V.I.P. Dasanayake_computer System

8/6/2019 V.I.P. Dasanayake_computer System

http://slidepdf.com/reader/full/vip-dasanayakecomputer-system 1/12

CS 2842-Computer Systems

ASSIGNMENT 02

Name: V.I.P. Dasanayake

Index Number: 090075M

Field: Electrical engineering

1



(1) Memory Hierarchy

This is mainly base on performances and cost of various kinds of memories such as

flash memory register, RAM, hard disks...etc. Memory hierarchy means put those

sections such as RAM, registers, hard drives …etc in a Oder based on performances and

prices, Other than that physical size, fast.

The principle of memory usage is

To optimize memory performances for a given cost set of technologies are arranged

in a hierarchy that contains a relatively small amount of fast memory and larger

amount of less expensive, but slower memory.

Data transfer speed in hard drive and registers are very different so they can’t

exchange data. So there should be some memories which speed is close to both speeds.

For that purpose there are some memories which supports to that. And the other thing is

when it is want to show some huge data (which have high memory capacity) it can’t be

shown due to the huge memory. Computer may take long time to load that data, so to

minimize that time there are several memories. These memories should have to put in

order. Memory hierarchy concept helps for that. So data transfer is happening b the

combination of all these memories. Here are some memories which are in the order of

the concept of Memory Hierarchy.

2

Cache

RAM

Register

Tape backup

Hard drive

Flash

..Price

..Fast

..Cost

ly

..Vola

til e

..S

mall

size



(2)i. Latency

Latency is the time that elapses between the start of an operation and the completion

(the delay between the time a bit is sent and the time the bit is received). This is usuallymeasured by nanoseconds (ns). This is a measure of the time required to perform a

transfer .

Latency doesn’t sufficient as a measure of memory performance. Latency is a

simplistic measure that doesn’t provide complete information.

ii. Memory Controller

There is a speed difference between the memory and the processor/CPU. So memory

controller supports to communication happens between processor and memory. While the

data transfers to processor from memory controller gets data from memory. So the datawhich receives from the controller is not fresh because data transfers to the processor

from controller are happening gradually.

Hardware use for memory access

To access memory a processor presents a read or write request to the controller.

Controller translates the memory address request to signals and passes to memory chips.

iii. Memory bus

This is a parallel interface which decides memory transfer size. Memory controller

provides the connection between a physical memory and a processer that uses the

memory. To achieve high performances, memory systems use parallelism. The processor

and the controller connected parallel with several wires and those wires can transfer one

data bit at any time. Those wires are called Memory bus.Control circuits in the memory use the address to select the bits at the specified

location in RAM and send a copy of the data back to the CPU over the memory bus.

(Technical name for the hardware connection between a processor and memory is

memory bus)

This can be shown as below.

Memory bus ( wires )

3

ControllerPhysical memory

Processor

ProcessorPhysical

memoryController



iv. Little-endian

When sending a byte of data over a network, both sending and receiving data should

be equaled as they will be transferred most significant bit or least significant bit first.

When a system that numbers bytes of an integer from least significant to most significant,we call it as little endian system.

Example:-

Number one expressed as thirty two bit

3 2 1 0

Little endian

(Least significant byte is zero)

v. Virtual Memory

Virtual memory means a mechanism that hides the details of the underlying physical

memory to provide a more suitable memory environment. We can simplify this as when

controller wants new data which aren’t include in the RAM, RAM replace some old data

and give the data to controller.

Illustration of a virtual memory system that an address space two physical memories

4

0000

0000

0000

0001

Memory 2

Memory 1Processor seesa singlecontiguousmemory

Address

0

999 1000

1999



(3)

i. Random Access Memory (RAM)

The term RAM is generally understood to refer to volatile memory, which can be

written as well as read. Mainly there are two types of RAM as Static RAM ( SRAM ) and

Dynamic RAM ( DRAM ).Those are alternatives to each other.

SRAM:-

This is the Easiest type for programmer because it is a straight forward extension of

digital logic. SRAM consists of transistors. Though this has a high speed, power

consumption and heat generating of this is relatively very high. Because when multiple

transistors operate continuously each transistor consumes power and generates heat in

small amount.

