A Quick Start Guide to CS/COE 0447

Informal discussion about basic concepts

We’ll refine and more formally define many of this

material over the next few weeks!!!

Digital Computer
•  Stores information and operates on discrete numbers, proceeding in
discrete steps from one “instruction” to the next

•  Basic representation is “binary number”

–  Each numeric binary digit (a bit) is a “0” or “1”. Base-2 number.
–  Put several digits together for a larger number (e.g., 0101b is 5d)
•  Binary numbers used for many items
–  Characters (letters, like ‘A’, ‘B’, ‘C’, …)
–  Positive & negative integer numbers (no decimal point)
–  Fractional numbers (with a decimal point)
–  Pixels (dots) on a graphics display
•  Binary number as a sequence of 0s and 1s
–  0100 0001 is a 8-bit binary number (byte)
–  Represents the character ‘A’ in the ASCII encoding (later)
–  Represents the decimal number 65

1
Binary Numbers
•  Computer operates on a “native” fixed-size binary quantity
–  4-bit computer: Native size is a binary number with 4 bits (nibble)
–  8-bit computer: Native size is a binary number with 8 bits (byte)
–  16-bit computer: Native size is a binary number with 16 bits (halfword)
–  32-bit computer: Native size is a binary number with 32 digits (word)
–  64-bit computer: Native size is a binary number with 64 digits
–  36-bit computer: Native size is a binary number with 36 digits

Binary Numbers
Really just a different representation for a quantity

How many ways can we represent these animals?

“Ten happy cows”
“Dix vaches heureux ”
“10 cows”
10 (decimal quantity)
1010b (base-2 quantity for 10 decimal)
110001b 110000b (base-2 quantity for characters ‘1’, ‘0’)
Ah (base-16 quantity for 10 decimal)
12o (base-8 quantity for 10 decimal)

2
Other Representations
•  Base-10 numbers (decimal)
–  Each digit represents one of ten values
–  0, 1, 2, 3, 4, 5, 6, 7, 8, 9

•  Base-16 numbers (hexadecimal)

–  Each digit represents one of sixteen values
–  0, 1, 2, 3, 4, 5, 6, 7, 8, 9, A, B, C, D, E, F

•  Base-8 numbers (octal)

–  Each digit represents one of eight values
–  0, 1, 2, 3, 4, 5, 6, 7

•  Do you get the idea??? ? 

A bag o’ pennies!
Consider a bag of 100 pennies.

How can we represent it?

$1, 0.63 GBP, 0.76 Euro, 1 Susan B Anthony coin

OK, let’s try our bases…

Base-10: 100
Base-2: 1100100b
Base-16: 64h
Base-8: 144o

3
Why all this fuss?
•  Binary numbers
–  Digital computers use “on/off” switches (transistors)
–  On is the binary digit 1
–  Off is the binary digit 0

•  Have you written down a 32 digit binary number?

11000010111101010001100010010011b

•  What a pain! Is there a better way?

•  Base-16 (hexadecimal) is convenient notation

–  Compact form
–  Easy and fast conversion binary <-> hexadecimal
–  Fixed sizes tend to be “chunkable” into 4s (consider slide 3)

Hexadecimal numbers
Compare: 11000010111101010001100010010011b
To: C2F51893h

What about a 64-bit number? Which would you rather write????

How can we convert from binary to/from hex?

Simple, quick approach: A table approach (we’ll be more formal later)
Construct a table that maps binary numbers to hex numbers

Hex digit is one of 16 numbers

Binary digit is one of 2 numbers

We need one hex digit for every four binary digits (4 digits, 2 per digit = 24 = 16)

4
How to count in binary?
Suppose, we add two 1-bit binary numbers:
0+0=0 0+1=1 1+0=1
But what about 1 + 1?

It’s just like decimal numbers. We must carry the one:

1 10 11 100
+1 +1 +1 + 1
--- --- --- ----
10 11 100 101

We have to carry the one since a single bit is only 0 or 1

Binary - Hex Table

Count from 0 to 15 in binary, then write hex digit
binary hex binary hex
0000 0 1000 8
0001 1 1001 9
0010 2 1010 A
0011 3 1011 B
0100 4 1100 C
0101 5 1101 D
0110 6 1110 E
0111 7 1111 F
remember how to add in binary, e.g.,:
0001 0101 1011
+ 0001 + 0001 + 0001
------ ------ ------
0010 0110 1100

5
Use table to convert
For every four binary digits, look up the hex symbol
Write the hex symbol in corresponding position for four binary digits
e.g., 00001110001111010000100011000100b

Grouped into “chunks” with four binary digits per chunk:

0000 1110 0011 1101 0000 1000 1100 0100

0 E 3 D 0 8 C 4

To go from hex to binary, simply do this process in reverse.

Converting Decimal to/from Binary

•  How can we convert between decimal and binary?

