Please use this identifier to cite or link to this item: http://hdl.handle.net/11455/72272
標題: 電腦動畫系統在廣域分散式計算環境之設計與實現
作者: 王宗銘 
林永森
關鍵字: 電腦動畫;分散式計算;光線追蹤法;電腦圖學
出版社: 國立中興大學工學院;Airiti Press Inc.
Project: 興大工程學報, Volume 10, Issue 1, Page(s) 1-14.
摘要: 
本文敘述設計與實現廣域分散式電腦動畫系統-ANA。系統整合了電腦動畫的模型輸入、成圖計算與動畫輸出三大步驟。在模型輸入方面,系統支援個人電腦上普遍被使用的動畫軟體3D Studio來建構三度空間模型,並經轉檔後作成圖計算。在成圓計算方面,系統採用光線追蹤成像語法,並結合廣域網路上的工作站群計算資源共同來完成。系統發展適切的工作分配方法,以達成計算資源間的負載平衡。此外,我們也提出有效的方法,使系統具備容錯能力。在動畫輸出方面,為了減少網路傳輸負載,大量成圖計算後之圖形檔案資料在當地區域網路內,依MPEG-1檔案格式先行壓縮後,再傳回主程式控制端。主程式控制端經解壓縮還原後,再以壓縮方式,並配合後製作方式作為最後的動畫輸出。實驗結果顯示:ANA系統成圖計算時間已大幅縮減:原本在單台SPARC-5機器上需要至少26小時14分才能完成的成圓計算工作量,系統在同時使用32台異質網路工作站連結之計算資源後,只需要約1小時46分便可完成,其時間節省之獲益達92.7%。此外,系統具備高度之容錯能力;在17台計算資源全數異常時,仍能正常運作。本文結論為:利用廣域網路的分散式計算環境,使ANA系統之工作效率獲得顯著的提昇;由於系統的容錯能力,提昇分散式電腦動畫系統在廣域分散式計算環境之穩定性。

Computer animation is considered as a new medium that has been applied widely to many fields including advertisement, entertainment, art, education, simulation, etc. This paper investigates design and implementation of distributed computer animation system on computing environments connected by a wide-area network. A computer animation system, ANA, has been developed which supports the model conversion, distributed rendering and animation output. 3D models constructed at PC-based 3D Studio computer animation software are converted as the input model. The time-consuming rendering is proceeded using ray tracing algorithm on heterogeneous networked workstations. A task distribution strategy is proposed to balance the loads among computing resources. Furthermore, a successful fault tolerance method has been presented to improve the system stability. To reduce the load of network transmission, computed images are encoded into MPEG-1 files at the local control node before they are transmitted back to the console for further processing. Experimental results demonstrate that when connecting 32 heterogeneous networked workstations, the rendering time has been significantly reduced to around 1 hour and 46 minutes, in compared to nearly 26 hours 14 minutes that are required when using a single SPARC-5 resource. The timing improvement has achieved to 92.7%. The system illustrates high degree of stabilization. It can be executed properly even if 16 computing resources are being aberrant. The concluding remarks are that the ANA system exhibits a potential to significantly improve the computing efficiency, while the fault tolerance method we proposed has greatly maintained the system stability.
URI: http://hdl.handle.net/11455/72272
ISSN: 1017-4397
Appears in Collections:第10卷 第1期
工學院

Show full item record
 
TAIR Related Article

Google ScholarTM

Check


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.