CUDA

This pa­per addres­ses the prob­lem of iden­tif­ying rhe­olo­gi­cal pa­ra­me­ters of wo­od using ar­ti­fi­ci­al neu­ral net­works with pa­ral­lel le­ar­ning al­go­rithm using Python prog­ram­ming lan­gua­ge, Cha­iner fra­me­work and CU­DA techno­logy. An in­tel­li­gent system for iden­ti­fi­ca­ti­on of rhe­olo­gi­cal pa­ra­me­ters of wo­od has be­en de­ve­lo­ped. The system cre­ated con­ta­ins the most user-fri­endly in­ter­fa­ce, all the ne­ces­sary set of to­ols for au­to­ma­ti­on of the pro­cess of vis­ua­li­za­ti­on and analysis of da­ta.

This article provides an overview of implementations of algorithms FFT (Fast Fourier transform) in software libraries, which are widely used in software development. Analyzed the most widely used software libraries (FFTW, Intel IPP, CUDA CuFFT), which implemented FFT, including selected such that would make the most efficient use of computing resources of a modern computer are analyzed.

This article deals with the principles of organization of hybrid architecture of cluster computing based on the use of computing opportunities of the CPU and GPU with support for GPGPU. The advantages and disadvantages of this method of calculation have been analyzed, and the feasibility of realization of parallel programs based on the use of a hybrid architecture of cluster computing has been demonstrated. A software framework that can be used to develop the applications has been implemented on the basis of the proposed programming model.

Здійснено короткий огляд технологій виконання обчислень на графічному процесорі (GPGPU) та особливостей написання програм при їх використанні. Перевірено ефективність способу виконання поданих потоковим графом алгоритмів на прикладі алгоритму швидкого перетворення Фур’є на графічному процесорі (GPU) з підтримкою технології GPGPU. Показано перспективу цього способу як для виконання на GPU, так і на CPU.

Analysis of characteristics of software implementation of parallelization of process of constructing of discrete dynamical models was conducted in this paper. SIMD-architecture was used for the task of parallelization. Technology CUDA and GPU NVIDIA was used for this software implementation.