Unit-2: Memory Organization

Prof. Shankar Mali

What is Computer Memory?

Computer memory refers to the hardware or system components used to

store data and instructions temporarily or permanently in a computer. It
plays a critical role in a computer's functionality by holding the data that
the CPU processes or needs to execute tasks. Memory can be classified
based on its speed, volatility, and permanence.
Memory Architecture
Memory Architecture

Memory architecture refers to how a computer's memory is organized and how data flows between
different types of memory and the processor.

Processor (CPU): The brain of the computer that processes data and instructions.

Cache Memory: A small, super-fast memory located inside or close to the CPU. It stores frequently used
data to speed up processing.

Main Memory (RAM): The computer's primary working memory. It temporarily holds data and
instructions the CPU is actively using. It's fast but loses data when the power is off.

Storage (Hard Drive or SSD): Long-term memory where all your files, programs, and data are stored.
It's slower than RAM but keeps data even when the computer is off.

Bus System: A pathway that connects the CPU, memory, and storage. It allows data to travel between
these components
Memory Hierarchy

The memory hierarchy is

the arrangement of
different types of memory
in a computer system
based on speed, size, cost,
and proximity to the CPU.
It helps ensure efficient
data storage and retrieval
by balancing performance
and cost.
Registers

Location: Inside the CPU.

Function: Temporarily holds data and instructions that the

CPU is actively processing.

Speed: The fastest type of memory, as it’s directly wired into

the CPU.

Capacity: Extremely small (typically 32 or 64 bits per register).

Cost: Very expensive due to high speed and integration with

the CPU.

Example: Registers store operands for immediate calculations,

like results of arithmetic operations.
Cache Memory

Location: Built into or near the CPU.

Function: Stores frequently accessed data and
instructions to avoid fetching them from slower
memory.
Speed: Very fast, but slightly slower than registers.

Capacity: Larger than registers but still relatively

small.
Cost: High, but less expensive than registers.
Cache Memory Levels

L1 Cache: Closest to the CPU,

very small (32KB–128KB), and
extremely fast.

L2 Cache: Larger (256KB–

12MB), shared between cores
in some architectures, slower
than L1.

L3 Cache: Largest (up to

64MB+), shared among all CPU
cores, slower but still faster
than RAM.
Main Memory (RAM)

Location: Connected to the motherboard, accessed by the

CPU via the memory bus.
Function: Temporarily holds data and programs currently in
use.
Speed: Slower than cache but much faster than secondary
storage.
Capacity: Typically 8GB–64GB in modern computers.
Cost: Moderate cost per byte, significantly cheaper than
cache.
Volatility: Volatile, meaning data is lost when power is off.
Example: When you open an application, its data is loaded
into RAM for quick access.
Secondary Storage

Location: Connected to the motherboard via data cables (e.g.,

SATA).
Function: Provides long-term storage for operating systems,
applications, and files.

Speed:

SSD (Solid-State Drive): Faster than traditional hard drives.

HDD (Hard Disk Drive): Slower due to mechanical parts.

Capacity: Typically 256GB–2TB or more.

Cost: Much cheaper than RAM per byte.

Volatility: Non-volatile, retains data even when power is off.
Example: Your operating system and documents are stored here.
Tertiary Storage

Location: Often external to the computer.

Function: Used for backups, archival, and
infrequently accessed data.
Speed: Very slow compared to other types of
memory.
Capacity: Large, often measured in terabytes or
higher.
Cost: Very cheap per byte.
Example: External hard drives, DVDs, Blu-rays, and
tape drives.
Cloud Storage

Location: Remote servers accessed via the internet.

Function: Provides off-site storage with virtually
unlimited capacity.
Speed: Dependent on internet bandwidth and
latency.
Capacity: Scalable, typically in gigabytes or
terabytes for individual users.
Cost: Subscription-based, depending on storage
size.
Example: Google Drive, Dropbox, and iCloud.
Characteristics of Memory

Speed
Size (Capacity)
Volatility
Cost
Accessibility
Latency
Bandwidth
Power Consumption
Physical Size
Reliability
Write Cycles
Speed

Definition: The rate at which data can be read from or written to memory.
Explanation: Faster memory allows quicker data access, improving overall
system performance.
Comparison:
Registers and cache are the fastest.
RAM is slower than cache but faster than secondary storage.
Hard drives, SSDs, and other secondary storage are much slower.
Size (Capacity)

Definition: The amount of data the memory can hold, typically measured in
bytes (e.g., KB, MB, GB, TB).
Explanation: Higher capacity memory can store more data, which is essential
for running complex programs or handling large datasets.
Comparison:
Registers and cache have the smallest size.
RAM has moderate capacity.
Secondary and tertiary storage offer the largest capacities.
Volatility Cost

Definition: The cost per byte of memory.