•  Binary numbers represent “sums of powers of 2”

•  We can exploit this property to easily convert

•  From decimal to binary

–  You are given the binary number size (e.g., 8 bits, 16 bits, 32 bits)
–  Write decimal number as a sum of powers of 2
–  Each exponent value determines a “position” in the binary number
–  Positions numbered 0 to number size – 1
–  Positions ordered right to left (0 is rightmost, number size – 1 is left)
•  8-bit number: bit7 bit6 bit5 bit4 bit3 bit2 bit1 bit0
–  Write “1” in exponent position for each power of 2 in sum
–  Write “0” in all other positions

6
Example: Decimal to Binary
•  Suppose 8 bit number
•  Consider 63 decimal

•  What are the powers of 2 for 63?

•  How many “positions” are in the binary number?
•  Which positions have 1s and which ones have 0s?
•  Write down the 1s and 0s for the number?

•  Answers:
–  63 = 32 + 16 + 8 + 4 + 2 + 1 = 25 + 24 + 23 + 22 + 21 + 20
–  8 positions b/c it’s an 8-bit binary number
–  1s in positions 0, 1, 2, 3, 4, 5 and 0s in positions 6, 7
–  00111111

From Binary to Decimal

•  There’s a “1” in any position with a power of 2 value
•  Thus, we can sum the powers of 2s that correspond to the “1s
positions” to get the decimal number!

•  Consider the 8-bit number: 00101101b

•  What positions (powers of 2) have 1s?

•  What is the sum of the powers of 2?

•  Answer:
–  Positions 0, 2, 3, 5
–  Sum is: 20 + 22 + 23 + 25 = 45 decimal

7
System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software
(Compiler, Assembler, Linker)

Operating System

Instruction Set Architecture (ISA)

Processor Implementation

System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software Applications written in

(Compiler, Assembler, Linker) a high-level
programming language
Operating System (e.g., Java, C, C++)

Instruction Set Architecture (ISA)

Processor Implementation

8
System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software
(Compiler, Assembler, Linker)
Converts high-level
Operating System language programs
into a form that can run
Instruction Set Architecture (ISA) on the computer
system
Processor Implementation

System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software
(Compiler, Assembler, Linker)

Operating System
Provides common
functionality needed
Instruction Set Architecture (ISA) when program executes
and has a way to share
Processor Implementation computer resources
among programs

9
System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software
(Compiler, Assembler, Linker)

Operating System
Gives an interface (a
Instruction Set Architecture (ISA) “contract”) between the
software & the hardware;
Hardware needs to know
Processor Implementation what to do. (0447!)
Different ISAs are possible!

System layers
•  Computer system consists of several “layers”
•  It’s stack of components that work together

App 1 App 2 App 3 App 4

System Software
(Compiler, Assembler, Linker)
A specific hardware
Operating System circuit (set of transistors
& wires) that implements
Instruction Set Architecture (ISA) the ISA. You can
implement it different
ways!
Processor Implementation

10
From a C (C++ or Java) program to running it
•  Software tools: Compiler, Assembler, Linker
–  Convert the HLL program into machine code
–  Processor (computer system) fetches & executes machine code
–  It doesn’t “know” C, C++ or Java!

•  Once program converted into machine code:

–  Program is stored in memory
–  Processor fetches one instruction at a time (the conveyor belt)
–  Processor executes the instruction until done (the control tower)
–  Continues executing instructions until program is done

•  This is the “fetch-execute cycle”. Fetch an instruction, execute it.

From a C (C++ or Java) program to running it

hello.c

compiler

hello.s

assembler

crt0.o
hello.o

linker hello

11
Assembly language and machine code
•  Instruction set architecture
–  Describes the “machine instructions” that the processor can execute
–  Instructions say what computation to do (e.g., add two numbers)
–  Gives other details, such as temporary storage locations
–  Registers are “temporary storage” that can be used in computations
–  Many different instruction set architectures (Intel x86, Sun SPARC, etc)

•  Assembly language
–  Simply a human readable form to write instructions
–  It’s a programming language, much like Java, C++ or C
–  Except it’s at the lowest level of the software stack
–  Many different assembly languages, usually one for each ISA

•  Machine instructions (a.k.a., “machine code”)

–  Simply a machine readable form of instructions
–  A machine instruction is simply a binary number!

A little example
Let’s add three numbers: 10, 20, 30
Here’s a C program:
int main(void) {
int sum;
sum = 10 + 20 + 30; // sum is 10+20+30=60
}
Here’s a MIPS program (assembly language):
addiu $5,$0,10 ; $5 is register, $5=0+10=10
addiu $5,$5,20 ; adds 20 to $5, $5=10+20=30
addiu $5,$5,30 ; adds 30 to $5, $5=30+30=60

The addition is done as three separate computations (0+10,10+20,30+30)

Register $5 is needed as temporary storage to hold sum between computations

12
A little example
Here’s a MIPS program (assembly language):
addiu $5,$0,10 ; $5 is register, $5=0+10=10
addiu $5,$5,20 ; adds 20 to $5, $5=10+20=30
addiu $5,$5,30 ; adds 30 to $5, $5=30+30=60

How about the machine code?

