2021 Chapter 2 CPU Lecture 2 Stud

Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 65

CHAPTER Two

THE CENTRAL PROCESSING


UNIT (CPU)
1
part 2
MACHINE INSTRUCTION CHARACTERISTICS
 The operation of the CPU is determined by
the instructions it executes

 We will investigate
 the design of the instruction set
 The impact of the set on the design of the
overall computer system

7
ELEMENTS OF A MACHINE INSTRUCTION
 Each instruction must contain all the information
required by the CPU for execution
 These elements are :
 Operation code (Op code)
 Specifies the operation to be performed
 Source Operand(s) reference(s)
 Specifies a register or memory location of operand data
 With this/these operands

 An operation may involve one or more source operands

 Inputs for the operation specified

8
ELEMENTS OF A MACHINE INSTRUCTION…
 Result Operand reference
 Where the result of the operation should be placed ?
 Put the answer here

 Next Instruction reference


 From where to fetch the next instruction
 When you have done that, do this...

 In most cases not explicitly stated in the instruction

 Implicitly the next instruction, the one that


logically follows the current one in the program

4
OPERAND LOCATIONS
 Where Can Operands Be?
 Source and result operands can be in one of the three
areas:
 Main memory
 Memory address must be supplied

 CPU register

 Implicitly referenced, if only one register exist

 Explicitly using its unique number, if


more than one register exists
 I/O device

 Specify the I/O device for the operation

5
INSTRUCTION REPRESENTATION
 With in a computer, each instruction
is represented by sequence of bits
 Each instruction has a unique bit pattern

 Instruction divided into fields


 Corresponding to the constituent elements
of the
instruction
 E.g. A simple instruction format

 With most instruction set, more than one format is used


6
SIMPLE INSTRUCTION FORMAT

7
INSTRUCTION REPRESENTATION…
 Difficult for the programmers to deal binary
with representation of machine instruction

 Hence, it is common to use a symbolic representation


of machine instruction
 Opcodes
 Represented by abbreviations called mnemonics --- that indicate the
operation
 Operands
 Also represented symbolically
 Example
 ADD R, Y

 It is possible to write a machine language program in


symbolic form
A simple program accepts this symbolic input ,convert opcode
and operand references to binary form and construct binary 13
machine instruction
INSTRUCTION TYPES
 An instruction set must be functionally complete
 Sufficient enough to express any of the instructions
from a high level language

 Categories of instruction types


 Data processing
 Arithmetic and logic instructions
 Data storage
 Memory instructions
 Data movement
 I/O instructions
 Control
 Test and
14
branch
instructions
 Program flow
3.3 INSTRUCTION SET DESIGN
 One of the most interesting and analyzed aspects
of computer design
 A very complex task

 Because it affects so many aspects of the


computer system
 Instruction set
 Boundary where computer designer
Computer programmer can and view the
machine same
 Programmer’s means of controlling the CPU
 Programmers requirement must be considered in designing
the instruction set
15
INSTRUCTION SET DESIGN …
 Most important issues related the design of
instruction
to set includes:
 Operation repertoire
 How many ops?
 What can they do?
 How complex are they?
 Data Types
 Built-In Data types supported
 Instruction formats
 How to encode as binary values
 Length of op code field
 Number of addresses
 Registers
 Number of CPU registers available
 Which operations can be performed on which registers? 16
 Addressing modes
TYPES OF OPERAND
 Machine instructions operate on data
 Categories of data
 Numbers
 Signed Integer / Unsigned Integer / Floating Point

 Characters

 ASCII, Unicode etc.

