1. Computer Science 1000
Terminology II

2. Storage a computer has two primary tasks
store data
operate on data
a processor's primary job is to operate on data
math operations
move operations
note that processors do have a very small amount of storage
how the majority of data stored by the machine?

3. Storage there are a variety of storage media available for computers:
RAM
hard drive
removable media
these storage types are differentiated by:
capacity
price
latency
first, we should determine what is being stored

4. Information Storage
ask a non-computer person what their computer stores
programs/apps/games
pictures
songs/videos
documents (text)
what does it mean to store an object, like a piece of text, in a computer?
in other words, how is it represented?

5. Information Storage consider a notebook (for comparison)
how is a piece of text stored/represented?
as a set of written symbols
the set of available symbols depends on your language
individual symbols can be combined into other objects (e.g. words, sentences)

6. Information Storage
in a computer, information is stored as a set of bits
a bit is short for binary digit
in simplest terms, a binary digit is either 0 or 1
hence, information stored by a computer is simply a set of 0s and 1s

7. Information Storage how does the computer store other information?
other information is encoded in binary
the way that information is stored in binary depends on the information type

8. Information Storage numbers
people typically use numbers in decimal format
represented by digits 0-9
any decimal number can be represented in binary form
for example, here are the first 16 integers in binary:
Dec
Bin 4
100 8
1000 12
1100 1 5
101 9
1001 13
1101 2 10 6
110
1010 14
1110 3 11 7
111
1011 15
1111

9. Information Storage numbers – notes
the entire number is typically coded in binary, not each individual digit
e.g. 49 in binary is , not
most numbers are stored as a fixed number of bits
e.g. 32-bit numbers
each number stored as a 32-bit sequence
smaller numbers are padded on left with zeroes (like decimal)
e.g (1110) as 32-bit number:

10. Information Storage text e.g. ASCII: 8-bit sequence
each character in a piece of text has a binary encoding
e.g. ASCII: 8-bit sequence
each character has a unique 8-bit sequence

11. Information Storage image
a digital picture is made up of pixels (tiny squares)
each pixel stored as its colour
each colour has a unique binary encoding
images will often indicate their colour depth
e.g. 24-bit colour uses 24 bits per colour
example (RGB): pure red:

12. Information Storage context
the previous representation of the colour red is also the binary representation of 16,711,680
so when we see that sequence, how do we determine what kind of data it is?
it's up to a program to interpret the number
often, the file type is used as a hint
different programs will interpret the same sequence differently

13. example: kev.png

14. Information Storage representations
the previous was a brief introduction to how information is encoded, to facilitate understanding of memory and storage
later in the semester, we will consider an entire chapter on how information is stored, with topics like:
binary representation of negative numbers, and numbers with a decimal point (3.4)
other text representations (e.g. Unicode)

15. Information Storage units and prefixes byte: 8 bits (typically)
most storage is measured in bytes, rather than bits
hence, a 100 byte file would contain 800 bits
bits and bytes are typically abbreviated as b and B
hence, 80 B = 80 bytes
= 640 b = 640 bits

16. Information Storage – Unit Prefixes
bits and bytes are often abbreviated using SI (metric) prefixes
for example:
K (kilo) e.g. kilobyte (KB)
M (mega) - e.g. megabit (Mb)
G (giga) e.g. gigabyte (GB)
T (tera) e.g. terabyte (TB)

17. Information Storage – Units
it is not always clear what the multiplier is
when referring to main memory, we typically use powers of two
hence, the prefix kilo means multiply by 210 , and not 1000
hence, 1 KB = 1024 bytes, 3 KB = 3072 bytes ...
other multipliers:
mega: =
giga: 230 =
when used in this context, known as binary prefixes

18. Information Storage – Units
when referring to other storage types, we typically use powers of 10
hence, the prefix kilo means multiply by 103, like you are used to
mega : 106
giga: 109
hence, 500 GB = 500,000,000,000 bytes
when used in this context, known as decimal prefixes

19. Information Storage – Units
the industry is not consistent
when you buy a 4 GB USB key, Windows will often report it as smaller, as it assumes that 4 GB = 4 x 230

20. Other Interesting Example

21. Storage Media
now that we know what is being stored, and how to define it, let's consider different ways to store it
types we will consider:
volatile storage
persistent storage

22. Volatile Storage typically referred to as memory
defined as storage that requires a continuous power source to maintain its state
in other words, when its power source is disconnected, all memory is erased
and you lose your data
your CPU cache discussed previously would be considered volatile memory
however, RAM is the primary volatile storage on most computers

