Location via proxy:   [ UP ]  
[Report a bug]   [Manage cookies]                
0% found this document useful (0 votes)
2 views

Computer Architecture

The document provides an overview of computer architecture, detailing key components such as the CPU, memory, and I/O systems, along with their functions and characteristics. It explains the Von Neumann architecture, digital logic gates, and various types of memory and storage, including RAM, ROM, and mass storage devices. Additionally, it covers I/O devices, interrupts, and polling methods used in CPU instruction cycles.

Uploaded by

hieuntd2410281
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
2 views

Computer Architecture

The document provides an overview of computer architecture, detailing key components such as the CPU, memory, and I/O systems, along with their functions and characteristics. It explains the Von Neumann architecture, digital logic gates, and various types of memory and storage, including RAM, ROM, and mass storage devices. Additionally, it covers I/O devices, interrupts, and polling methods used in CPU instruction cycles.

Uploaded by

hieuntd2410281
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Computer Architecture

Wednesday, December 11, 2024 11:12 PM

Some terms:
- PCI (Peripheral Component Interconnect): 32 and 64 bit versions, replaced by PCI express
- AGP (Accelerated Graphics Port): port for Video cards and 3d accelerators
- IDE (Integrated Drive Electronics): Standard E-interface between motherboard data path/bus and disk storage
- EIDE (Enhanced IDE): data rates from 4 - 16.6 MBps

CPU:
- Overview:
○ Does the processing, made of transistors, millions of transistors per chip organized into circuits in each chip, accommodates binary values
- Functions:
○ Four basic functions: Adding, Decoding, Shifting, Storing
○ Four corresponding transistor circuits:
▪ Adder: 4 arithmetic operations
□ Adds two binaries, output results
□ Inputs: two bits (x, y) for addition and a carry-in bit (ci)
□ Output: sum bit(s) and a carry-out bit (co)
▪ Decoder: react to specific bit patterns
□ Address memory, select I/O devices
□ Specific input pattern = output line selected
□ Circuit with n inputs can control 2 n lines
▪ Shifter: Move bits Left-Right
□ Support math operations like multiplication and division
□ Shifts input (bits) left to right and vice versa
▪ Flip-flop (latches): storage of bits
□ Special form of latch circuit
□ Holds value at output even if input change
□ Inputs: S (set), R (reset)
□ Outputs: Q and Q'
□ Used for high speed memory in CPU, Static RAM (SRAM) and cache
▪ Other Circuits:
□ Multiplexer, Parity Generator, Counter
□ Design process:
 Construct truth table for I/O
 Build Boolean expression for truth table
 Represent expression in circuit diagram
□ Integrated Circuits (ICs):
 Whole circuits etched onto single semiconductor chip
 VLSI - Millions of transistors making up CPU circuits

- Transistors:
○ Material: Silicon or Germanium
○ Parts (3): Emitter, Collector, Base
▪ Base turn currents on/off, translating to binary values
○ Size: ~130 nm side (Pentium IV)
- Digital Logic gates:
○ One level above transistors, performs operation on Boolean algebra
○ Boolean Algebra:
▪ Chief operators: AND, OR, NOT
□ AND: Both x and y input must be (1) for output to be one (x*y=z)
□ OR: One or both of x and y must be (1) for output to be 1 (x + y=z)
□ NOT: Output is the reverse of input (x' = z)
▪ Variables: 1 or 0
▪ Truth tables: Tabular representation of Boolean expressions (Leftmost columns represent possible combo of inputs, last column contain corresponding output (2n rows for n inputs)

- Gates:
○ Transistor circuits implement Boolean operators
○ Binary values maintained with constant voltage
○ 6 fundamental gates: AND, OR, NOT, NAND, XOR, NOR (last three are constructed from chief operators)
▪ NAND: Reverse output of AND
▪ NOR: Reverse output of OR
▪ XOR: Inputs with opposing values produce (1), but (0) when they are the same (0-0 or 1-1)
○ Output for any given combination of gates can be predicated on truth tables
○ Complex circuits can be constructed from gates, output of one can be input of another gate

Von Neumann Architecture:


- 3 main characteristics:
○ Four main sub-systems: Memory, ALU (Arithmetic/Logic Unit), Control Unit, I/O system
○ Program stored in memory during execution
○ Instructions are executed sequentially

- Subsystems:
○ Memory:
Architecture Page 1
○ Memory:
▪ RAM: made of many memory cells with addresses associated to each cell
□ All accesses to memory are to a specified address, cell is smallest unit of access, time to access/store same for all cells
□ Both data and programs are stored in memory
□ Memory width (W) and Address width (N) (address width = N-bits then address space = 2 N ie. 0,1,…,2N)
□ Typical capacity for PC: 64-256MB (1KB = 210 Bytes, 1MB = 220 Bytes, 1GB = 230 Bytes)
□ Access time: 50-75 ns (1*10-9s)
□ Volatile and relatively expensive
▪ Operations (2):
□ Fetch: copy of content in memory cell w/ specified address, non-destructive - copies value in memory cell
□ Store: Store Specified value into memory cell w/ specified address, destructive - overwrites previous value in cell
□ Memory system interfaced with:
 Memory Address Register (MAR)
 Memory Data Register (MDR)
 Fetch/Store signal
▪ Structure (2):
□ Fetch: Load address to MAR, decode address in MAR, copy content of memory cell with specified address into MDR
□ Store: Load address into MAR, Load value into MDR, Decode address in MAR, Copy content of MDR into cell w/ specified address

