port connections today
• Firewire/IEEE1394
• Network (usually Ethernet)
  

• Plug specifics:
• Telephone plug/jack has 4 wires/connectors (RJ-11)
• Network plug/jack has 8 wires/connectors  (RJ-45)

• Front ports: for your camera, memory cards, etc.

• Notepad ports
  

• Handheld ports (above)
• Hot swapping peripheral devices
  

Sorry, I could not resist:

• ALU and FPU : carry out arithmetic and logical operations
  
• Control Unit: coordinates the other units' activities
  
• Prefetch Unit: gets instructions from cache or main memory
  
• Decode Unit: decodes the instruction for the ALU
  
• Registers: fast memory for ALU operations
  
• Internal Cache: L1, maybe L2, storing we hope the next instructions to be carried out.
  
• Bus Interface: connects to the system bus 
  

• Evolution of CPU Speed
  
• 1980: 4.77MHz
  
• 2009: ~4GHz
• A 800x speed increase

• The machine cycle is the series of operations involved in the execution of a single, machine-level instruction.
• Machine Cycle broken down:
• Instruction stage
• Fetch instruction
• Decode
• Execution stage
• Execute instruction
• Store result
• Example:
  

• The CPU instruction set=the CPU's machine language
• RISC and CISC (Reduced and Complex Instruction Set Computing)
• Clock Speed v. Processing Speed
• Clock Speed: measured in Hertz
• 1 Hertz = 1 cycle per second
• 1 gigahertz = 1 billion cycles per second
• Flops (floating point instructions per second)
• 1 gigaflops = 1 billion floating point instructions per second
• 1 teraflops = 1 trillion floating point operations per second
• 1 petaflops = 1 quadrillion floating point operations per second
• each one of which may take one or several clock cycles to carry out
• Other performance measures

Linpack	for scientific and mathematical processing
Ramp-C	for commercial transaction processing

• More on this jargon below
  
 

• Prefixes to know:

 

Multipliers	Multiple
Kilo-	thousand	x1,000	x103
Mega-	million	x1,000,000	x106
Giga-	billion	x1,000,000,000	x109
Tera-	trillion	x1,000,000,000,000	x1012
Peta-	quadrillion	x1,000,000,000,000,000	x1015
Divisors	Fraction
Milli-	thousandth	/1,000	x10-3
Micro-	millionth	/1,000,000	x10-6
Nano-	billionth	/1,000,000,000	x10-9
Pico-	trillionth	/1,000,000,000,000	x10-12

• If you memorize the two sets of prefixes in order, you can figure them out by
• Starting with x1,000 at kilo- and with /1,000 at milli-
• Increasing the multiplier or divisor by adding 3 zeros at each step.
• That's how I do it anyway
• Examples:
• Typical PC's sold today have ~500 megabytes (mB) of RAM and ~40=-100 gigabytes (gB) of disk.
• 4 terabytes (tB) of storage can be addressed by a 32 bit processor.
• A picosecond is about the shortest period of time we can currently measure accurately.
• Gigabit Ethernet communicates at 1 GB/second
  
• (Exa-, Zetta-, Yotta- I recall reading somewhere that all recorded knowledge on earth is on the order of an exabyte.)
• Just about everyone knows about "nano" now...
  
  
  
  • A nanosecond is...?
  
  

Speeding Up Your System Today

• Add more memory
• Remove malware (or badware): adware, spyware, viruses, etc.)
  
• Perform system maintenance.
• Uninstall unused programs?
• Run Windows Disk Cleanup? (But do this if you are low on disk space.)
  
• Archive large files?
• Clear out Recycle Bin (Trash)?
• Delete temporary files?
• Defragment ("defrag")?
• How do these make the computer faster?
• Buy a larger(?) or second(!) hard drive.
• Upgrade your Internet connection.
• Upgrade your video card.
• Upgrade your CPU.

• Any one of the above may have disappointing results if it was not in fact a throughput bottleneck in your system.
• The psychology of speed: speed gains of <+100% are often disappointing anyway.

• Get a new system.

Strategies for Making Computers Faster and Better

• Improved architecture
  
• Moving circuits closer together
  
• Making components smaller: nanotechnology
  
• Increasing register size/word size; 64 bit and 128 bit computing
• Faster memory
  
• Faster and wider buses
• Improved materials: copper on silicon, gallium arsenide
• Improved instruction set design
• Pipelining:

• Multiple CPU's
  
• Parallel Processing: multiple CPU's, each carrying out the same instruction simultaneously on a different bundle of data
  
• "vector processing"
  
• Massively parallel processing: supercomputers with thousands of CPU chips
  

• Multiprocessing
  
• Symmetric Multiprocessing (SMP)  or n-way CPU's: mainframes and servers
• Dual Core and Quad Core CPU chips now on PC's
  

• Hyperthreading:  Introduction to Multithreading, Superthreading and Hyperthreading
• 64 bit and larger word CPU's
• Clustering
• Grid Computing
  

Future Trends

• Organic computers
• Biotechnology
• Nanotechnology: creating circuits and machines of microscopic, molecular, or atomic size.
• New materials: Copper/germanium
• Optical processing
• Silicon Photonics
• Terascale computing
  
• Superconductive materials
• Quantum computers

• 3D chips: mutilayered circuitry

---

Octal (base 8) 31 = Decimal 25
Programmers use octal numbers and a lot of hexadecimal (base 16)  numbers too.







.