USING CUDA TECHNOLOGY COMPANY NVIDIA FOR CAD NEURAL NETWORK

Denis Kalitin, the Candidate of Technical Science, Associate Professor, MSMU

Abstract
The development of computer technology throughout the world are in various ways. The current reality requires an increase in computing power, but often, increasing productivity by an order or more, require significant financial investment. This article describes a new parallelization of data processing, which can give a significant gain in performance, but it does not require huge financial investments. CAD neural networks, in essence, an object with a high proportion of tasks, which are based on parallel computing. It is obvious that, using the technologies can be several orders of magnitude speed up a CAD system.

KEYWORDS: neural networks, CAD, parallel processing, programming, CUDA, GPGPU.

Download article: USING CUDA TECHNOLOGY COMPANY NVIDIA FOR CAD NEURAL NETWORK