Miniature SRAM circuit that stores one data bit

DRAM:-

DRAM consumes less power because it consists of capacitors. When a value is

written to DRAM the hardware charges or discharges the capacitor to reflect the digital

value. Computer that uses DRAM contains an extra hardware circuit called a refresh

circuit which can perform fast writing and reading a bit. External refresh circuit

periodically read data and writes it back again or charge will dissipate and the value will

be lost.

When consider all the things most computer memories are composed of DRAM rather

than SRAM

A bit in a DRAM

5

Refresh

Circuit for

one bit

Writeenable

Output

Input

Circuit for

one bit

Writeenable

Output

Input



ii. Magnetic disks

Floppy and hard disk drives store information on magnetic disks. The disk is a thin,

flexible piece of plastic with little magnetic particles inlayed in its surface. The read-write

head creates a small magnetic field that aligns the magnetic poles of the particles on thesurface of the disk directly beneath the head to write data to the disk. While particles

aligned in one direction represent a “0”, particles aligned in the opposite direction

represent a “1”.The drive head scans the surface of the disk to read data from a disk. The

magnetic fields of the particles in the disk induce an alternating electric current in the

read-write head, which is then translated into the series of “1”s and “0”s that.

The main components of a disk drive are

a) The motor , which rotates the disk,

b) The read-write mechanism,

c) The logic board , (which receives commands from the operating, system to place or

retrieve information on the disk)

Floppy and hard drives use a small magnetic head to magnetize portions of the disk surface, to read or write information to a magnetic disk

iii. Tape drive

Tape drive is a data storage which can read and write data on a magnetic tape.

A tape drive provides random access storage. Tape drive can stream data from to tape

very quickly, because it spends much time winding tape between reels to read and

write any piece of data.

Now in market there are several tape drives which can store data by

compressing it to half of its original capacity. So if there has been mentioned 80/160,

it means that its true storage is 80 and it can store nearly 160 by compressing data to

80. Companies like IBM, Sony use higher compression ratio for their tape drives.

Compression ratio is depends on what type of data is being compressed. The true

storage capacity is known as the native capacity or the raw capacity.

Tape drive can be connected to computer using SCSI , Fiber channel , SATA,

USB. Earlier tape drive is designs as an alternative for disk drives and those are little

bit cheap than disk drives.

iv. Registers

Information is stored in a CPU memory location called a register. Registers can be

thought as the CPU’s tiny pad, temporarily storing instructions or data. Register capacity

is the total number of bits or bytes that a register in a microprocessor can store. Register isa volatile memory. There are two types of registers called program counter and

instruction register.

Program counter:-

When a program is running, register called the program counter keeps track of which

program instruction comes next by maintaining the memory location of the next program

instruction to be executed. The CPU’s control unit coordinates and times the CPU’s

functions, and it uses the program counter to locate and retrieve the next instruction from

memory. In other words we can say that it stores the address of instructions which are

coming through address bus.

Instruction register:-

The instruction travels along the memory bus from the computer’s memory to theCPU, is stores in instruction register

6



(4)Von Neumann Architecture

When a computer follows Von Neumann Architecture it uses the stored program

approach in which program resides in memory. The hardware that follows Von NeumannArchitecture consists of three principle components that interact,

1. Processor,

2. Memory,

3. Input Output facilities (I/O facilities)

Von Neumann Architecture concept can be shown as below.

Computer

(5)

I.

Fixed logic processors selectable logic processors

Can Perform only single

operation

Can’t change the function

Parameters of the function can’t

be changed

Less flexible

Multiple operations

Function is specified when the

processor is invoked


Parameters of the function can’t be

changed

More flexible

7

Processor

Input Output Facilities

Memory



II.

Fixed logic processors Parameterized logic processors.

Can Perform only singleoperation


Parameters of the function can’t

be changed

Less flexible

Performs pre determined operation Accepts a set of parameters that

control the operation

Mostly like fixed logic processors

Parameters of the function can be

changed

More flexible

(6)

Structure of the CPU:-

Information from an input device or from the computer’s memory is communicated

via the bus to the central processing unit (CPU), which is the part of the computer that

translates commands and runs programs. The CPU is a microprocessor chip that is, a single

piece of silicon containing millions of tiny, wired electrical components.

There are five major component of the CPU.