Assembler converts the above form into binary numbers
0x2405000a binary for addiu $5,$0,10
0x24a50014 binary for addiu $5,$5,20
0x24a5001e binary for addiu $5,$5,30
note: 0x notation means the number that follows is hexadecimal

A Loop on Two Architectures

How much does the assembly language differ between ISAs?
int main(void)
{
int i, sum;
sum = 0;
for (i = 0; i < 10; i++) // sum numbers 0 to 9
sum = sum + i;
}

The assembly language and machine instructions for this program were
created for two different instruction sets:

Sun SPARC processor

MIPS processor

13
 A Loop on Two Architectures
C program

int main(void) {
int i, sum;

sum = 0;
for (i=0;i<10; i++)
sum=sum+i;
}

Let’s see this loop on the

SPARC and MIPS processors.

Careful look at the assembly

and the machine code!

A Loop on Two Architectures

C program SPARC Assembly

int main(void) { Compiler

int i, sum; mov 0,%o1
mov 0,%g1
sum = 0;
for (i=0;i<10; i++) L2: add %o1,%g1,%o1
sum=sum+i; add %g1,1,%g1
} cmp %g1,10
blt L2
...

Compiler translates the C

program into a sequence of
assembly language
instructions for the Sun
SPARC processor.

14
 A Loop on Two Architectures
C program SPARC Assembly MIPS Assembly

int main(void) { Compiler

int i, sum; mov 0,%o1 move $5,$0
mov 0,%g1 move $3,$0
sum = 0;
for (i=0;i<10; i++) L2: add %o1,%g1,%o1 L2: add $5,$5,$3
sum=sum+i; add %g1,1,%g1 addi $3,$3,1
} cmp %g1,10 slti $2,$3,10
blt L2 bne $2,$0,L2
... ...

The same C program can

be translated by a compiler
for the MIPS processor as
well.

Do you see differences?

A Loop on Two Architectures

C program SPARC Assembly MIPS Assembly

int main(void) { Compiler

int i, sum; mov 0,%o1 move $5,$0
mov 0,%g1 move $3,$0
sum = 0;
for (i=0;i<10; i++) L2: add %o1,%g1,%o1 L2: add $5,$5,$3
sum=sum+i; add %g1,1,%g1 addi $3,$3,1
} cmp %g1,10 slti $2,$3,10
Assembler

blt L2 bne $2,$0,L2

... ...

SPARC Machine Code

0x9de3bf90
The assembler for the 0x92102000
SPARC converts the 0x82102000
human readable assembly 0x92024001
into actual machine code 0x82006001
that the processor is able 0x80x06009
to execute (binary 32-bit ...
numbers).

15
 A Loop on Two Architectures
C program SPARC Assembly MIPS Assembly

int main(void) { Compiler

int i, sum; mov 0,%o1 move $5,$0
mov 0,%g1 move $3,$0
sum = 0;
for (i=0;i<10; i++) L2: add %o1,%g1,%o1 L2: add $5,$5,$3
sum=sum+i; add %g1,1,%g1 addi $3,$3,1
} cmp %g1,10 slti $2,$3,10

Assembler
blt L2 bne $2,$0,L2
... ...

SPARC Machine Code MIPS Machine Code

0x9de3bf90 0x00002821
Similarly, the assembler for 0x92102000 0x00001821
MIPS turns the assembly 0x82102000 0x00a32821
into machine code. 0x92024001 0x24630001
0x82006001 0x2c62000a
MIPS also has 32-bit instr. 0x80x06009 0x1440fffc
... ...
Do you see differences?

A Loop on Two Architectures

What is the correct value for sum?

•  45 in decimal

•  101101 in binary

•  2D in hexadecimal

•  55 in octal (base-8  )

16
Try Running the Assembly Program with MARS
The assembly file is on the CS 0447 web site:
–  http://www.cs.pitt.edu/~childers/CS0447/examples/loop-mars.asm

Steps to run the program

1.  Save this file to a local directory on your computer (say “foo”)
2.  Download the MARS simulator (see CS0447 web site).
3.  Double click on the simulator icon (the jar file)
4.  Select “File” -> “Open”, then navigate to directory “foo”, select the
file “loop-mars.asm”.
5.  Select “Run” -> “Assemble”. This assembles your program.
–  You are now in the Execute window, which shows the machine code.
6.  Select “Run” -> “Go”. This runs your program. It stops with:
-- program is finished running –
7.  Look in the Registers window. The value of $a1 should be 45. This
is the total sum of the numbers 0 to 10.

17