Upload
aashita-irani
View
217
Download
0
Embed Size (px)
Citation preview
8/8/2019 03 System Architecture
1/19
ELE 3230 - Part 3 1
Par t 3
8088 Sys tem Arc h it ec t u re
(Hal l : c h2; Brey : c h2; Tr iebel : c h2)
ELE 3230Mic roproc essors and Com put er Sys tems
8/8/2019 03 System Architecture
2/19
ELE 3230 - Part 3 2
His t o r i c a l Bac k ground
1969/70 Intel 4004, first Microprocessor (M.E.Hoff) 4 bit microprocessor, originallydeveloped for Busicom, a small Japanese calculator company. Limited to 4096 memorylocation (of 4 bit data), 45 instructions; integrated 2300 PMOS transistors.
1971 Intel 8008, first 8 bit microprocessor (16K x 8bit)
1973 Intel 8080, 10 x faster than 8008, more memory (64k)
1974 Other 8 bit processors: Motorola 6800, Fairchild F8,
1975 Signetic 2650, MOS Technology 6502 (used in Apple II), Rockwell PPS-8
1976 National IMP-PACE, first 16 bit microprocessor, followed by Texas Instrument,TMS9900, Zilog Z80 (8 bit) (used in Radio Shack TRS-80)
1977 Intel 8085 (8080 with built-in clock & system controller) 1978 Motorola 6809 (8 bit), Intel 8086 (16 bit processor, 1M)
1978/79 Intel 8088 - variant of 8086 with 8 external data pins
1981 IBM adopts Intel 8088 for IBM PC/XT
(see http://bwrc.eecs.berkeley.edu/CIC/archive/cpu_history.html )
2 most popular microprocessor series:
INTEL 8086, 80186, 80286, 80386, 80486, Pentium
Motorola 68000, 68010, 68020, 68030, 68040
8/8/2019 03 System Architecture
3/19
ELE 3230 - Part 3 3
Mic roproc essor Com put er Sys t em
CentralProcessing
Unit (CPU)
InputOutput
Unit
(I/O Unit)
Memory
Data Bus
Address Bus
Control Bus
Control Bus
Processor Bus
Co-Processor(e.g. Floating point unit)
Control Unit (CU)
Registers
Arithmetic & LogicUnit (ALU)
Data Bus
Address Bus
8/8/2019 03 System Architecture
4/19
ELE 3230 - Part 3 4
Mic roproc essor Com put er Sys t em
Control Unit (CU) generates all the control signals within theCPU. It initializes the registers on power-up, generates thesignal to fetch instructions for the ALU.
The Control unit may be implemented (i) completely byhardware (hard-wired controller e.g. using a state counterand a Programmable Logic Array) or (ii) by a mixture of
software instructions (microcode stored in CPU) andhardware (microprogrammed control). Both the Intel 8086family and Motorola 68000 family use microprogrammedcontrollers.
Registers - small, fast memory which usually store data andaddresses associated with the instruction being carried out.
ALU performs arithmetic and logic operations
CentralProcessing
Unit (CPU)
Control Unit (CU)Control Unit (CU)
RegistersRegisters
Arithmetic & LogicUnit (ALU)
Arithmetic & LogicUnit (ALU)
8/8/2019 03 System Architecture
5/19
ELE 3230 - Part 3 5
Ins t ruc t ion Ex ec ut ion Cyc le
Two main steps in the cycle:
1. Fetch the next instruction from main memory
2. Decode and Execute the instructionThe Fetch cycle consists of
i) use the instruction pointer (IP) to set the address bus with the
address of the next instruction and increment the instruction pointer;
ii) wait (few hundred nanoseconds) for data to be transferred to the
data bus from memory; and
iii) read the data from the data bus.
The Execution Cycle consists of
i) Decode the instruction and generate the correct sequence of internal
and external signals
ii) Execute the instruction and restart the Fetch Cycle
8/8/2019 03 System Architecture
6/19
ELE 3230 - Part 3 6
Bas ic Ins t ruc t ion Cyc le
HALT
Fetch the
NextInstruction
Executethe
Instruction
START
Fetch Cycle
Execute Cycle
8/8/2019 03 System Architecture
7/19
ELE 3230 - Part 3 7
Pipel ined Inst ruc t ion Fet c h and
Ex ec ut ion Cyc lesInstruction Fetch and Execution pipeline
Fetch Fetch FetchExecute Execute Execute Fetch
The Fetch and Execution are implemented by two process unitsinside CPU:
Bus Interface Unit (BIU) fetches instructions from memory, passes theinstruction to the instruction stream byte queue and starts to fetch thenext instruction immediately
Execution Unit (EU) removes instructions from the instruction queue
What are the advantages of allocating the functions to two separate units?