 Logical Data

 Bits or flags

 A group of bit where each bit has information

12
TYPES OF OPERATIONS
 Different machines support different opcodes:
 Data Transfer
 I/O
 Arithmetic
 Logical
 Conversion
 System Control
 Transfer of Control

13
DATA TRANSFER
 Most fundamental type of instructions
 Must specify the following things
 Location of source and destination operands
 memory, register …
 Length of data to be transferred
 Full word, half word …
 Addressing mode of each operand
 E.g.
 Move, Store, Load, Push, Pop
 In terms of CPU action
 data transfer operations are the simplest type
14
ARITHMETIC
 Performed by ALU
 Basic arithmetic operations provided by
most
machines are:
 Add, Subtract, Multiply, Divide
 Signed Integer
 Floating point

 May include
 Increment
 Decrement

 Negate
Absolute
15
LOGICAL
 Operations that manipulate individual bits of a word
 Bitwise operations
 AND, OR, exclusive-OR (XOR)
 NOT (one’s complement)
 In addition a variety of shifting and rotating functions
 Logical shift
 Shifts bits of word either to the left or right
 On one end the bit shifted out is lost
 Arithmetic shift
 Treats the data as a signed integer and does not shift the sign
bit
 Rotate/Cyclic shift
 Preserves all of the bits being operated on
21
SHIFT AND ROTATE OPERATIONS
0

logical
shift
shift in 0

arithme
tic
shift
keep sign !

rotate
17
CONVERSION
 Change the format of data or operate on
the format of data
 Binary to Decimal
 ASCII to EBCDIC

 E.g. Binary to BCD


 binary 00001111 (1510)
 packed BCD 0001 0101
1

18
INPUT/OUTPUT
 Transfer data
 from the computer to peripheral devices
 From peripheral devices to the computer system

 To be seen in detail in another chapter

19
SYSTEM CONTROL
 Reserved for use by the operating system
 Instructions executed while the processor is
in
certain privileged mode
 Privileged instructions
 CPU needs to be in specific state

20
TRANSFER OF CONTROL
 Alter the flow of program
 i.e. change the sequence of instruction execution
 Update the PC to a specific address
 Most common such type of operations are:
 Branch, Skip, Procedure call

21
TRANSFER OF CONTROL…
 Branch
 Also called jump instruction
 Its operand is the address of the next instruction to be fetched
and executed
 Two types of branches
 Conditional and Unconditional

 Conditional Branch
 A branch is made if certain condition is met
 E.g.
 BRP X
 Branch to instruction at location X if result is positive
 BRZ X
 Branch to instruction at location X if result is zero
 BRE R1,R2,X
 Branch to X if contents of R1 is equal to contents of
R2
 Unconditional branch
 E.g. 27
 BR X
 Branch to instruction at location X
TRANSFER OF CONTROL…
Branch Instructions

28
TRANSFER OF CONTROL …
 Skip
 Implies that the next instruction be skipped
 Contains an implied address
 Can be of two types:
 Unconditional
 Skip (i.e. unconditionally increment PC to skip the next

instruction)
 Conditional

 Test some condition and skip if met/satisfied

 E.g. Increment and skip next instruction if result is

zero ISZ R1  increment & skip if zero


Branch X  skip this if result was zero
ADD A
 Procedure call
 [Reading 29
Assignment]
ADDRESSING MODES
 The manner in which each address field
specify operand location

 Notations:
 A = Contents of an address field in the instruction
 R = Contents of an address field in the instruction that refers to a
register
 (X) = Contents of memory location X or register X
 EA =Effective address of the location containing the referenced
operand

25
ADDRESSING MODES…
 Types of addressing modes
 Immediate
 Direct
 Indirect
 Register
 Register Indirect
 Displacement
 Stack

26
IMMEDIATE ADDRESSING
 The instruction itself contains the operands value
 Operand is part of instruction

 Operand = A
 E.g. ADD 5
 Add 5 to contents of accumulator
 5 is operand

 No additional memory reference required


after the fetch of the instruction itself
 The value that can be specified is limited
 Size/value of the operand is limited/limited range
 Fast 27
IMMEDIATE ADDRESSING DIAGRAM

