GPU parallel computing architecture and CUDA programming model

This article consists of a collection of slides from the author's conference presentation on NVIDIA's CUDA programming model (parallel computing platform and application programming interface) via graphical processing units (GPU). Some of the specific topics discussed include: the special features of GPUs; the importance of GPU computing; system specifications and architectures; processing capabilities; parallel memory sharing; CUDA programming models; transparent scalability; and major applications supported.