3D Gaussian splatting (3DGS) has drawn significant attention in the architectural community recently. However, current architectural designs often overlook the 3DGS scalability, making them fragile for extremely large-scale 3DGS. Meanwhile, the VR bandwidth requirement makes it impossible to deliver high-fidelity and smooth VR content from the cloud. We present Nebula, a coherent acceleration framework for large-scale 3DGS collaborative rendering. Instead of streaming videos, Nebula streams intermediate results after the LoD search, reducing 1925% data communication between the cloud and the client. To further enhance the motion-to-photon experience, we introduce a temporal-aware LoD search in the cloud that tames the irregular memory access and reduces redundant data access by exploiting temporal coherence across frames. On the client side, we propose a novel stereo rasterization that enables two eyes to share most computations during the stereo rendering with bit-accurate quality. With minimal hardware augmentations, Nebula achieves 2.7$\times$ motion-to-photon speedup and reduces 1925% bandwidth over lossy video streaming.

翻译：三维高斯溅射（3DGS）近期在计算机图形学界引起了广泛关注。然而，当前系统设计往往忽视3DGS的可扩展性，导致其难以支撑超大规模3DGS场景。同时，虚拟现实（VR）的高带宽需求使得从云端传输高保真流畅VR内容面临巨大挑战。本文提出星云（Nebula）——一个面向大规模3DGS协同渲染的连贯加速框架。该框架通过传输细节层次（LoD）搜索后的中间结果替代传统视频流，成功将云端与客户端间的数据通信量降低1925%。为进⼀步优化运动到光子延迟体验，我们在云端设计了时序感知的LoD搜索机制，通过利用帧间时序相关性来规整不规则内存访问并减少冗余数据读取。在客户端侧，我们提出新型立体光栅化算法，使双眼在立体渲染过程中共享绝大部分计算，同时保持比特级精度质量。通过最小化的硬件增强，星云系统在保持无损画质的前提下，较有损视频流方案实现了2.7倍的运动到光子加速，并降低1925%带宽消耗。