Introduction to Parallel Approach Problems:
When different processors work try to solve a problem by working simultaneously on different problems. They can set range from one to million. It requires correct algorithms, well programs, compilers and operating systems. It is used because it reduces the time of computation. In parallel processing we can select our own architecture to solve any problem. This decreases the solution time enormous time. There are different methods available:
SISD (single instruction single data): it has one single processor connected to control and memory in series, with instruction and data stream.
MISD (multiple instruction single data): it has n number of processors connected to memory and every processor is connected to control unit with instruction stream.
SIMD (single instruction multiple data): it has n number of processor connected to the memory but in this case all processors are connected to single control unit.
MIMD (multiple instructions multiple data): it has n number of processors connected to memory and every processor is connected to its own control unit. It means n number of processor is connected to n number of control unit.
Speed: this technique is basically for speed. Speed is calculated by dividing worst case algorithm with worst case running time of program. Much the speed betters the parallel algorithm.
Numerical problems are solved in different manner like matrix transposes SIMD gauss Jordon, SIMD root search, computing Eigen values.
Conclusion: it is used in real time systems that provide solution for our real life world. For many simultaneous problems, parallel is the best one. It has beaten all the technology barriers.
Download A Technical Seminar Topic on Parallel Approach Problems .