23. RAM Random Access Memory also referred to as main memory
the location of your program and associated data when your program is running
example: consider a running web browser
stores:
instructions (for your processor)
images and text from the webpage
things that you can't see (e.g. cookies, passwords)

24. RAM
the most defining feature of a system's main memory is its capacity
the amount of information that it can store
modern consumer systems typically have 2-16 GB of RAM
4-8 GB is very common
in 8 GB of RAM, you could store:
~4.2 million pages of text (~129 Encyclopaedia Britannica 2010 ed.)
~2000 songs
remember: for main memory, 1 GB = 230 bytes, not 109

25. CPU – RAM

26. Why does RAM capacity affect performance?
recall that RAM stores programs and data
hence, the bigger the RAM, the more programs and data it can store
this means:
more programs can be loaded into memory at once*
more data can be stored in main memory (important for large media items like movies)
certain programs (e.g. newer games) have minimum memory requirements just to run
* this ignores a concept called virtual memory, discussed later

27. Why Random Access Memory?
named because any location on RAM chip can be accessed in (nearly) the same amount of time
compare this to sequential access memory
example: magnetic tape storage
items directly under the reader can be accessed quite quickly
feeding the tape to find other locations is extremely slow
hence, RAM devices are typically much faster

28. Persistant Storage sometimes referred to as non-volatile memory
defined as storage that maintains its state even when no power source is connected
in other words, state is maintained between power interruptions
although there are other potential forms of data corruption
many types of persistent storage
hard drives
optical drives
key drives

29. Hard Drive also referred to as hard disk or simply disk
the primary source of persistent storage on modern machines
like RAM
can store programs, documents, images, videos, etc
unlike RAM:
items in persistent storage are typically not in use
they are loaded into RAM from your hard drive in order to be used

30. Hard Drive
like RAM, the most defining feature of a hard drive is its capacity
typical consumer hard drives range in size from 500 GB to 4 TB
consider 2 TB of hard disk space:
~1 billion pages of text (~30000 Encyclopaedia Britannica)
~ songs (mp3)
remember: for persistent storage, 1 GB = 109 bytes, not 230

32. Hard Drive vs RAM
RAM and hard drives store data in fundamentally different ways
details beyond scope of the class
one of the ways in which they differ is price
by price, let's consider $/GB (to be fair)
note that certain things can affect this range (e.g. laptop RAM is usually more expensive than desktop RAM)

33. RAM – Example
8 GB = $50-$60  $6.25/GB - $7.50/GB

34. Hard Drive - Example
1 TB = $70-$75  $0.07/GB - $0.075/GB

35. Hard Drive vs RAM persistence capacity price
hard drives are persistent, no data is lost when power is disrupted
RAM is volatile, loss of power = RAM is erased
capacity
most consumer hard drives: 500GB – 2TB of HD
most consumer RAM: 2GB – 16GB
price
hard drives cost pennies per GB of storage
RAM costs dollars (about a 100 times more)
what is the advantage of RAM over an HD?

36. Hard Drive vs RAM answer: speed!! RAM is fast compared to HD
performance measured in two ways
access time
transfer rate

37. Storage – Access Time time to retrieve a single random piece of data
for modern RAM:
50 – 150 nanoseconds*
for modern hard drives:
5 – 15 milliseconds*
hence, RAM is the clear winner
performance can vary depending on how data is accessed** * **

38. Storage – Transfer rate
how much data can be transferred in a second
for modern RAM:
6-17 GB/s
for modern hard drives:
MB/s*
again, RAM is the clear winner
better technologies (e.g. SSD drives) improve HD performance, but still much slower than RAM *

39. RAM vs. Hard Drive
in summary, RAM has the ability to access and transfer data much quicker
for running programs, it is critical that data latency be minimized
otherwise, your processor would always be waiting
although more expensive and less spacious, RAM makes your current computer experience possible

40. Hard Drive – RPMs
one other common feature listed with typical hard drives is their RPMs
common values: 5400, 7200, 10000
RPMs stand for revolutions per minute
basically, more RPMs = better performance
to understand why, we must how consider how hard drives are constructed

41. Hard Drive – Construction
data stored magnetically on platters, which are just smooth round surfaces
data is read/written by the head, which is at the end of the arm mechanism that you see
these platters spin, and the arm moves to a particular location and reads the data that passes under it
hence, the faster it spins, the faster that data can be accessed

42. Hard Drive – SSD a newer technology than magnetic drives
no moving parts (quiet)
considerable performance improvement over magnetic hard drives
throughput: MB/s
considerably more expensive
over $1/GB