本論文提出一個平行彩繪 (parallel rendering) 系統的架構，並利用此系統來完成多重解析度模型的彩繪工作，尤指資料量龐大或複雜度高的精細模型。該平行彩繪系統乃是採用分散式系統的網路架構，並在叢集電腦 (PC cluster) 上實作而成。 在該平行彩繪系統中，各電腦之間以100Mbps的乙太網路相連結，同時利用我們發展的通訊協定，透過Winsock來達到資訊的交換，以提供處理大量資料時所需的運算能力與記憶體空間。 我們的平行彩繪系統所採用的影像平行演算法是sort-last演算法，該演算法可以提供較佳的擴充性，以利後續的研究發展。同時我們也採用OpenGL來完成實際彩繪的工作。 我們的平行彩繪系統支援離散性多重解析度 (discrete multiresolution) 的模型彩繪，以更進一步提昇彩繪的效率。 A parallel rendering system is presented to render multiresolution models on a PC cluster. The system provides great memory capacity and calculation power for real-time visualization on large-amounted data such as terrain models. The rendering system is implemented on a distributed architecture; all PCs in the cluster is connected by a 100Mbps Ethernet network. The information are exchanged using the protocol that we developed through Winsock to achieve the cooperation. The parallel rendering algorithm is adopted from the sort-last strategy due to its high scalability and the rendering algorithm is implemented by OpenGL library. The parallel rendering system supports the parallel rendering for discrete multiresolution models to improve the rendering performance.
