You are on page 1of 5

Submitted to: Sir Irfan Ali

Name: Arsalan Shakeel

Roll No: 14CS06 5th semester

Curse name: Computer Architecture

Flynn Taxonomy

About 50 years ago 1966 Michael J.Flynn gives his perception about the categories of execution in
system. He said that only two types of information is flow in system one is data and another one is
instructions. So the reasons are the causes of data hazard we can solve it easily in productive way of
architecture in system. His entire theory is based on parallel instruction streaming and parallel data
streaming. Basically he improves the architecture of computer system in order to type of execution
which is going to be occurring in system. The perception of Michael J.Flynn was just want to make the
execution of system more and more advance and smarter than all the systems which had been launch
before it.

His theory was based on the categories of execution occur in system. And the information flow in system
what type intention it had after giving his theory about system is still using this phenomenon in
architecture of computer. According to him sometime we have more than one set of instruction
streaming is executes in system. And the Michael J.Flynn was the one who gives the multiple processing
executions in the system and for this purpose he said that we need to improve our data streaming
segment in the architecture of computer because if we want to execute more than one instruction set
then we need to use this type of strategy for system. He has further said that if we always utilized
multiple processors for the execution of single instruction that is not beneficial for the system health
and it would lose of power system performance system executions segments. So we need to solve also
this type of endurance in our processing unit.

Michael J.Flynn also gives the perception about the single data streaming can handle multiple
instruction stream at the same time but it is depend upon the size and the type and the intensity of that
instruction streams. So he was not only focus on the single or multiples instruction execute in system
but he was also focuses on the need of data streaming according to the use of that particular set of
instruction stream going to be executing in system.

Michael J.Flynn also gives his perception about the intensity, power, usability and their effect of
execution to the entire system and whole architecture of system. Which means that he have some a
time just a single instruction stream but their intensity, power, usability and their effect of execution to
the entire system has very high chances of create deadlock in the system. So we need to improve out
system’s data streaming section for this we can use more powerful data streaming processor on the
other hand we use to multiple data streaming for that particular heavy instruction stream. But now
have mostly use multiple data stream because it supportable for both types of instruction stream on
the system and according to survery in 2006 this is declared that the maximum types of computer
manufacture in different corners of this world and made on the bases of Flynn Taxonomy in the entire
world. The Taxonomy of Flynn and Basically the standard of level he had defined in their theory is was
categories in the classification of Flynn’s Taxonomy.

Classification’s of Flynn Taxonomy

i) SISD (Single Instruction Single Data)

In this classification of Flynn our whole system is execution is based on the single instruction stream is
fetch from the memory and for this single instruction set stream we use only one data streamer. Some it
is slower than other classification process of execution in the system. We have a example of SISD we
often saw in our daily life that is Structure of Von Neauman which is basically made up of single
instruction stream and against of this single data streaming which has just capacity of one instruction set
one data stream on the bases of this way of architecture it use single instruction set and simple data
stream in whole process of execution in the CPU. When we are working on array of GPU system we
always face SISD classification because it is not capability of is works on multiple set of instruction.
ii) SIMD (Single Instruction Multiple Data)

In this classification of Flynn the way of execution in entire system is much different than SISD
classification in this classification system uses multiple processor for single set of instruction we have.
This classification gives result much faster than SISM classification and performance has brilliantly better
SISD. Their way of execution is really faster because the as we know that why it is much brilliant, faster,
smarter and swifter than SISD because SISD we have jus t one data stream and here we have multiple
data stream we can utilize our instruction in multiple processor to improve pipelining way of execution
in system. We have example of SIMD (Single Instruction Multiple Data) is Vector Faster computer. It has
using multiple processors for single set of instruction fetch from memory the reason behind this,
stupendous, fabulous and fantastic performance is that is has only working on a single set of instruction
and for their executions it has really powerful multiple data streaming execution.
iii) MISD (Multiple Instruction Single Data)

Third classification of Michael Flynn theory which is really rarely used in manufacturing of computer.
This classification is for particular places of work as per need of work which is not so popular in market
and any other place of architect of computer system development but we could not say that this is not
touch professional people. It is using but not in that way as other are running in different aspect of
computer architecture development. We have a example of MISD is Arguably none.

iv) MIMD (Multiple Instruction Multiple Data)


The final and the most popular classification of Flynn taxonomy MIMD in this classification we have
privilege of run more than one execution set of instruction fetch from memory and against of it we have
multiple data streams for execute it. Our working strategy in that system is really smarter and the
performance in execution is really faster we are successful in enhancement of pipelining and many
other aspect of execution process. We further divide this MIMD in two classifications Single program,
multiple data streams (SPMD) and as well as Multiple program, multiple data streams (MPMD). We have
example of MIMD is multiple processor and multiple computers.

You might also like