1. Controller

2. Computational Engine (ALU)

3. Local Data Storage

4. Internal Interconnections

5. External Interface

1. Controller

Controller is known as the heart of the computer because it Coordinates

operations. As it is mentioned earlier memory controller supports to communication

happens between processor and memory. Though there are four other structures of CPU, this is the most important thing.

2. Computational Engine (ALU)

An ALU consists of a single, complex hardware unit that can perform various

kinds of operations. (It provides operation on integer arithmetic, operation on bits

such as left or right shift. And hardware that performs logical operation such as

Boolean and, or, exclusive or and not) when ALU perform some operation such as

addition, subtraction, and bit shifting, remain functional units which supports to the

operation provide mechanisms that interface to the ALU to the rest of the system. The

computational engine doesn’t perform multiple or initiate activities. The engine only

operates at the request of the controller.

8



3. Local Data Storage

Mainly it is about hardware registers. As it is mentioned before information is stored

in a CPU memory location called a register. Register is a volatile memory. In here it is

dealing with memory size bites. This is a most important part of memory in the computer.

There are several register as instruction register, program counter. This registers get andsupply data to perform some operation and show.

4. Internal Interconnections

This is called data path also. The instruction code that tells the computer which

operation to perform also specifies which memory address or addresses to access. An address

is sent through wires called an address bus from the CPU to the main memory (RAM).

Control circuits in the memory use the address to select the bits at the specified location in

RAM and wires called a data bus, send a copy of the data back to the CPU. These address bus

and data bus are internal interconnections.

Internal interconnection is used to move data values from local storage to the

computational engine or move results from the computational engine to local storage. Internalinterconnections are in the CPU.

5. External Interface

This unit handles all communication between processor and the rest of the computer

system. Especially it is happening between the process and the external memory. While

there are several connections between external interface and ALU, controller, local storage,

there is also an external connection.

9



1100 01 10

(7)

Memory AccessingHere it can be use to write to the memory or read from the memory .

MBR:-Memory Buffer Register (MBR) is also called as Memory Data Register (MDR). All

the data in the MBR are situated according to an order. There are addresses in MAR in

HIGH, LOW manner. As Left number in MDR represents the HIGH part of MAR, right

number in MDR represents the LOW part.

MAR:-

MAR means Memory Address Register. To access a certain memory address for

MAR the controller give the address to MAR then is divided to low and high. As an example

when left bites of the address becomes the HIGH part of MAR, right part becomes LOW part

of MAR

The identifying of the address is can be explained as follows.

Example:-address=1101

The address = 10 | 01

The address

10

00

01 10 11

HIGH LOW

0, 0 0, 1 0, 2 0, 3

1, 0 1, 1 1, 2 1, 3

2, 0 2, 1 2, 2 2, 33, 0 3, 1 3, 2 3, 3

MDR



(8)Fetch –Execute cycle

Though the details among programmable processors are varying, all those processors

are following this fetch execution cycle.

The main idea of this is

When there is a stored point in a program, after accessing the next step of

The program from that point, it will be executed to perform steps of the program. This

whole thing will repeat forever

There are two main parts of this as Fetch cycle and Execute cycle. In fetch cycle there

are 3 steps and in Execute cycle there are two steps.

• In Fetch cycle

1) Fetch instruction

2) Decode opcode

3) Fetch operands

•

In Execute cycle4) Perform operation

5) Store result

In the Von Neumann Architecture the loop which is described earlier will continue

among processor, memory, Input Output facilities. The loop can briefly describe by following

picture.

11

Cycle happen of the Von Neumann Architecture



(9)There are two types of registers

IR – Instruction Register

PC – Program Counter (IP – Instruction Pointer)

While the Program counter stores the address of the instructions, Instruction Register

stores the instruction. Execution Flow Control happens using the value on the PC.

Program counter:-

When a program is running program counter keeps the track of which program

instruction comes next by maintaining the memory location of the next program

instruction to be executed. Address bus give these instruction addresses to the program

counter

Instruction register:-

While this program counting happening instruction coming along memory bus stored

in the instruction register.

Controller is the logical thing CPU (ALU is just doing some calculation only)

in CPU. When busses travel addresses and instruction to PC and IR, they store it. Whilethis happening memory controller supports to communication happens between processor

and memory. So when we consider execution of a program

Execution of a program with the control of the controller

12

ALU

Store theinstructions

Store the

address of

the

instruction

V.I.P. Dasanayake_computer System

Documents