Both BIU and EU can be working simultaneously without waiting for thecompletion of the other task (pipelined parallel processing)
Is this efficient? How to improve it?
8/8/2019 03 System Architecture
8/19
ELE 3230 - Part 3 8
Tim ing Diagram for Inst ruc t ion
Pipel ine Operat ion
Instruction 1
Instruction 2
Instruction 3
Instruction 4
Instruction 5
Instruction 6
Instruction 7
Instruction 8
Instruction 9
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
FI DI CO FO EI WO
1 2 3 4 65 7 8 9 10 11 12 13 14
Time
I: InstructionO: Operand
Fetch Instruction Decode Instruction Check Operand Fetch Operand (if needed) Execute Instruction Write Output
(Operand is defined in the Instruction Set part)
8/8/2019 03 System Architecture
9/19
ELE 3230 - Part 3 9
In t roduc t ion t o In t e l 8086/8088
Microprocessors
8086 pin diagram
GND
AD14AD13AD12
AD10AD11
AD8AD9
AD6AD7
AD4
AD5
AD3AD2AD1AD0
NM1INTRCLK
GND
(ALE)
(HLDA)
(HOLD)
(WR)
(IO/M)(DT/R)
(DEN)
(INTA)
27
7
14
11
1
23
45
6
8
910
12
13
1615
17
1918
20
34
30
40
3938
37
36
35
33
3231
29
28
2526
24
2223
21
8086CPU
MAXMODE
{ }MIN
MODE
A19/S6
A17/S4A18/S5
A16/S3AD15
Vcc
READYRESET
BHE/S7MN/MX
RDRQ/GT0
RQ/GT1LOCK
S2S1
S0
QS1
QS0
TEST
8088 pin diagram
GND
A14A13A12
A10A11
A8A9
AD6AD7
AD4
AD5
AD3AD2AD1AD0
NM1INTRCLK
GND
ALE
HLDA
HOLD
A19/S6
A17/S4A18/S5
A16/S3A15
Vcc
READYRESET
SS0MN/MX
RD
WR
IO/MDT/R
DEN
INTATEST
27
7
14
11
1
23
45
6
8
910
12
13
1615
17
1918
20
34
30
40
3938
37
36
35
33
3231
29
28
2526
24
2223
21
8088CPU
MINMODE
{ }MAX
MODE
(HIGH)
(RQ/GT0)
(RQ/GT1)(LOCK)
(S2)(S1)
(S0)
(QS0)
(QS1)
Some pins have different functionsfor two different operation modes.
8/8/2019 03 System Architecture
10/19
ELE 3230 - Part 3 10
In t roduc t ion t o In t e l 8086/8088
Microprocessors8088 and 8086 are almost identical except that 8088 has only 8
external data lines whereas the 8086 has 16 external data lines.
Both have
16-bit wide data bus internally in microprocessor
20 address pins, including 16 address/data (AD0-AD15) + 4address/status (A16/S3-A19/S7) for 8086, allowing a maximummemory address range of 1MByte
multiplexed address/data pins (8088 only multiplexes 8 pins, AD0-
AD7) 2 modes of operation (the maximum and minimum modes)
Same instruction set
8/8/2019 03 System Architecture
11/19
ELE 3230 - Part 3 11
Int e rnal Arc h it ec t u re of t he 8088
Both 8088/8086 employ parallel processing.
Contain two processing units: Execution unit (EU) and Businterface unit(BIU); operate at the same time.
The BIU sends out addresses, fetches instructions frommemory, reads data from ports and memory, and writes datato ports and memory, i.e. the BIU handles all transfers of dataand addresses on the buses for the execution unit.
The EU tells BIU where to fetch instruction or data from,
decodes instructions, and executes instructions.
8/8/2019 03 System Architecture
12/19
ELE 3230 - Part 3 12
8086 In t erna l B loc k Diagram
8/8/2019 03 System Architecture
13/19
ELE 3230 - Part 3 13
Bus Int e rfac e Uni t (BIU)
Perform bus operation such as instruction fetching, reading/writing ofdata operand for memory, inputting/outputting data for I/O peripherals.
