在线片元剔除的多片元效果高效绘制方法

系统仿真学报 ›› 2016, Vol. 28 ›› Issue (10): 2407-2414.

在线片元剔除的多片元效果高效绘制方法

周果^1,2,3, 朱登明^1,2, 王兆其^1,2, 魏毅^1,2

1.中国科学院计算技术研究所,北京 100190;
2.移动计算与新型终端北京市重点实验室,北京 100190;
3.中国科学院大学,北京 100049

收稿日期:2016-05-09 修回日期:2016-07-22 出版日期:2016-10-08 发布日期:2020-08-13
作者简介:周果(1987-),男,陕西,博士生,研究方向为计算机图形学;朱登明(1973-),男,安徽,博士,副研究员,研究方向为自然现象模拟和可视化。
基金资助:
国家自然科学基金(61173067,61379085,61532002),国家863计划项目(2015AA016401)

Online Fragment Culling for Efficient Rendering of Multi-Fragment Effects

Zhou Guo^1,2,3, Zhu Dengming^1,2, Wang Zhaoqi^1,2, Wei Yi^1,2

1. Institute of Computing Technology, Chinese Academy of Sciences, Beijing 100190, China;
2. Beijing Key Laboratory of Mobile Computing and Pervasive Device, Beijing 100190;
3. University of Chinese Academy of Sciences, Beijing 100049, China

Received:2016-05-09 Revised:2016-07-22 Online:2016-10-08 Published:2020-08-13

摘要/Abstract

摘要： 多片元效果有实时透明等重要应用,它按深度顺序处理每个像素的所有片元。已有的k-buffer算法在单遍绘制场景时,剥取并缓存距离视点最近的多个片元,这需要避免缓存更新时的读写冲突,降低了在复杂场景上的计算效率。对此提出利用帧间的时空相关性,允许单遍绘制时在线地剔除缓存外片元的方法。描述建立半堆结构的片元缓存的过程,分析最坏情况下的时间复杂度,在绘制每帧时反向重投影构造剔除界面。由于减少了缓存更新操作且不需要预处理,复杂场景特别在被漫游或发生变化时的绘制效率被显著提高。

关键词: 深度剥离, 堆排序, 片元临界区, 反向重投影

Abstract: Multi-fragment effects such as real-time transparency are rendered by visiting the per-pixel fragments in depth order. The k-buffer algorithm rendered the scene in a single pass to obtain several layers of fragments. The read-modify-write hazard while updating the buffer should be avoided thus the performance was likely to be hit hard. The temporal and spatial coherence among frames were exploited to cull the input fragments in an online manner. A fragment cache based on a semi-heap was proposed. The worst case scenario of the time complexity was analyzed. The reverse reprojection was employed to construct the culling interface on the fly while rendering a frame. Since the buffer updating operations are effectively reduced and no preprocessing is required, the complexity scene during walkthrough or changing can be rendered efficiently.

Key words: depth peeling, heap sort, fragment interlock, reverse reprojection

中图分类号:

TP391.9

周果, 朱登明, 王兆其, 魏毅. 在线片元剔除的多片元效果高效绘制方法[J]. 系统仿真学报, 2016, 28(10): 2407-2414.

Zhou Guo, Zhu Dengming, Wang Zhaoqi, Wei Yi. Online Fragment Culling for Efficient Rendering of Multi-Fragment Effects[J]. Journal of System Simulation, 2016, 28(10): 2407-2414.

参考文献 19

[1]	Bavoil L, Callahan S P, Lefohn A, et al.Multi-fragment effects on the GPU using the k-buffer [C]// Proceedings of the Symposium on Interactive 3D Graphics and Games. New York, USA: ACM Press, 2007: 97-104.
[2]	Everitt C.Interactive Order-Independent Transparency [R]. Santa Clara, USA: Nvidia Corporation, 2001.
[3]	Bavoil L, Myers K.Order independent transparency with dual depth peeling [R]. Santa Clara, USA: Nvidia Corporation, 2008.
[4]	Liu F, Huang M C, Liu X H, et al.Efficient depth peeling via bucket sort [C]// Proceedings of the Symposium on High Performance Graphics. Saarbrucken, Germany: ACM Press, 2009: 51-57.
[5]	刘芳, 黄梦成, 刘学慧, 等. 基于桶内动态融合的透明现象的高效绘制[J]. 计算机辅助设计与图形学学报, 2010, 22(3):382-387.
[6]	Yang J C, Hensley J, Grun H, et al.Real-time concurrent linked list construction on the GPU[J]. Computer Graphics Forum (S0167-7055), 2010, 29(4): 1297-1304.
[7]	Liu F, Huang M C, Liu X H, et al.FreePipe: a programmable parallel rendering architecture for efficient multi-fragment effects [C]// Proceedings of the Symposium on Interactive 3D Graphics Games. Washington D C, USA: ACM Press, 2010: 75-82.
[8]	黄梦成, 刘芳, 刘学慧, 吴恩华. 基于CUDA 渲染器的顺序独立透明现象的单遍高效绘制[J]. 软件学报, 2011, 22(8): 1927-1933.
[9]	Maule M, Comba J, Torchelsen R, et al.Hybrid transparency [C]// Proceedings of the Symposium on Interactive 3D Graphics Games. Orlando, USA: ACM Press, 2013: 103-118.
[10]	Wyman C.Exploring and Expanding the Continuum of OIT Algorithms[C]// Proceedings of the Symposium on High Performance Graphics. Dublin, Ireland: The Eurographics Association, 2016.
[11]	Vasilakis A A, Papaioannou G, Fudos I. An Efficient, Memory-Friendly and Dynamic k-buffer Framework[J]. IEEE Transaction on Visualization and Computer Graphics (S1077-2626), 2015, 21(6): 688-700.
[12]	Salvi M, Montgomery J, Lefohn A.Adaptive transparency [C]// Proceedings of the Symposium on High Performance Graphics. Vancouver, Canada: ACM Press, 2011: 119-126.
[13]	Salvi M, Vaidyanathan K.Multi-layer alpha blending [C]// Proceedings of the Symposium on Interactive 3D Graphics and Games. San Francisco, USA: ACM Press, 2014: 151-158.
[14]	谢国富, 王文成. 单遍数据读取的GPU上的多片元效果绘制[J]. 计算机学报, 2011, 34(3): 473-481.
[15]	Wang W, Xie G.Memory-efficient single-pass GPU rendering of multifragment effects[J]. IEEE Transaction on Visualization and Computer Graphics (S1077-2626), 2013, 19(8): 1307-1316.
[16]	Kumar R, Vassilvitskii S.Generalized distances between rankings [C]// Proceedings of the 19th international conference on World wide web. Raleigh, USA: ACM Press, 2010: 571-580.
[17]	Scherzer D, Yang L, Mattausch O, et al.Temporal Coherence Methods in Real-Time Rendering[J]. Computer Graphics Forum (S0167-7055), 2012, 31(8): 2378-2408.
[18]	Nehab D, Sander P, et al.Accelerating Real-Time Shading with Reverse Reprojection Caching[C]// Proceedings of the 22nd ACM symposium on Graphics hardware, Sarajevo, Bosnia and Herzegovina. USA: ACM, 2007: 25-35.
[19]	Yang L, Nehab D, Sander P, et al.Amortized Supersampling[J]. ACM Transactions on Graphics (S 0730-0301), 2009, 28(5): 238-250.