Introduction to parallel processing
computing tech that involves simultaneously executing multiple
tasks in parallel
aims to divide
larger problem into smaller and solve them
concurrently
improves overall performance efficiency of amp system
can beapplied to indu processors w in single computer to
cluster
of interconnected comp
ability to process tasks simultaneously rather than sequentially
Adj
Improvedperformance reduced execution time
processing units
Increased scalability more
by adding
Enhanced
fault tolerance
Real time processing capability like simulations signal processing etc
Disady
, algos complex
PARALLEL PROCESSING IN UNI PROCESSOR
Icomponents
CPU
Main mem
I O
subsystem
a common synchronous bus architecture fr Ito m m
If
it is a
system w single prolesson perform two or more task
sim
Parallelism increases effiency reduces time of protessing
types Pipline Processing
not in syllabus
Array
Vector in J
and
Funclassifatinof comp
proposed
by Mikael J Flynn
categories comp architecture based on no
of instruction streams
data streams they can handle concurrently
, SISD
use Em
Single atiosingleData
Single stream of instruction operates on a
single
stream
of
data
represents tradional henna
CPU executes single stream inst
of on
single piece
of data atating
do inherent parallelism
Ihave
system falls into this class
Most conventional uniprocessor
SingleinstructioniMultipleDatafsinDI
executes single instruction on
multiple data elementssimultaneou
sanest applied to Lata simously
used in
vectyprocessing multi apps
GPL uses starch to
aubatation
computing tech that involves simultaneously executing multiple
tasks in parallel
aims to divide
larger problem into smaller and solve them
concurrently
improves overall performance efficiency of amp system
can beapplied to indu processors w in single computer to
cluster
of interconnected comp
ability to process tasks simultaneously rather than sequentially
Adj
Improvedperformance reduced execution time
processing units
Increased scalability more
by adding
Enhanced
fault tolerance
Real time processing capability like simulations signal processing etc
Disady
, algos complex
PARALLEL PROCESSING IN UNI PROCESSOR
Icomponents
CPU
Main mem
I O
subsystem
a common synchronous bus architecture fr Ito m m
If
it is a
system w single prolesson perform two or more task
sim
Parallelism increases effiency reduces time of protessing
types Pipline Processing
not in syllabus
Array
Vector in J
and
Funclassifatinof comp
proposed
by Mikael J Flynn
categories comp architecture based on no
of instruction streams
data streams they can handle concurrently
, SISD
use Em
Single atiosingleData
Single stream of instruction operates on a
single
stream
of
data
represents tradional henna
CPU executes single stream inst
of on
single piece
of data atating
do inherent parallelism
Ihave
system falls into this class
Most conventional uniprocessor
SingleinstructioniMultipleDatafsinDI
executes single instruction on
multiple data elementssimultaneou
sanest applied to Lata simously
used in
vectyprocessing multi apps
GPL uses starch to
aubatation
