high performance computing on graphics processing units: hgpu.org

The paper deals with potentialities of various up-to-date software development tools for making use of graphic processor (GPU) parallel computing resources. Examples are given to illustrate the use of present-day software tools for the development of applications and realization of algorithms for scientific-technical calculations performed by GPGPU. The paper presents some classes of hard mathematical problems of scientific-technical calculations, for which the GPGPU can be efficiently used. is possible. To reduce the time of calculation program development with the use of GPGPU capabilities, various dedicated programming systems and problem-oriented subroutine libraries are recommended. Performance parameters when solving the problems with and without the use of GPGPU potentialities are compared.