Definition: Whether the memory retains
Explanation: Faster and more advanced
data when power is turned off.
memory types are more expensive.
Types:
Comparison:
Volatile Memory: Loses data when
Registers and cache are the most
power is off (e.g., RAM, cache). expensive.
Non-Volatile Memory: Retains data RAM is moderately priced.
even without power (e.g., ROM, SSD, Secondary storage like SSDs and HDDs
HDD). are cheaper per byte.
Accessibility Latency

Definition: The time it takes to retrieve

Definition: How the CPU accesses the data from memory.
memory. Explanation: Lower latency means faster

Types: data access.

Direct Access: Registers, cache, and Comparison:

Registers and cache have the lowest
RAM are directly accessible by the CPU.
latency.
Indirect Access: Secondary and tertiary
RAM has moderate latency.
storage require additional steps to
Secondary storage has higher latency due
access.
to mechanical or network delays.
Power Consumption Bandwidth

Definition: The amount of power the

memory uses during operation.
Definition: The amount of data that can
Comparison:
be transferred in a given amount of
Registers and cache consume less power
due to their smaller size and proximity to
time.

the CPU. Explanation: Higher bandwidth allows

RAM consumes more power, especially in more data to be transferred, increasing

large capacities. performance.
SSDs are energy-efficient compared to
traditional HDDs.
Physical Size Reliability

Definition: The physical dimensions of

Definition: The likelihood of the
the memory component.
memory operating without errors over
Explanation: Smaller devices require
time.
more compact memory solutions.
Example: Explanation: Non-volatile memory like

Registers and cache are microscopic ROM and SSDs is generally more

and embedded within the CPU. reliable for long-term data storage than

RAM modules and storage drives are volatile memory like RAM.
physically larger.
Memory Characteristics

Characteristic Registers Cache RAM SSD HDD

Speed Fastest Very Fast Fast Moderate Slow

Size Smallest Small Moderate Large Very Large

Volatility Volatile Volatile Volatile Non-Volatile Non-Volatile

Cost Highest High Moderate Moderate Lowest

Accessibility Direct Direct Direct Indirect Indirect

What is RAM?

Random Access Memory, is a type of computer memory that allows data to be read and written
randomly, meaning that the computer can access any location in the memory directly rather
than having to read the data in a specific order. This makes RAM an essential component of a
computer system, as it enables the CPU to access data quickly and efficiently.

RAM is volatile in nature, which means if the power goes off, the stored information is lost. RAM
is used to store the data that is currently processed by the CPU. Most of the programs and data
that are modifiable are stored in RAM.

Mainly RAM have 2types

SRAM (Static RAM)

DRAM (Dynamic RAM)

SRAM (Static Random-Access Memory)

Operation: SRAM uses bistable latching circuitry (flip-flops)

•

to store each bit of data. It does not need to be refreshed as

long as power is supplied.
Speed: SRAM is faster than DRAM because it does not
•

require refreshing.
Power Consumption: SRAM consumes less power when
•

idle but more power when active compared to DRAM.

Density: SRAM is less dense, meaning it takes up more space
•

on a chip for the same amount of memory.

Cost: SRAM is more expensive to produce due to its complex
•

structure.
Use Cases: SRAM is typically used in cache memory (e.g.,
•

CPU caches) where speed is crucial.

DRAM (Dynamic Random-Access Memory)

Operation: DRAM stores each bit of data in a separate

•

capacitor within an integrated circuit. Since capacitors leak

charge, the data in DRAM must be refreshed periodically
(typically every few milliseconds).
Speed: DRAM is slower than SRAM due to the need for
•

refreshing.
Power Consumption: DRAM consumes more power when
•

idle because of the refresh cycles but less power when active
compared to SRAM.
Density: DRAM is more dense, allowing for larger memory
•

capacities in the same physical space.

Cost: DRAM is cheaper to produce, making it suitable for
•

larger memory requirements.

