METHODS OF OPTIMIZATION 3D GRAPHICS RENDERING ON TILE-BASED GRAPHICSPROCESSORS BY THE EXAMPLE OF A VISUALIZATION MODELING N-BODY GRAVITATIONAL INTERACTIONS
Keywords:
tile-based processors, TBDR architecture, optimization, graphics engine, state update tree, geometry merging, geometry instancing, mesh difference, N-body simulation.Abstract
Visualization of physical phenomena such as gravitational interactions of celestial bodies requires rendering a large number of identical 2D and 3D objects. Unfortunately, tile-based deferred architectures of graphical accelerators, popular on the mobile market, are not suitable to process a large number of rendering requests. On such devices, it is necessary to use various optimization techniques to minimize the number of draw calls. In this work, we present implementation of a graphical engine that maintains a tree of state changes for every scene object to allow resubmission of only necessary sets of data to perform rendering. Furthermore, our system minimizes the number of objects to transfer by merging meshes and performing geometry instancing. The system also offloads the data bus by computing and sending only the differences between geometry meshes. The speedup relative to an unoptimized version is demonstrated by measuring the speed of systems to visualize modeling results of gravitational interactions between a large number of celestial bodies.
References
Верзунов С.Н. Разработка программных средств для вейвлет-анализа одномерных временных рядов // Проблемы автоматики и управления. – 2014. – № 2(27). –
С. 62–71.
Shebanow, M. An evolution of mobile graphics. Презентация конференции High Per-formance Graphics. 2013.
Imagination Tech Limited. PowerVR Hardware Architecture Overview for Developers // PowerVR SDK Manuals. 2017.
Navik, Ankit P., и другие. Microbenchmark Based Performance Evaluation of GPU Rendering. // Emerging Research in Computing, Information, Communication and Ap-plications. Springer, New Delhi, 2015.
McCaffrey, Jon. Exploring Mobile vs. Desktop OpenGL Performance. // OpenGL In-sights. 2012.
Arnau, Jose-Maria, Joan-Manuel Parcerisa, и Polychronis Xekalakis. Parallel frame rendering: Trading responsiveness for energy on a mobile gpu. // Proceedings of the 22nd international conference on Parallel architectures and compilation techniques. IEEE Press. 2013.
Ma, Xiaohan, и другие. Characterizing the performance and power consumption of 3D mobile games. // Computer 46.4. 2013. (С. 76–82).
http://iait.kg/wp-content/uploads/2018/04/Ver_Toks.png (дата обращения: 2. 04.2018).
Downloads
Published
Issue
Section
Categories
License
Copyright (c) 2022 Сергей Николаевич
This work is licensed under a Creative Commons Attribution 4.0 International License.