GPU Graphics Processing Unit
GPU stands for graphics processing unit and is a relatively new mechanism used for parallel approaches in order to analyse big data. This is particular the case for data parallelism and task parallelism. In comparison to standard multi-core CPUs the GPUs consist of a many-core architecture with hundreds to even thousands of very simple cores. Simple means that they have a high throughput computing-oriented architecture using massive parallelism by executing a lot of concurrent threads rather slowly. A CPU instead typically executes a single long thread as fast as possible. Parallelism means that many-core GPUs are able to handle an ever increasing amount of multiple instruction threads.They are used in large clusters and within massively parallel supercomputers today.
Since GPU started to emerge they were considered as graphics coprocessor or accelerator that have been mounted on a graphics card or video card of a normal desktop computer. The first GPU was available as Geforce 256 from NVidea in 1999 and NVidea dominated the market ever since. The initial key goal of a GPU was to offload the CPU from numerous graphics tasks that often exist in graphics or video editing applications. The reason is because GPU chips are able to process many million of polygons per second. More recently, general-purpose computing on GPUs (GPGPUs) emerged that is used within high performance computing environments. This GPU computing drives many advances in various business and scientific application fields with an ever increasing performance but reduced power requirements.
More details about GPU
We refer to the interesting following video in order to obtain more pieces of information: