We offer in-depth practical course on NVIDIA CUDA technology for developers and researchers who use parallel computing. In the first part of the series introduces the basics of CUDA programming model in language C and Fortran, information about the types of memory and GPU methods for the effective use of shared memory by the example of some algorithms. It then provides an overview of the main application libraries and language vredstv with embedded computing on GPU. Individual lectures are devoted to the elements of professional development - analysis tools, debugging and diagnostics. The methods of managing multiple GPU workstations and distributed cluster systems.

We offer to a listeners a practical course of NVIDIA CUDA technology for developers and researchers who use parallel computing . The basics of programming using technology CUDA, information about the types of memory GPU, especially the use of standard libraries , with issues of profiling, debugging and optimizing code for CUDA and the application of technologies and development programs Thrust and OpenACC will be presented to all listeners. The methods of managing multiple GPU within a single workstation in a system with distributed memory will be also discussed during the track.