Unit 4: Memory

MEMORY

Memory is a storage unit of a computer system used to store:

Data
Instructions
Intermediate results
Final output

It allows the CPU to access information quickly during program execution.

Characteristics of Memory

Capacity – amount of data it can store
Access time – time to read/write data
Cost – price per bit
Volatility – whether data is lost on power off

Memory Hierarchy

Concept of Memory Hierarchy

Memory hierarchy is an arrangement of different memory types in a system based on:

Speed
Cost
Storage capacity

Goal: To provide the CPU with fast access to data at minimum cost.

Levels of Memory Hierarchy

Level	Memory Type	Speed	Cost	Capacity
1	Registers	Very High	Very High	Very Small
2	Cache Memory	High	High	Small
3	Main Memory (RAM)	Medium	Medium	Medium
4	Secondary Memory	Low	Low	Very Large

Locality of Reference

Memory hierarchy works on:

Temporal locality – recently used data is likely to be reused
Spatial locality – nearby memory locations are likely to be accessed

Semiconductor RAM Memories

RAM (Random Access Memory)

RAM is a read/write memory where any location can be accessed in the same amount of time.

Types of Semiconductor RAM

(a) Static RAM (SRAM)

Uses flip-flops
No refresh required
Very fast
Low density
Expensive
Used in cache memory

(b) Dynamic RAM (DRAM)

Uses capacitors
Requires periodic refresh
Slower than SRAM
High density
Cheaper
Used as main memory

SRAM vs DRAM (Important for Exams)

Feature	SRAM	DRAM
Speed	Very fast	Slower
Refresh	Not required	Required
Cost	High	Low
Density	Low	High
Usage	Cache	Main memory

Memory Organization

Two-Dimensional (2D) Memory Organization

In 2D memory organization, memory cells are arranged in:

Rows
Columns

Components:

Row decoder
Column decoder

Working:

Address is divided into row and column parts
Row decoder selects a row
Column decoder selects a column

Advantage:

Fewer decoder lines
Efficient for medium-size memories

Two and Half Dimensional (2½D) Memory Organization

In 2½D memory organization, memory is divided into:

Blocks (or modules)
Each block has its own 2D structure

Address Division:

Block selection bits
Row selection bits
Column selection bits

Advantages:

Reduced decoder complexity
Better scalability
Used for large memory systems

Comparison: 2D vs 2½D

Feature	2D Organization	2½D Organization
Structure	Rows & Columns	Blocks + Rows + Columns
Decoder complexity	Medium	Low
Scalability	Limited	High
Usage	Small/medium memory	Large memory

ROM Memories

ROM (Read Only Memory)

ROM is a non-volatile memory used to store permanent data.
Data is retained even when power is off.

Types of ROM

Type	Description
Mask ROM	Programmed during manufacturing
PROM	Programmable only once
EPROM	Erased using UV light
EEPROM	Electrically erasable
Flash ROM	Block-wise erase, faster

Uses of ROM

BIOS programs
Booting instructions
Firmware in embedded systems

Conclusion (Exam-Ready)

Memory stores data and instructions required by the CPU.
Memory hierarchy balances speed, cost, and capacity.
Semiconductor RAM includes SRAM and DRAM.
2D and 2½D memory organizations improve addressing efficiency.
ROM provides permanent storage for system programs.

CACHE MEMORY

Cache memory is a small, high-speed semiconductor memory placed between:

CPU and
Main Memory (RAM)

Purpose of Cache Memory

Reduce average memory access time
Improve CPU performance
Bridge the speed gap between CPU and main memory

Principle Used

Cache memory works on the principle of locality of reference:

Temporal locality – recently accessed data is likely to be accessed again
Spatial locality – nearby memory locations are likely to be accessed soon

Cache Design Issues

While designing cache memory, the following important issues are considered:

Cache Size

Larger cache → higher hit ratio
But increases cost and access time

Mapping Technique

Determines how main memory blocks are placed in cache
Direct, Associative, or Set-Associative

Replacement Policy

Decides which block to remove when cache is full

Write Policy

Write-through
Write-back

Cache Organization

Unified cache
Split cache (Instruction & Data cache)

Cache Performance

Cache Hit and Miss

Cache Hit: Data found in cache
Cache Miss: Data not found in cache

Hit Ratio

H i t R a t i o = \frac{N u m b e r o f C a c h e H i t s}{T o t a l M e m o r y A c c e s s e s​}

Miss Penalty

Time required to fetch data from main memory when a miss occurs.

Effective Memory Access Time (EMAT)

E M A T = H i t T i m e + (M i s s R a t e \times M i s s P e n a l t y)

This formula is very important for exams.

Cache Address Mapping Techniques

Address mapping defines how a main memory block is mapped to a cache line.