Opcode Operand

28
DIRECT ADDRESSING
 Address field contains address of operand
 Effective address EA = A
 Operand = (A)
 E.g. ADD A
 Add contents of memory location A to accumulator
 One more memory access needed to fetch the operand
 No additional calculations required to work out
effective address
 Number of memory locations can be referenced,
that limited, due to limited width of
the field
 Common
 Limited addressgeneration
on earlier space of computers
29
DIRECT ADDRESSING DIAGRAM
Opcode A
Memory

Operand

30
INDIRECT ADDRESSING
 A (the address field) refers to a memory location
which contains the address of operand
 Multiple (two) memory accesses to find
operand
 Fetch EA
 Fetch operand
 Access the memory twice, hence slower
 EA = (A)
 Operand = ((A))
 E.g. ADD (A)
 Look in A, find address (A) and look there for operand
 Range of EA increased
36
 Large address space
 2n ,where n = word length
INDIRECT ADDRESSING DIAGRAM

Opcode A
Memory

Operand

32
REGISTER ADDRESSING
 Similar to direct addressing
 The address field refers to a register than a main memory address
 EA = R
 Operand = (R )
 Very small address field needed
 Shorter instructions
 Faster instruction fetch
 No memory access required
 Very fast execution
 Very limited address space
 Small number of registers can be referenced
33
REGISTER ADDRESSING DIAGRAM

Opcode R
Registers

Operand

39
REGISTER INDIRECT ADDRESSING
 Similar to indirect addressing mode
 Operand is in memory pointed to by contents
of
register R
 EA = (R)

 Operand =((R ))

 Large address space (2n)


 Where n is the width of the register
 Address space limitation overcome
 Uses one less memory access than
indirect addressing 35
REGISTER INDIRECT ADDRESSING
DIAGRAM
Opcode R
Memory

Registers

Operand

36
DISPLACEMENT ADDRESSING
 Very powerful addressing mode
 Combines
 direct addressing and
 Register indirect addressing
 Instruction needs to have two address fields
 At least one of which is explicit
 EA = A + (R)
 Address field hold two values
 A = base value
 R = register that holds displacement
 or vice versa 41
DISPLACEMENT ADDRESSING DIAGRAM

Instruction
Opcode R A
Memory

Registers

+ Operand

38
RELATIVE ADDRESSING
 A version of displacement addressing mode
 Implicitly references the PC register
 R = Program counter, PC
 EA = A + (PC)
 Saves bits

39
BASE-REGISTER ADDRESSING
 Referenced register contains memory address
 R holds pointer to base address

 Address field contains the displacement


from that address
 A holds displacement

 Exploits locality of memory references

40
INDEXED ADDRESSING
 Address field contains memory address
 A = base
 The referenced register contains a
displacement from that address
 R = displacement
 EA = A + ( R )
 Good for arrays and performing
accessing iterative
operations
 Auto indexing
 increment and

decrement 41
INDEXED ADDRESSING…
 Combinations
 Indirect addressing with
indexing
 Post indexing
 The indexing performed after the indirection
 EA = (A) + (R)
 Good to access a block of data of a fixed format
 Pre indexing
 The indexing performed before the indirection
 EA = (A+(R))
 Used to construct a multiway branch table 47
STACK ADDRESSING
 A stack
 The stack mode of addressing is a form of implied
addressing
 The machine instructions need not include a memory
reference but implicitly operate on the top of the
stack

43
INSTRUCTION FORMAT
 It defines the layout of the bits of an instruction,
in terms of its constituent parts

 More that one format used in a given instruction


set

 What are the factors affecting the


instruction format ?
 READING ASSIGNMENT!!!

