Introduction to Technical Seminar on Mesh Algorithms in Parallel Processing:
The paper is about parallel processing. Parallel processing is defined as processing of the tasks with several processing units or with number of processors. Generally the powerful computer for sharing the jobs contains more processing units.
Operations which are done simultaneously take less time for computation so that more tasks can be done. Generally a task is divided into subtasks. And the operations of subtasks are done at the same time on different processors. These subtasks are then joined to form a task i.e. a solution to the original problem. Thus we can conclude that parallel processing is used for increasing speed.
The speedup equation is given by speedup = ws/wp where ws denotes the worst case running time of fastest sequential algorithm and wp denotes the worst case running time of the parallel algorithm for the same problem.
A mesh algorithm is defined as, a x b grid which has processor at each grid point. And each processor of mesh is defined as a tuple (i, j),where 1 £ i £ a and 1 £ j £ b. Also each processor will be having a RAM with local memory. At the same time each processor is capable of making calculations related to manipulation like addition, in one unit of time.
The packet routing is defined as the communication with in a process in a fixed connection network. Where the processor has information regarding the other processor it has to be delivered to.
We can conclude that, parallel processing is used in research and development areas. The mesh algorithms discussed in the paper increases the performance of the parallel algorithms.
Download IT Technical Seminar on Mesh Algorithms in Parallel Processing with PPT.