Direct Mapping

Each memory block maps to only one cache line
Simple and fast

Address Format:


| Tag | Line | Word |

Advantages:

Simple hardware
Fast access

Disadvantages:

High conflict misses

Fully Associative Mapping

A block can be placed in any cache line
Uses associative memory (CAM)

Address Format:


| Tag | Word |

Advantages:

Lowest miss rate

Disadvantages:

Expensive hardware
Slower than direct mapping

Set-Associative Mapping

Cache divided into sets
Block maps to a specific set but can go to any line within that set

Address Format:


| Tag | Set | Word |

Advantages:

Balance between cost and performance

Disadvantages:

Slightly complex

Mapping Comparison (Very Important)

Feature	Direct	Fully Associative	Set-Associative
Hardware cost	Low	High	Medium
Speed	Very fast	Slower	Medium
Miss rate	High	Low	Medium
Complexity	Low	High	Medium

Cache Replacement Policies

Replacement policy decides which cache block is replaced during a cache miss.

Least Recently Used (LRU)

Replaces the block not used for the longest time
Best performance
High overhead

First In First Out (FIFO)

Oldest block is replaced
Simple implementation
Not always optimal

Random Replacement

Randomly selects a block
Easy to implement
Unpredictable performance

Comparison of Replacement Policies

Policy	Performance	Complexity
LRU	Best	High
FIFO	Medium	Low
Random	Low	Very Low

Exam-Ready Summary Points

Cache memory improves CPU speed by reducing memory access time.
It works on the principle of locality of reference.
Cache performance is measured using hit ratio and EMAT.
Mapping techniques decide block placement in cache.
Replacement policies decide which block to remove.

Frequently Asked Exam Questions

Define cache memory and explain its working principle.
Explain cache mapping techniques with diagrams.
What is hit ratio? Derive EMAT formula.
Compare direct, associative, and set-associative mapping.
Explain cache replacement policies

AUXILIARY MEMORIES & VIRTUAL MEMORY

Auxiliary Memory (Secondary Storage)

Meaning of Auxiliary Memory

Auxiliary memory is non-volatile storage used to store large amounts of data permanently.
It is slower than main memory but has very high capacity and low cost per bit.

Examples:

Magnetic Disk
Magnetic Tape
Optical Disk

Magnetic Disk

A magnetic disk stores data on rotating circular platters coated with magnetic material.

Examples: Hard Disk Drive (HDD)

Structure of Magnetic Disk

Platters – circular disks
Tracks – concentric circles on a platter
Sectors – divisions of tracks
Read/Write Head – accesses data
Cylinder – set of tracks at same position on all platters

Disk Access Time

Disk access time consists of:

Seek Time – time to move head to correct track
Rotational Latency – waiting for sector to reach head
Transfer Time – time to read/write data

Advantages of Magnetic Disk

Random access
Large storage capacity
Low cost

Disadvantages of Magnetic Disk

Mechanical parts → slower than SSD
Prone to physical damage

Magnetic Tape

A magnetic tape is a sequential access storage device where data is stored in a long magnetic strip.

Used mainly for:

Backup
Archiving

Characteristics

Sequential access only
Very low cost
High storage capacity

Advantages

Cheapest storage
Suitable for large backups
Long data retention

Disadvantages

Very slow access
Not suitable for real-time data access

Optical Disks

Optical disks store data using laser technology.

Examples:

CD (Compact Disk)
DVD (Digital Versatile Disk)
Blu-ray Disk

Types of Optical Disks

Type	Description
CD-ROM	Read only
CD-R	Write once
CD-RW	Rewritable
DVD	Higher capacity
Blu-ray	Very high capacity

Advantages

Portable
Durable
Low cost

Disadvantages

Slower than hard disk
Limited rewrite cycles

Comparison of Auxiliary Memories (Important Table)

Feature	Magnetic Disk	Magnetic Tape	Optical Disk
Access Type	Random	Sequential	Random
Speed	Medium	Very Slow	Slow
Cost	Low	Very Low	Low
Usage	OS, data storage	Backup	Media distribution

VIRTUAL MEMORY

Virtual memory is a memory management technique that allows execution of programs larger than physical main memory.

It creates an illusion of a large main memory by using secondary storage.

Need for Virtual Memory

Programs are large
Physical RAM is limited
Better memory utilization
Supports multitasking

Implementation of Virtual Memory

Paging (Most Common Method)

Main memory divided into frames
Virtual memory divided into pages
Page table maps pages to frames

Page Table

Stores mapping of virtual pages to physical frames

Contains:

Frame number
Valid/Invalid bit
Protection bits

Page Fault

Occurs when:

Required page is not present in main memory

Steps:

OS detects page fault
Required page is loaded from disk
Page table is updated
Execution resumes

Performance Issues in Virtual Memory

Thrashing

Excessive page faults
System spends more time swapping than executing

Page Replacement Algorithms

FIFO
LRU
Optimal (theoretical)

Advantages of Virtual Memory

Efficient RAM utilization
Large programs can run
Improved system performance

Disadvantages of Virtual Memory

Page fault overhead
Slower than physical memory
Complex OS design

Exam-Ready Conclusion

Auxiliary memory provides permanent storage at low cost.
Magnetic disks offer random access and are widely used.
Magnetic tape is best for backup and archiving.
Optical disks are portable storage media.
Virtual memory extends main memory using disk storage and paging.

Most Important Exam Questions

Explain magnetic disk structure and access time.
Compare magnetic disk, tape, and optical disk.
Define virtual memory and explain its implementation using paging.
What is a page fault? Explain its handling