Parallel computation using GPU with architectures CUDA and OpenCL

We use GPGPU technology to develop software that allows to increase computation capacity by using Graphics Processing Unit. CUDA and OpenCL architectures are used in various areas including video stream processing, astrophysics, computational biology and chemistry, computational fluid dynamics, electromagnetic coupling, image registration in tomography, seismic analysis, ray tracing etc.

However GPGPU application always requires thorough reasoning. Exempli gratia for artificial neural networks training we used CUDA architecture and ran computational experiments and optimization approaches (paper in Russian).

Realtime 3D face reproduction using GPGPU:

Nvidia CUDA realtime simulation: