Course (certified) on Tech Diplomacy

The course presents the possibility of general purpose use of the computational power of graphics boards thanks to a generalized model of their GPUs. The hardware architecture of graphical processors is presented together with the general purpose OpenCL software development environment. Algorithms suitable to massively parallel implementation are presented using practical examples. Topics studied in deatils include: operations on big amount of data, parallel primitives in the OpenCL environment, solution of a set of linear equations, physical simulation on GPU, hash based parallel algorithms, Monte Carlo methods in GPU, optimization issues of GPGPU algorithms, effective cooperation with graphical APIs, special questions of multi GPU and distributed systems.