Perform other functions such as instruction queuing and dataacquisitions.
8-bit (16-bit) bi-directional data busfor 8088 (8086).
20-bit address bus can address any one of the 220 (1,048,576) byte-
memory .
Contain segment register, instruction pointer, address generation adder,bus control logic, and an instruction queue.
Use instruction queue to implement a pipelined architecture (prefetchup to 4 (6) bytes of instruction code for 8088 (8086) and then store andaccess the codes in FIFO order).
See Part-9 for the detail discussion on instruction set and segment
registers.
8/8/2019 03 System Architecture
14/19
ELE 3230 - Part 3 14
Ex ec ut ion Un it (EU)
Responsible for decoding and executing instruction.
Contains: arithmetic logicunit (ALU), status and control flags, general
purpose registers, and temporary-operand register.
EU accesses the instruction from output end of the instruction queueand data from general-purpose register.
It reads one instruction at a time, decodes them, generates operandaddress if necessary, passes them to BIU and requests to perform theread/write cycle to memory or I/O, and performs the operation specifiedby the instruction on operand.
During execution, EU may test the status and control flags and updatethese flags based on the results of execution.
8/8/2019 03 System Architecture
15/19
ELE 3230 - Part 3 15
Flag Regist er
The flag indicates the condition of the microprocessor as well ascontrols its operations.
A flag register is a flip-flop which indicates some conditionsproduced by the execution of an instruction or controls certainoperations of the EU. A 16-bit flag register in the EU contains nineactive flags. (Each flag occupies one bit in the flag register.)
Two types of flags:conditional flags: Six flags are conditional flags. They are set or reset
by the EU on the basis of the results of some arithmetic operation.
control flags: The three remaining flags in the flags register are used
to control certain operations of processor. They are called the controlflags.
8/8/2019 03 System Architecture
16/19
ELE 3230 - Part 3 16
Flag Regist ers
bit 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
U=Undefined Carry Flag (CF)- set by carry out of MSB.Parity Flag (PF)- set if result has even parity.
Auxiliary carry Flag (AF)- for BCDZero Flag (ZF)- set if results = 0
Sign Flag (SF) = MSB of result
Overflow Flag (OF)- overflow flag
IF- interrupt enable flag
DF- string direction flag
TF- single step trap flag
U U U U OF DF CFPFAFZFSFTFIF UUU
MSB: Most Significant Bit
See Examples in Part-6
Conditional Flags
Control Flags
8/8/2019 03 System Architecture
17/19
ELE 3230 - Part 3 17
Condi t ional Flags
carry flag (CF)- indicates a carry after addition or a borrow aftersubtraction, also indicates error conditions.
parity flag (PF)- is a logic 0 for odd parity and a logic 1 for even parity.
auxiliary carry flag (AF)- important for BCD addition and subtraction;holds a carry (borrow) after addition (subtraction) between bit-3 and bit-4.Only used for DAA and DAS instructions to adjust the value of AL after aBCD addition (subtraction).
zero flag (ZF)- indicates that the result of an arithmetic or logic operation
is zero. sign flag (SF)- indicates arithmetic sign of the result after an arithmetic
operation.
overflow flag (OF)- a condition that occurs when signed numbers areadded or subtracted. An overflow indicates that the result has exceeded
the capacity of the machine.
(See Part-6 for more details)
8/8/2019 03 System Architecture
18/19
ELE 3230 - Part 3 18
Cont ro l Flags
The control flags are deliberately set or reset with specificinstructions YOU put in your program. The three control flags are:
trap flag (TF) - used for single stepping through a program (fordebugging);
interrupt flag (IF) - used to allow or prohibit the interruption ofa program;
direction flag (DF) - used with string instructions.
No specific instruction to set TF. See example 12-1 in Breys formore details.
8/8/2019 03 System Architecture
19/19
ELE 3230 - Part 3 19
General -Purpose Reg ist ers
EU has eight 8-bit general-purpose registers, labeled AH, AL, BH, BL,CH, CL, DH, and DL. These registers can be used individually fortemporary storage of 8-bit data.
Register pairs AH-AL, BH-BL, CH-CL, and DH-DL can be used togetherto form register AX, BX, CX, and DX and can be used to store 16-bit datawords.
The AL register is also called the accumulator. It has some features that
the other general-purpose registers do not have.
The advantage of using internal registers is :
It can be accessed more quickly than from external memory. No memoryreference or memory cycle is needed to get the data.
(See Part-6 for more details)