Parallel Computing : Effectiveness and Classes

Parallel Computing implements architectures of digital electronics and algorithms specialized for them to process information simultaneously.The purpose of Parallel Computing is to perform, with a machine, the largest number of operations in the smallest possible time. In order to do this, the operations must be done in parallel, that is to say simultaneously by several processing units.

The task at hand is broken down into multiple sub-tasks those are executed at the same time making up each of the parallel architectures in Parallel Computing. Parallel Computing, as defined, applies to multiple types but without specification it is to parallelize the processing units.

 

Effectiveness of Parallel Computing

 

In order Parallel Computing to be effective, the methods used for programming tasks in Parallel Computing that constitute a program specific to this method of calculation, must be made with this in mind. These methods were originally developed on sophisticated machines – from supercomputers.

In addition, Parallel Computing architectures have become the dominant paradigm for all computers from 2000. Indeed, the processing speed that is related to the increased frequency of processors has its limits due to increase heat production that causes errors in calculations.
The creation of multi core processors, treating multiple instructions simultaneously within the same component, solves this dilemma for office machines since the mid-2000.

 

Parallel Computing and Parallel Computers can be roughly classified according to the level at which the hardware level supports parallelism. On the one hand, there are common machines with either multi-core machines or multiprocessor, massively parallelized machines and structures formed from grid computing, thousands of ordinary computers, not vested in this particular task and connected by a network.

 

Types or Classes of Parallel Computing

 

• Bit level Parallel Computing
• Instruction level Parallel Computing
• Data parallelism
• Task parallelism

 

Usage of Parallel Computing

 

Achievement of adequate programs on architectures designed for Parallel Computing helps in weather predictions, modeling and simulation of problems of larger dimensions, Medical science and Health research, image processing and computer graphics rendering, artificial intelligence and automated manufacturing.

 

Further Reading on Parallel Computing : Introduction to Parallel Computing by Blaise Barney, Lawrence Livermore National Laboratory.