SRINIVAS

• Apr 2nd, 2005

 

There are 3 kinds of Parallelism: 
1) Data Parallesim  
2)Componnent Paralelism 
3) Pipeline. 
 
When the data is divided into smalll chunks and processed on different components simultaneously we call it DataParallelism 
 
When different componens work on different dataa sets it is called Component parallelism 
 
When a graph uses multiple components to run on the same data simltaneously we call it Pipeline parallelism

malleswara reddy

• Oct 17th, 2005

 

SRINIVAS Wrote: There are 3 kinds of Parallelism: 
1) Data Parallesim  
2)Componnent Paralelism 
3) Pipeline. 
 
When the data is divided into smalll chunks and processed on different components simultaneously we call it DataParallelism 
 
When different componens work on different dataa sets it is called Component parallelism 
 
When a graph uses multiple components to run on the same data simltaneously we call it Pipeline parallelism

Ankita1

 Profile Answers by Ankita1 

• Jul 6th, 2006

 

The 3 kinds of parallelism are:

Component parallelism: occurs when program components executes simuteneously on diff branches of the graph.

Pipeline parallelism: occurs when connected sequence of program components

executes simulteneously on same branch of the graph.

Data parallelism: occurs when you seperate data in to multiple divisions,allowing multiple copies of pogram components to operate on data in all the divisions simulteneously.

sixtodsilva

• Sep 18th, 2007

 

Three kind of parallelisms

1.Component Level Parallelism --> different components acting on different data at same time.

2.Data Parallelism --> A graph that deals with data divided into segments and operates on each segement

3. Pipeline parallelism --> A graph with multiple components running simultaneously on the same data .

JYOTI

• Dec 31st, 2017

 

Pipeline parallelism
Data is pushed from component to component, via OS pipes, avoiding writing to disk.
Component Parallelism
Within the same flow, different sub-flows can execute at the same time performing different tasks, making best profit of the available CPU.
Data Parallelism
The same flow can contain multiple streams of data. This activity is backed up by a multiple file system. (i.e. End 2 End parallelism supported).