Applied Realization of Neural Network and Neurolike Parallel-Hierarchical System Based on GPGPU
Автор
Yaroviy, A. A.
Яровий, А. А.
Дата
2010Metadata
Показать полную информациюCollections
- Наукові роботи каф. КН [806]
Аннотации
The paper considers main aspects and features
of applied realization of neural network and neurolike parallelhierarchical systems based on GPGPU technologies. The
research performs the analysis of the architecture of modern
GPU and models of GPU parallel programming, as well as
considers ways of GPU programs optimization for parallel
computing in general-purpose parallel-hierarchical systems,
proposed as a networking model of neurolike scheme of
information processing. On the basis of research performed
and results obtained software modules are offered for
realization at CPU and GPU of large-scale neural and
neurolike parallel-hierarchical networks of different
topologies.
Пожалуйста, используйте этот идентификатор, чтобы цитировать или ссылаться на этот ресурс:
http://ir.lib.vntu.edu.ua//handle/123456789/18033