44
3.4 INSTRUCTION CYCLE REVISITED
 Instruction cycle so far consists of the
following sub cycles:
 Fetch
 Execute
 Interrupt

 Additional sub cycle to be added:


 Indirect cycle

45
INSTRUCTION CYCLE ,REVISED …
 Indirect addressing
 May require memory access to fetch operands
 Indirect addressing requires more memory accesses
 Can be thought of as additional instruction subcycle

46
3.5 INSTRUCTION CYCLE DATA FLOW
 The exact sequence of events during an
instruction cycle depends on the design of the
CPU

 In general terms, assuming a CPU that employs


the following registers
 MAR
 MBR
 PC
 IR
 The possible Data Flow is indicated next 47
DATA FLOW (INSTRUCTION FETCH)
 Fetch
 PC contains address of next instruction
 Address moved to MAR
 Address placed on address bus
 Control unit requests memory read
 Result placed on data bus, copied to MBR, then to IR
 Meanwhile PC incremented by 1

48
DATA FLOW (FETCH DIAGRAM)

49
DATA FLOW (DATA FETCH)
 IR is examined
 If indirect addressing, indirect cycle is performed
 Right most N bits of MBR transferred to MAR
 Control unit requests memory read
 Result (address of operand) moved to MBR

50
DATA FLOW (INDIRECT DIAGRAM)

51
DATA FLOW (EXECUTE)
 May take many forms
 Depends on instruction being executed

 May include
 Memory read/write
 Input/Output
 Register transfers
 ALU operations

52
DATA FLOW (INTERRUPT)
 Simple and Predictable
 Current PC saved to resumption after
allow interrupt
 Contents of PC copied to MBR
 Special memory location (e.g. stack
pointer) loaded to MAR
 MBR written to memory
 PC loaded with address of interrupt
handling routine
 Next instruction (first of interrupt handler)
can
be fetched 53
DATA FLOW (INTERRUPT DIAGRAM)

54
3.6 PROCESSOR STRUCTURE AND
FUNCTION
 Processor Organization
 To understand the organization of the CPU,
recall the requirements placed on the CPU:
 Fetch instruction
 Interpret instruction

 Fetch data

 Process data

 Write data

 To do all these things


 Processor needs to store data temporarily
 Temporary data storage locations --- Registers
--- are needed 60
PROCESSOR ORGANIZATION…
 Simplified view of a CPU
 Indicates its connection to the system via the system
bus

56
PROCESSOR ORGANIZATION…
 Detailed view of a CPU
 Indicates
 Data transfer and logic control paths
 Internal processor bus

 Used to transfer data b/n the ALU and various registers

57
REGISTER ORGANIZATION
 Registers
 CPU must have some working space
 Temporary storage
 Top level of memory hierarchy
 Number and function vary between processor designs
 One of the major design decisions

58
REGISTER ORGANIZATION…
 Registers classified into two groups:
 User visible registers:
 Can be referenced by assembly language instructions

 Control and status registers:


 Used by control unit to control the operation of the CPU

59
USER VISIBLE REGISTERS
 Can be categorized as follows:
 General purpose registers
 Data registers
 Address registers
 Condition codes

60
USER VISIBLE REGISTERS…
 General Purpose Registers
 Can be assigned variety of function by
the
programmer
 May be true general purpose
 May be restricted
 May be used for data or addressing

 Data Registers
 Used only to hold data
 Accumulator
61
USER VISIBLE REGISTERS…
 Address registers
 Hold addresses
 Can be
 General purpose or
 Devoted to particular addressing mode

 Index registers, stack pointer, segment register

62
CONDITION CODES
 Sets of individual bits , also called flags
 Each bit set by CPU hardware as the result
of
operations:
 E.g. Arithmetic operation results could be
 Positive, negative, zero, overflow
 Can be read (implicitly) by programs
 e.g. Jump if zero
 Can not (usually) be set by programs

63
CONTROL & STATUS REGISTERS
 Refers to different registers employed to
control the operation of the CPU
 Most not visible to the user

 Register essential during instruction cycle are


 Program Counter (PC)
 Instruction Register (IR)
 Memory Address Register (MAR)
 Memory Buffer Register (MBR)

64
CONTROL & STATUS REGISTERS…
 Program Status Word (PSW)
 A register that contain status information
 Usually contains
 Condition codes
 Status information
 Common fields or flags include
 Sign, zero, carry, equal, overflow, interrupt
supervisor, enable/disable

 Note that CPU design and Operating


system design are closely linked.
65

You might also like