CLM: Removing the GPU Memory Barrier for 3D Gaussian Splatting
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866912693181480960 |
|---|---|
| author | Zhao, Hexu Min, Xiwen Liu, Xiaoteng Gong, Moonjun Li, Yiming Li, Ang Xie, Saining Li, Jinyang Panda, Aurojit |
| author_facet | Zhao, Hexu Min, Xiwen Liu, Xiaoteng Gong, Moonjun Li, Yiming Li, Ang Xie, Saining Li, Jinyang Panda, Aurojit |
| contents | 3D Gaussian Splatting (3DGS) is an increasingly popular novel view synthesis approach due to its fast rendering time, and high-quality output. However, scaling 3DGS to large (or intricate) scenes is challenging due to its large memory requirement, which exceed most GPU's memory capacity. In this paper, we describe CLM, a system that allows 3DGS to render large scenes using a single consumer-grade GPU, e.g., RTX4090. It does so by offloading Gaussians to CPU memory, and loading them into GPU memory only when necessary. To reduce performance and communication overheads, CLM uses a novel offloading strategy that exploits observations about 3DGS's memory access pattern for pipelining, and thus overlap GPU-to-CPU communication, GPU computation and CPU computation. Furthermore, we also exploit observation about the access pattern to reduce communication volume. Our evaluation shows that the resulting implementation can render a large scene that requires 100 million Gaussians on a single RTX4090 and achieve state-of-the-art reconstruction quality. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2511_04951 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | CLM: Removing the GPU Memory Barrier for 3D Gaussian Splatting Zhao, Hexu Min, Xiwen Liu, Xiaoteng Gong, Moonjun Li, Yiming Li, Ang Xie, Saining Li, Jinyang Panda, Aurojit Computer Vision and Pattern Recognition D.4; I.3.2; I.3.7 3D Gaussian Splatting (3DGS) is an increasingly popular novel view synthesis approach due to its fast rendering time, and high-quality output. However, scaling 3DGS to large (or intricate) scenes is challenging due to its large memory requirement, which exceed most GPU's memory capacity. In this paper, we describe CLM, a system that allows 3DGS to render large scenes using a single consumer-grade GPU, e.g., RTX4090. It does so by offloading Gaussians to CPU memory, and loading them into GPU memory only when necessary. To reduce performance and communication overheads, CLM uses a novel offloading strategy that exploits observations about 3DGS's memory access pattern for pipelining, and thus overlap GPU-to-CPU communication, GPU computation and CPU computation. Furthermore, we also exploit observation about the access pattern to reduce communication volume. Our evaluation shows that the resulting implementation can render a large scene that requires 100 million Gaussians on a single RTX4090 and achieve state-of-the-art reconstruction quality. |
| title | CLM: Removing the GPU Memory Barrier for 3D Gaussian Splatting |
| topic | Computer Vision and Pattern Recognition D.4; I.3.2; I.3.7 |
| url | https://arxiv.org/abs/2511.04951 |