CSC 35 Final Exam

Where does the processor maintain the stack?

In memory

What size integers does the stack store?

Size of the integer is dependent on bit size of the system

64 bit system will store a 64 bit integer

fixed location pointer (SO)

defines the bottom of the stack

stack pointer (SP)

gives the location of the top of stack

growing upwards

bottom pointer (SO) is the lowest address in the stack buffer, and stack grows towards higher addresses

growing downwards

bottom pointer (SO) is the highest address in the stack buffer and stack grows towards lower addresses

if the data exceeds the allocated space, a _______________________ occurs

stack overflow

stacks are implemented using __________ buffers which are __________ in size

stacks are implemented using MEMORY buffers which are FINITE in size

How does is the stack essential for subroutines to work?

* used to save the return addresses for call instructions
* backup and restore registers
* pass data between subroutines

what does the call instruction do?

transfers control to a subroutine

what is the return instruction used for?

to mark the end of a subroutine

if you forget to use the return instruction what will happen?

execution will simply continue, in memory, until a return instruction is encountered

instruction to call

call address

instruction to return

ret (there is no argument)

what happens when the return instruction is executed?

the old instruction pointer is read form the system stack and the current instruction pointer is updated to restore execution after the initial call

the operating system is...

a series of programs

the programs that make up operating systems run with ________________ that are needed to talk to all the hardware and run instructions

special privileges

processors support two modes for executing programs, these are...

* privileged (supervisor) mode
* user mode

privileged/supervisor mode can...

* run special instructions
* -talk to all the hardware

user mode can...

* only execute certain instructions
* can't talk to all the hardware

vector tables store

addresses of interrupt service routines

when a processor is interrupted/alerted that something must be handled...

it runs a special program that handles the event

during an interrupt, the processor sends an ___________________

interrupt number

the interrupt number

a pointer to a location in a vector tables that contains the address of the ISR to execute

all interrupt service routines belong to which part of the operating system?

the kernal

the three steps the processor takes to execute subroutines are..

1. backup the register file
2. execute interrupt service routine
3. restore the original executing program and register file once completed

software programs "talk" to the operating system using an _________________

API : application program interface

benefits of API's

* applications are faster and smaller
* makes the system more secure since apps do not directly talk to IO

syscall instruction

calls interrupt number reserved for programs needing attention

The rate in which instructions are executed is controlled by ____________

the CPU clock

How is the clock rate measured?

Hertz - number of oscillations per second

true/false - the faster the clock rate, the faster instructions will be executed

true

true/false - some instructions require multiple cycles to execute

true

true/false - the increase in performance increases at a linear rate

false - speed and density improve exponentially, not linearly

Moore's Law

states the performance doubles every 18 months

two major design philosophies

CISC vs RISC

RISC stands for

Reduced Instruction Set Computer

CISC or RISC?

instructions can take multiple clock cycles depending on how complex

CISC

RISC philosophy emphasizes what design principles?

* simple hardware
* software should contain the complexity rather than hardware
* minimalize memory access

CISC advantages

* generally requires fewer instructions than RISC to perform same computation
* software is easier to write
* programs tend to take less space in memory
* variable sized instructions make it possible to add more instructions

CISC or RISC?

access to memory is restricted to load/store instructions

RISC

CISC or RISC?

the number of bytes used by instructions tend to vary in sizes

CISC

CISC or RISC?

instructions typically take one clock cycle each

RISC

CISC or RISC?

bytes used by instructions tend to be fixed in size

RISC

RISC advantages

* simpler instructions simplify hardware, making processors easier to manufacture
* produces less heat and requires less energy
* memory access is minimized

CISC stands for

Complex Instruction Set Computer

CISC or RISC?

operands are generalized - each can access memory, immediate or registers

CISC

CISC or RISC?

many registers - since all instructions can only use registers for calculations

RISC

CISC or RISC?

very few general-purpose registers

CISC

CISC philosophy emphasizes what design principles?