○ I/O:
▪ Handle devices that communicate with outside (keyboard, screen, …) and store info (Hard-drives, floppies, CDs, …)
▪ Mass storage device access methods:
□ Direct Access Storage Devices (DASDs): Hard-drives, floppies, CD-ROMs
□ Sequential Access Storage Devices (SASDs): Magnetic tapes
▪ Controllers:
□ Speed: Hard-drives: ~10ms
□ I/O controller:
 Small memory buffer, control logic for controlling the devices, sends interrupt signal to CPU when read/write complete
 Data transferred between RAM and memory buffer
 Processor free while I/O controller read/write data into buffer

○ ALU:
▪ Perform arithmetic and logic operations, integrated into today's CPUs
▪ Components (3):
□ Circuits (for operations): Array of circuits to do operations
□ Registers (for storage of computational data): Very fast memory cells, store operands and intermediate results, CCR (condition code register) stores "<,=,>" operations
□ Bus (connection path between first two): Data path interconnecting ALU registers and circuitry

○ Control Unit:
▪ Execute programs repeatedly:
□ Fetch next instruction from memory
□ Decode instruction
□ Execute instructions by issuing signals to ALU, memory and I/O controllers
□ Continue until HALT
▪ Machine Language Instructions:
□ Made of Operation code (which operation to perform) and Address Field(s) (memory addresses on which the operation works)
▪ Instruction Set Design (two approaches):
□ Reduced Instruction Set Computers (RISC): small & simple as possible, minimizes circuitry = faster computer
□ Complex instruction Set Computers (CISC): more and many complex instructions, can do more work but require more circuitry
▪ Typical instructions:
□ Notations: X, Y, Z for RAM cells, R for registry
□ Transfer: LOAD X, STORE X, MOVE X, Y
□ Arithmetic: ADD X, Y, Z (output = Z), ADD X, Y (output overwrites Y), ADD X (Sum of R and X), SUBTR, OR, …
□ Compare: COMPARE X, Y (Compare contents of X and Y cells and set condition codes)
□ Branch: JUMP X (load next instruction from memory), JUMPGT X (load next instruction from location X if GT flag in CCR set, otherwise load from same location)
□ Control: HALT - stops execution
▪ Structure of Control Unit:
□ PC (Program counter): Stores address of next instruction to fetch
□ IR (Instruction Register): Store instruction fetched from memory
□ Instruction Decoder: Decode instruction and activate circuitry
▪ Overview of operation:
□ PC is set to address of first instruction in memory
□ Repeat till HALT or fatal error:
 Fetch Instruction:
◊ PC put address in MAR
◊ Signal to fetch memory into MDR
◊ Add more value to IR
◊ Increase address to program counter
 Decode:
◊ Decode instruction in IR
◊ Decoder generate signal to circuitry to carry out instruction
 Execute:
◊ Differs between instruction

Buses:
- Set of wires and rules for data transfer, connecting system components
- 3 signal groups: Control, Address, Data
- Modern Standard: PCI

Peripheral Buses:
- Connect different I/O devices to computer
- Allow CPU to pass control to other devices

Storage:
- Primary and secondary storage

Memory:
- ROM: Etched in chip, can't be modified, BIOS is an example
- RAM: Allows direct memory reference, reading and writing, volatile memory, CPU fetch instructions from RAM
○ Types:
▪ DRAM: Circuits using one transistor per bit, needs constant refreshing to maintain data, 10-70ns in speed

Architecture Page 2
▪ DRAM: Circuits using one transistor per bit, needs constant refreshing to maintain data, 10-70ns in speed
▪ SRAM: Flip-flop circuits, fastest memory type, mainly use in registers and cache

Mass storage:
- Greater capacity and cheaper than ROM or RAM (per megabyte)
- Non-volatile but slower than Ram or ROM
- Hard-disks: most common, store data in magnetic platters than spin at 7200 RPM, low cost-unit storage compared to RAM, RAID (Redundant Array of Independent Discs)
- Optical storage: Examples: CDs, DVDs; stores data using optical tech - pits burned into disc represent binary data and is read while spun like a hard drive
- Flash drives: Portable, plug into USB port, replacement of floppy drives, non-volatile

I/O devices:
- Keyboards: keystrokes converted to binary signals transported through USB and then to CPU
- Mouse: Senses movement translated in to binary code
- Monitors:
○ CRTs: Use raster scanning techniques, Quality based on resolution and refresh rate
○ LCD: thinner, cooler than CRTs, use transistors instead of electron beams
- Printers: types: inkjet and laser; quality measured in Dots per Inch and Pages per Minute
- Sound cards: Fit into PCI expansion slot, Digitize sound for storage, Converts binary sound files to analogue

Interrupts and Polling:


- CPU instruction cycle = clock speed
- Processing determined by two methods: Polling and Interrupt Handling

Architecture Page 3

You might also like