Sie sind auf Seite 1von 28

DEPARTMENT OF TECHNICAL EDUCATION

ANDHRA PRADESH
Name : V BAPUJI NAIK
Designation : LECTURER
Branch : COMPUTER ENGINEERING
Institute : GOVT. POLYTECHNIC, PRODDATUR
Year/Semester : III SEMESTER
Subject : COMPUTER ARCHITECTURE
Code No : CM-303
Major Topic : PIPELINE & VECTOR PROCESSING
Duration : 150 Mins.
Sub Topic : VECTOR PROCESSING & ARRAY
PROCESSING
Teaching Aids : ANIMATIONS, DIAGRAMS
CM303.66to68 1
OBJECTIVE

On completion of this period, you would be able to


understand

• Concept of Parallel Processing


• Types of Parallel Processors
• Description of various Parallel Processors

CM303.66to68 2
What is Parallel Processing?

• Parallel means simultaneous or concurrent


i.e happening at a time

• Processing means Executing or Computing


i.e doing

CM303.66to68 3
Parallel Processing
(Cont..)

So Parallel Processing executes multiple


tasks at the same time to speed up
computations

CM303.66to68 4
Where do we come across such work in general?

• A 10 pages of text is given for typing.

• How does one can plan to complete the typing


faster?
• 10 typists are asked to type all at the same time
each one page.

CM303.66to68 5
Parallelism can be viewed at two levels
Instructional Level Parallelism
• Executing multiple instructions concurrently from the
same program
Processor level Parallelism
• Sharing workload(tasks) by multiple processors within a
single computer system
• distributing workload(tasks) among multiple computer
systems on a network

CM303.66to68 6
Instances of Parallel Processing

• Parallel events may occur in multiple resources


during the same time of intervals

• Simultaneous events may occur at the same


time instant

• Pipelined events may occur in overlapped time


spans

CM303.66to68 7
What is the ultimate benefit of Parallel
Processing?

To speed up computer processing capability at


lower cost

CM303.66to68 8
Classification of Parallel Processors

According to Michael J Flynn, Parallel


Processors can be four types, based on the multiplicity
of instruction streams and data streams in a computer
system
• An Instruction stream is a sequence of instructions
as executed by the machine
• A data stream is a sequence of data including input,
temporary results, called for by the Instruction
stream

CM303.66to68 9
Flynn’s classification of Parallel Processors

1. Single Instruction stream – Single Data stream ( SISD)


2. Single Instruction stream – Multiple Data stream ( SIMD)
3. Multiple Instruction stream – Single Data stream ( MISD)
4. Multiple Instruction stream –Multiple Data stream( MIMD)

CM303.66to68 10
Single Instruction stream – Single Data stream
( SISD)
CP : Control Unit MM : Memory Module
PU : Processor Unit IS : Instruction Stream
DS : Data Stream

IS

CU IS DS
PU MM

CM303.66to68 11
Block diagrams illustrates
• Four types of System Components are needed:
1. Control Unit(CU)
2. Processor Unit(PU)
3. Memory Modules(MM)
4. I/O Modules(IOM)
• Instructions & data are fetched from the Memory Modules
• The instructions are decoded by the CU, & sent to the PU for
execution
• Data streams flow between Processors and memory bi-
directionally
CM303.66to68 12
Block diagrams illustrates
(Contd..)

• Multiple memory modules may be used in the shared


memory subsystem
• Each instruction stream is generated by an independent
control unit
• Multiple data streams originate from the subsystem of
shared memory modules
• I/O facilities are not shown in these simplified block
diagrams
CM303.66to68 13
SISD Computer Organization

• It is a traditional Uniprocessor

• It has a single control unit & a single execution unit but


may be more than one functional unit
• Instructions are executed sequentially but may be
overlapped in their execution stage(pipeline)

Examples:
One Functional Unit: IBM 701(1), PDP VAX11/780(1)
Multiple Functional Units: CDC 6600(1), Cray-1(4)
CM303.66to68 14
Single Instruction stream – Multiple Data
stream ( SIMD)
SM : Shared Memory

DS1
PU1 MM1

DS2
PU2 MM2
CU
IS
o SM o
o o
DSn
PUn MMm

IS
CM303.66to68 15
SIMD Computer Organization
• It is an Array Processor type
• Multiple Processing Elements(PE) are supervised by
the same control unit
• All PEs receive the same instruction broadcast from the
control unit but operate on different data sets from
distinct data streams.
• The shared memory subsystem may contain multiple
modules.

• Examples: BSP(6), STARAN(1), DAP(1)

CM303.66to68 16
Multiple Instruction stream – Single Data
stream ( MISD)
IS1

IS1
CU1 PU1 DS

IS2
CU2 PU2 SM
o o MM1 MM2 o o MMm
o o
ISn
CUn PUn

DS
ISn
IS2

CM303.66to68 17
MISD Computer Organization

• Practically this type of system is not feasible

• N Processors receive distinct instructions


operating over the same data
• The result of one processor become the input of
the next processor

CM303.66to68 18
Multiple Instruction stream – Multiple Data
stream ( MIMD)
IS1

IS1 DS1
CU1 PU1 MM1
DS2
IS2
CU2 PU2 MM2

o o SM o
o o o
DSn
ISn MMm
CUn PUn

IS2 ISn

CM303.66to68 19
MIMD Computer Organization

• Most Multiprocessor Systems & Multiple computer


systems are this type.
• N Processors & n control units are there

• Interaction among processors is possible because all


memory streams are derived from the same data space
shard by all processors
• Examples: IBM 370/168 MP(9), Univac 1100/80(9), Cray-
2(9), Cray-x MP(9).
CM303.66to68 20
Summary
• Parallel Processing executes multiple tasks at the same
time to speed up computations
• Parallelism can be viewed at two levels - Instructional Level
& Processor level
• Flynn’s classified the parallel processors into four types
 Single Instruction stream Single Data stream ( SISD)
 Single Instruction stream Multiple Data stream ( SIMD)
 Multiple Instruction stream Single Data stream ( MISD)
 Multiple Instruction stream Multiple Data stream ( MIMD)

CM303.66to68 21
Summary

(Cont..)

• SISD is a traditional Uniprocessor


• SIMD is an Array Processor type
• Practically MISD of system is not feasible
• Most Multiprocessor Systems & Multiple
computer systems are of the MIMD type

CM303.66to68 22
Frequently Asked Questions
4 Marks
1. Define Parallel Processing

2. Write the advantages of Parallel


Processing.

3. Write the classification of the Flynn’s


Parallel Processors.

CM303.66to68 23
Frequently Asked Questions
(Cont..)

6 Marks
1. Write briefly on the Flynn’s classification of
Parallel Processors

12 Marks
1. Explain the Flynn’s classification of Parallel
Processors

CM303.66to68 24
Quiz
1. Which of the following are right with Parallel
Processing?

c) It is the simultaneous execution of


programs
d) It is the execution of instructions overlapped
within a program
e) Execution of a program can be distributed
over a set of processors
f) None of the above

CM303.66to68 25
Quiz
Contd..
2. Which of the following are of the
uniprocessor systems ?

c) SISD
d) SIMD
e) MISD
f) MIMD

CM303.66to68 26
Quiz
Contd..
3. Which of the following are not practically
feasible ?
a) SISD
b) SIMD
c) MISD
d) MIMD

CM303.66to68 27
Quiz
Contd..

4. Which of the following have the MIMD


Computer Organization ?

a) DAP(1)
b) Cray-2(9)
c) IBM 701(1)
d) None of the above

CM303.66to68 28

Das könnte Ihnen auch gefallen