Specifically, GPUs are designed to solve problems that can be expressed as data-parallel computations—programs executed in parallel on many data elements with extremely high computational density (ratio of mathematical operations to memory operations). Since all data elements execute the same program, there is less demand for precise flow control; and because it operates on many data elements and has a high computational density, memory access latency can be hidden through computation without having to use large data caches.

Data parallel processing maps data elements to parallel processing threads. Many applications that process large data sets can use data parallel programming models to accelerate computation. In 3D rendering, large sets of pixels and vertices are mapped to parallel threads. Similarly, image and media processing applications (such as post-processing of rendered images, video encoding and decoding, image scaling, stereo vision, and pattern recognition) can map image blocks and pixels to parallel processing threads. In fact, many algorithms outside the field of image rendering and processing are also accelerated by data parallel processing – from general signal processing or physical simulation to mathematical finance or mathematical biology. In the above fields, GPU computing has been successfully applied and achieved incredible acceleration results.