Use Cases: DRAM is commonly used as the main memory in
•

computers (e.g., RAM in PCs and servers) where larger

capacities are needed.
SRAM Vs DRAM

Feature SRAM (Static RAM) DRAM (Dynamic RAM)

Storage Mechanism Uses flip-flops (transistors) to store data. Uses capacitors to store data.
Requires periodic refreshing
Refresh Requirement No refresh needed (static).
(dynamic).
Speed Faster access times (nanoseconds). Slower due to refresh cycles.
Lower power when idle, higher when
Power Consumption Higher power due to refresh cycles.
active.
Less dense (takes more space for same More dense (higher capacity in
Density
memory). same space).
Cost More expensive to produce. Cheaper to produce.
Cache memory (e.g., CPU L1, L2, L3 Main system memory (e.g., RAM in
Use Cases
caches). PCs).
Volatile (loses data when power is
Volatility Volatile (loses data when power is off).
off).
What is DDR? (Double Data Rate Memory)

DDR (Double Data Rate) is a type of synchronous dynamic RAM (SDRAM) that can transfer data
twice per clock cycle—once on the rising edge and once on the falling edge of the clock signal. This
doubles the data transfer rate compared to older single data rate (SDR) memory. Over time, DDR
technology has evolved to provide higher speeds, lower power consumption, and better efficiency:

Introduction
DDR Type Speed (MT/s) Voltage Bandwidth
Year
DDR1 200–400 MT/s 2.5V 1.6–3.2 GB/s 2000
DDR2 400–1066 MT/s 1.8V 3.2–8.5 GB/s 2003
DDR3 800–2133 MT/s 1.5V 6.4–17 GB/s 2007
DDR4 1600–3200 MT/s 1.2V 12.8–25.6 GB/s 2014
DDR5 3200–8400 MT/s 1.1V 32–67.2 GB/s 2021
 Cache Memory : Cache Memory Principles

Cache memory is intended to give

memory speed approaching that of
the fastest memories available, and
at the same time provide a large
memory size at the price of less
expensive types of semiconductor
memories.
The concept is illustrated in Figure
4.3a.
There is a relatively large and slow
main memory together with a
smaller, faster cache memory. The
cache contains a copy of portions of
main memory.
When the processor attempts to
read a word of memory, a check is
 Cache Memory
(Cont.) determine if the word is in the
cache.
If so, the word is delivered to the
processor. If not, a block of main
memory, consisting of some fixed
number of words, is read into the cache
and then the word is delivered to the
processor.
Figure 4.3b depicts the use of multiple
levels of cache. The L2 cache is slower
and typically larger than the L1 cache,
and the L3 cache is slower and typically
larger than the L2 cache
Cache Memory

Cache is a type of memory that is used to increase the speed of data access. Normally,
the data required for any process resides in the main memory. However, it is
transferred to the cache memory temporarily if it is used frequently enough.

A diagram to better understand the data transfer in cache management is as follows −

Cache Performance
The cache performance can be explained using the following steps −
If a process needs some data, it first searches in the cache memory. If the data is
▪

available in the cache, this is termed as a cache hit and the data is accessed as
required.
If the data is not in the cache then it is termed as a cache miss. Then the data is
▪

obtained from the main memory. After that the data is transferred to the cache
memory under the assumption that it will be needed again.
The performance of the cache is measured using the hit ratio. It is the number of
▪

cache hits divided by the total cache accesses. The formula for this is:
 Disadvantages of Cache Memory

Some of the disadvantages of cache memory are as follows −

Since the cache memory is quite fast, it is extremely useful in any computer system. However, it is also
quite expensive.

The cache is memory expensive as observed from the previous point. Also, it is located directly on the
processor chip. Because of these reasons, it has a limited capacity and is much smaller than main
memory.
Read Only Memory (ROM)

ROM (Read-Only Memory) is a type of non-volatile memory that permanently stores

data. Unlike RAM, which loses data when the power is off, ROM retains information
even after power loss. It is mainly used for firmware, system boot processes, and
embedded applications.
Read Only Memory (ROM)

What is it?

ROM is a pre-programmed memory that comes with data already stored at the time of manufacturing.

The data in ROM cannot be modified or erased.

How it Works?

Since ROM is non-volatile, it stores critical data such as the BIOS (Basic Input/Output System), which helps
start a computer.

When the system is powered on, the processor reads instructions from ROM to boot up the system.

Uses of ROM

Used in BIOS, embedded systems, calculators, and old gaming consoles.

Example: The firmware in old computers and video game consoles like Nintendo cartridges.
PROM (Programmable Read-Only Memory)

what is it?

PROM is a blank ROM chip that allows the user to write data once using a special device called a PROM programmer.

Once written, the data cannot be changed or erased.

How it Works?

Inside a PROM chip, there are tiny fuses that can be "burned" (blown permanently) when programmed.

The pattern of burned and unburned fuses represents binary data (1s and 0s).

Since the fuses cannot be repaired, data is permanent after programming.

Uses of PROM

Used in firmware, game cartridges, and microcontrollers where data needs to be stored permanently.

Example: Used in old gaming consoles like Atari and arcade machines.
EPROM (Erasable Programmable Read-Only Memory)

What is it?

EPROM is a type of PROM that can be erased and reprogrammed multiple times using ultraviolet
(UV) light.

It has a transparent quartz window on top through which UV light can pass to erase the data.

How it Works?

When exposed to UV light for 10-30 minutes, the stored data is erased, making the chip reusable.
After erasing, new data can be written using a special EPROM programmer.

Uses of EPROM

Used in BIOS chips, microcontrollers, and industrial equipment where updates are required.
Example: Early computer BIOS chips were often EPROM-based, allowing manufacturers to update
system firmware.
EEPROM (Electrically Erasable Programmable Read-Only Memory)

What is it

EEPROM is an advanced version of EPROM that can be erased and rewritten electrically, without UV light.

It allows byte-level modification, meaning specific sections of data can be changed without erasing the entire
memory.

How it Works?

Uses electrical signals to erase and rewrite data.

Unlike EPROM, which needs a UV light and a dedicated erasing process, EEPROM can be reprogrammed while
still inside a system.

Uses of EEPROM

Used in modern BIOS chips, smart cards, RFID tags, and microcontrollers for firmware storage and updates.

Example: Used in modern computers’ BIOS and USB flash drives (a form of EEPROM-based memory).
EEPROM (Electrically Erasable Programmable Read-Only Memory)

DMA (Direct Memory Access)

DMA (Direct Memory Access) is a technique that allows certain hardware components, like disk
controllers, network cards, or GPUs, to transfer data directly to and from the main memory
(RAM) without involving the CPU. This improves system performance by freeing the CPU from
managing repetitive data transfer tasks.
Why DMA is needed?

DMA (Direct Memory Access)

In a system without DMA, the CPU is responsible for handling all data transfers between
memory and I/O devices, which leads to:

High CPU utilization.

Slower system performance due to CPU involvement in data transfer.

With DMA, the CPU initiates the transfer but is free to perform other tasks while the data
transfer happens independently.
Components of DMA

DMA Controller (DMAC): A dedicated hardware module that manages data transfers.

Memory: The primary location where data is stored.

I/O Device: The source or destination of the data transfer.

Bus System: Includes the data bus, address bus, and control bus to facilitate communication.
How DMA Works (Steps in a DMA Transfer)

1. CPU Initiates Transfer: The CPU sets up the DMA controller by specifying:
Source and destination addresses.
Amount of data to transfer.
Type of transfer (read or write).
2. DMA Controller Takes Over: The DMA controller directly communicates with the memory and I/O device
to perform data transfer.
3. Cycle Stealing (If Needed): The DMA controller temporarily takes control of the system bus from the CPU
to perform data transfer.
4. DMA Transfer Completion: Once the transfer is complete, the DMA controller sends an interrupt signal
to notify the CPU.
Types of DMA transfer

1. Burst Mode (Block Transfer Mode):

DMA transfers a complete block of data at once.
CPU is halted during the transfer.
Suitable for high-speed data transfer.
2. Cycle Stealing Mode:
DMA controller transfers one data unit at a time.
CPU execution is only briefly interrupted.
Used when continuous CPU operation is required.

3. Transparent Mode (Hidden DMA):

DMA only transfers data when the CPU is not using the bus.
No CPU performance impact but slower data transfer.
Cycle Stealing in DMA

Definition: A process where the DMA controller momentarily "steals" the bus cycles from the
CPU to transfer data.

Effect: CPU performance is slightly affected, but it continues execution with minimal delay.

Example: Used in audio/video processing where real-time data transfer is needed.

Cycle Stealing in DMA

Advantages of DMA

Reduces CPU workload.

Faster data transfer compared to CPU-controlled I/O.

Enables real-time data processing (e.g., audio and video streaming).

Supports multi-device communication efficiently.

Disadvantages of DMA

Complex hardware (requires a DMA controller).

Potential memory access conflicts between CPU and DMA.

Needs proper bus arbitration to avoid data corruption.

Virtual Memory

Virtual Memory is a technique in modern operating

systems that allows a computer to use part of its
storage (hard disk or SSD) as if it were RAM. It creates
an illusion that the system has more memory than it
physically does by managing memory efficiently
between RAM and disk space.
Without virtual memory, if a system runs out of RAM,
it would either crash or stop running new programs.
Virtual memory prevents this by temporarily moving
data from RAM to disk storage, allowing processes to
run smoothly.
Why we need Virtual Memory?
Memory Isolation & Protection:
Limited Physical RAM:
Prevents one process from interfering with
Many applications require more memory than the another’s memory space, enhancing security and
system's physical RAM can provide. stability.

Virtual memory enables systems to run large Process Execution Without Loading Entire
applications smoothly. Program:

Efficient Multitasking: Some applications (e.g., large software like

Photoshop, databases, or games) don’t need to be
Allows multiple programs to run at the same time fully loaded into RAM at once.
by swapping inactive parts of memory to disk.
Virtual memory loads only the necessary portions,
optimizing memory usage.
How does need Virtual Memory Works?

A. Address Translation C. Swapping

The CPU generates virtual addresses instead of When RAM is full, the OS moves inactive pages
physical addresses. from RAM to a reserved area on disk called swap
space.
A hardware unit called the Memory Management Unit
(MMU) translates these virtual addresses into physical
This process, called paging in (from disk to RAM)
addresses in RAM or storage.
and paging out (from RAM to disk), helps manage
memory efficiently.
B. Paging (Most Common Virtual Memory Technique)

Memory is divided into fixed-size pages (in virtual

memory) and page frames (in physical RAM).

The Page Table maps virtual pages to physical frames.

If a required page is not in RAM, a page fault occurs,

and the system loads it from disk (swap space).
Components of Virtual Memory

RAM (Physical Memory) – Stores active pages and frequently accessed data.

Hard Disk/SSD (Swap Space) – Stores less frequently used data when RAM is full.

Page Table – A mapping system that links virtual memory addresses to physical memory
locations.

Memory Management Unit (MMU) – Handles address translation between virtual and physical
memory.

CPU & Operating System – Manages virtual memory operations and ensures efficient data
access.
Virtual Memory Techniques

Paging
Demand Paging
Divides memory into fixed-size blocks (pages).
Loads pages into memory only when they are
Reduces external fragmentation but can cause needed (on-demand).
internal fragmentation.
Reduces memory waste but may cause page
Segmentation faults.

Divides memory into logical units (segments) like

functions, arrays, or stacks.

More flexible than paging but leads to external

fragmentation
Advantages of Virtual Memory

Allows Running Large Applications: Programs that require more memory than available RAM can still
run smoothly.

Enhances Multitasking: Multiple applications can run without exhausting RAM.

Provides Memory Isolation & Protection: Prevents unauthorized access between processes.

Cost-Effective: Allows cheaper systems to run demanding applications without needing expensive
RAM upgrades.

Efficient Memory Usage: Optimizes RAM usage by keeping only necessary data in active memory.
Disadvantages of Virtual Memory

Slower Performance Compared to RAM: Since accessing data from disk storage is much slower than
accessing RAM.

Page Thrashing: When too many pages are swapped between RAM and disk, leading to performance
degradation.

Increased Wear on SSDs: Frequent read/write operations to disk can shorten the lifespan of SSDs.

Extra Disk Space Required: Swap files take up storage space, which could be used for other data.
What is Internal Fragmentation?

Internal fragmentation happens when the

memory is split into mounted-sized blocks.
Whenever a method is requested for the memory,
the mounted-sized block is allotted to the
method. In the case where the memory allotted
to the method is somewhat larger than the
memory requested, then the difference between
allotted and requested memory is called internal
fragmentation. We fixed the sizes of the memory
blocks, which has caused this issue. If we use
dynamic partitioning to allot space to the
process, this issue can be solved.
What is External Fragmentation?

External fragmentation happens when there’s a

sufficient quantity of area within the memory
to satisfy the memory request of a method.
However, the process’s memory request cannot
be fulfilled because the memory offered is in a
non-contiguous manner. Whether you apply a
first-fit or best-fit memory allocation strategy
it’ll cause external fragmentation.