* emphasizes flexibility in instructions
* hardware should contain complexity rather than the software

as the rate of processor growth has been greater than memory, the relative speed of memory to the processor's speed has gotten much _________

slower

why is memory the bottleneck of CISC functionality?

CISC can access memory with every instruction and it is far more efficient now to do all the work on the processor and use memory only when absolutely necessary

true/false - memory is slow compared to register-to-register operations, making is less favorable as an architecture design

true

single operand processors are also known as

accumulators

two operand processor operands

both operands are typically treated as input, one is used to store the result

single operand processors are what generally which style architecture?

CISC

three operand processors are generally which style architecture?

RISC

compact encoding

* takes as little space as possible to store instructions
* programs require less storage

simple encoding

* requires little logic to decode
* minimizes the circuitry on the process and reduces cost

the more bits available for instructions...

the more complex the opcode and unique instruction potential

in order to optimize circuitry..

the format of opcodes and instructions aid circuitry

a common trick is to use parts of the opcode to store fields.. the example in notes allows ____________ to be stored in upper nibble and _____________ contained in lower nibble

Load opcodes (one for each register), register number (used to access register file)

signed integer

represents both positive and negative numbers

unsigned integer

only represents positive

what is the value of 00001101 represented in signed magnitude? (in decimal)

13

What is the value of 10001101 represented in signed magnitude? (in decimal)

\-13

true/false - signed magnitude has one zero

false

true/false - 1's complement has two zeros (positive and negative)

true

true/false - 2's complement has two zeros (positive and negative)

false

how does 1's complement represent positive and negative values?

* there is no sign bit
* the value can be made negative by inverting each

what is the decimal value of 00001101 in 1's complement?

13

what is the decimal value of 11110010 in 1's complement?

-13

how are numbers represented in 2's complement?

after the number is inverted (like 1's complement), 1 is added to result

what is the decimal value of 111111111111111111 in 2's complement?

-1

true/false - processors know the difference between unsigned or signed integers

false, it is the programmer's responsibility to keep track if it's signed or unsigned

addition instruction

add target, value

target = register or memory

value = immediate, register, memory

subtraction instruction

sub target, value

target = register or memory

value = immediate, register, memory

negation instruction in 2's complement

neg register

which two registers are used in intel to store a product using multiplication?

RAX - contains the lower 8 bytes

RDX - contains the upper 8 bytes

which register must be used to store the input for multiplication in intel?

RAX - must store data

signed multiply instruction

imul operand

operand = register or memory only

unsigned multiply instruction

mul operand

operand = register or memory only

sign magnitude extension

(msb stores the negative sign)

* copy the old sign bit to the new sign bit and fill the rest of the new bits with zeroes (including old sign bit)

sign magnitude extension of 01001101 (77) is...

00000000 01001101

sign magnitude extension of 11001101 (-77) is...

10000000 01001101

2's complement extension

copy the old most significant bit to all the new bits

2's complement extension of 01001101 (77) is...

0000000001001101

2's complement extension of 10110011 (-77) is...

1111111110110011

which two registers does division on the x86 processor use for the numerator?

RAX - contains the lower 8 bytes

RDX - contains the upper 8 bytes

which two registers are used for the result?

RAX - contains the quotient (whole number)

RDX - contains the remainder

signed divide instruction -

idiv denominator

denominator = register or memory / the divisor

unsigned divide instruction

div denominator

sign-extension for divide instruction

cqo

sign extends rax into rdx

how does compare work?

the second argument is subtracted from the first

the result is used to determine how the operands compare and the result is discarded

compare instruction

cmp arg1, arg2

arg1 = register or memory

arg2 = immediate, register, memory

what are flags?

a boolean value that indicates the result of an action

where are flags stored?

in the status register, as individual bits

zero flag

true - if last computation resulted in zero / the two operands are equal

sign flag

true - if the msb of the result is 1

carry flag

true - if a 1 is "borrowed" when subtraction is performed

important for unsigned numbers as it indicates it exceeded the size of register or an underflow