Parallel Processing
Parallel processing is an efficient and effective approach to work with big data. It performs computing with a number of ‘compute elements’ (e.g. cores) that solve a problem in a cooperative way. The goal is to tackle scientific, engineering, and business related problems in parallel in order to enable the ‘best performance’ possible. Especially supercomputers and large computing clusters depend heavily on parallelism for applications today. The parallelization can be twofold. In High Performance Computing (HPC) several cores in one big supercomputer or large cluster work in parallel on a problem with many message exchanges between them. In High Throughput Computing (HTC) several computing systems are part of a larger distributed computing network and solve problems in parallel with a limited number of message exchanges between them. Both of these parallel approaches are used often in the context of big data processing.
Parallel Applications
Scientific applications are typically created by using parallel software programming according to numerical models and known physical laws. This often includes the intensive re-use of proven mathematical and physical libraries and various Compilers specific for parallel systems. The Goal is to perform computational-intensive simulations of ‘real world phenomena’ in parallel on many cores. Often application domains have requirements on storage, main memory, or computational speed that are typically not met with standard Desktop PCs. One prominent example is Numerical Weather Prediction (NWP) that simulates the future weather and is daily used for predictions in weather forecast systems. Another famous example is the study of earthquakes using parallel processing over a large area of the earth to understand its source and impact. Other examples are Computational Fluid Dynamics (CFD), pharmaceutical drug design, or theoretical solid state physics. There is a wide variety of application areas. Examples include quantum chromodynamics, materials science, structural mechanics, and medical image processing.
More Information about parallel processing
The following video provides a short introduction to this topic:
Follow us on Facebook:
