Introduction to Numerical Problem a Parallel Approach Topic Presentation:
We know that in sequential processing we need to process the data sequentially that is one by one which is very time consuming and in sequential approach there is a lack of interprocessor communication. Parallel processing is solving a problem by working at a time on different parts of that problem. We require algorithms, programming languages, operating systems, compilers and so many numbers of processors that may be in tens or in millions also. Parallel processing reduces the time taking to solve the problems.
1. Single instruction stream, single data stream (SISD) Computer:
In SISD computer we have only one control and one processor instruction stream is sent to processor from the control and processed the data stream is sent to the memory.
2. Multiple instruction stream, single data stream (MISD) computer:
In MISD computers we have number of processors and controls. A single data stream is sent from the memory to the processor then numbers of instruction stream are sent to the controls.
3. Single instruction stream, multiple data stream (SIMD) computer:
In SIMD computer we have number of processors and a single control. Numbers of data streams are sent to the processors from the memory and a single instruction stream is sent to all the processors from the single control.
4. Multiple instruction stream, multiple data stream (MIMD) computer:
In MIMD we have number of processors and controls. Numbers of data streams are sent from the shared memory or interconnection network to the processor and numbers of instructon streams are sent to the processor from the number of controls.
In parallel processing we use number of parallel algorithms. A parallel algorithm is a solution method to be solved on a parallel computer. The quality of the parallel algorithm depends on the speedup it produces. The larger the speedup, the better the parallel algorithm.
Download Numerical Problem a Parallel Approach Topic for CSE Technical Seminar with PPT.