The largest in Ukraine supercomputer complex SKIT with total peak performance of reaching 43 tril lion operations per second operates at the Glushkov Institute of Cybernetics of NAS of Ukraine. The supercomputer effectively solves the tasks of the widest possible spectrum, from the study of the Earth's interior to linguistic studios, from genetics to modeling the evolution of black holes. The complex has found one of its applications in solving photogrammetric problems within the framework of R&D project of the NAS of Ukraine in partnership with Pixelated Realities NGO that deals with the promotion of digital methods for restoration, reconstruction, and research of objects of cultural heritage and with planning and revitalization of urban space. The purpose of the project is to create a specialized hardware and software complex based on the SKIT supercomputer to solve the resource-intensive photogrammetric problem. The software and hardware complex is based on original architecture of the SKIT supercomputer and photogrammetric method of building 3D scenes, which is reproducing the 3D objects by photographs. The task of photogrammetry involves the processing of several ten thousand photographs to create 3D scanned images of monuments, historic buildings, archaeological excavations, and urban objects using Reality Capture software. The project innovative result is the creation of the first in Ukraine specialized software and hardware system that uses high-performance computations to solve the problems of 3D reconstruction by photographs. Also, the Institute of Cybernetics, together with its partners, has developed Cloud Studio specialized hardware and software system. Conclusions: the creation of specialized supercomputer segment and examples of its use have demonstrated the effectiveness of such solutions in 3D modeling and animation, as well as the presence of a significant interest and a large market of supercomputer technologies. The Cloud Studio continues a successful experience of creating specialized computing systems designed to solve important applied R&D problems.

• Головинський Андрій Леонідович (технічні науки)
• Маленко Андрій Леонідович (фізико-математичні науки)

  Якщо, ви не знайшли інформацію про автора(ів) публікації, маєте бажання виправити або відобразити більш докладну інформацію про науковців України запрошуємо заповнити "Анкету науковця"