# TideGS: Scalable Training of Over One Billion 3D Gaussian Splatting Primitives via Out-of-Core Optimization

arXiv: https://arxiv.org/abs/2605.20150 · HTML: https://mlsanity.com/paper/2605.20150/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 32 of its 40 references that are in the map.
- First posted on arXiv 19 May 2026.

## Builds on (most cited first, 30 of 32)

- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Instant Neural Graphics Primitives with a Multiresolution Hash Encoding](https://mlsanity.com/paper/2201.05989/) (2022)
- [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021)
- [Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering](https://mlsanity.com/paper/2312.00109/) (2023)
- [ZeRO-Offload: Democratizing Billion-Scale Model Training](https://mlsanity.com/paper/2101.06840/) (2021)
- [VastGaussian: Vast 3D Gaussians for Large Scene Reconstruction](https://mlsanity.com/paper/2402.17427/) (2024)
- [Taming 3DGS: High-Quality Radiance Fields with Limited Resources](https://mlsanity.com/paper/2406.15643/) (2024)
- [CityGaussian: Real-time High-quality Large-Scale Scene Rendering with Gaussians](https://mlsanity.com/paper/2404.01133/) (2024)
- [MatrixCity: A Large-scale City Dataset for City-scale Neural Rendering and Beyond](https://mlsanity.com/paper/2309.16553/) (2023)
- [Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives](https://mlsanity.com/paper/2412.00578/) (2024)
- [FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering](https://mlsanity.com/paper/2408.07967/) (2024)
- [DOGS: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus](https://mlsanity.com/paper/2405.13943/) (2024)
- [CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes](https://mlsanity.com/paper/2411.00771/) (2024)
- [On Scaling Up 3D Gaussian Splatting Training](https://mlsanity.com/paper/2406.18533/) (2024)
- [3DGS-LM: Faster Gaussian-Splatting Optimization with Levenberg-Marquardt](https://mlsanity.com/paper/2409.12892/) (2024)
- [FastGS: Training 3D Gaussian Splatting in 100 Seconds](https://mlsanity.com/paper/2511.04283/) (2025)
- [RecSSD: Near Data Processing for Solid State Drive Based Recommendation Inference](https://mlsanity.com/paper/2102.00075/) (2021)
- [CityGS-X: A Scalable Architecture for Efficient and Geometrically Accurate Large-Scale Scene Reconstruction](https://mlsanity.com/paper/2503.23044/) (2025)
- [Rip-NeRF: Anti-aliasing Radiance Fields with Ripmap-Encoded Platonic Solids](https://mlsanity.com/paper/2405.02386/) (2024)
- [Drone-assisted Road Gaussian Splatting with Cross-view Uncertainty](https://mlsanity.com/paper/2408.15242/) (2024)
- [3DGS2: Near Second-order Converging 3D Gaussian Splatting](https://mlsanity.com/paper/2501.13975/) (2025)
- [Characterizing & Finding Good Data Orderings for Fast Convergence of Sequential Gradient Methods](https://mlsanity.com/paper/2202.01838/) (2022)
- [RetinaGS: Scalable Training for Dense Scene Rendering with Billion-Scale 3D Gaussians](https://mlsanity.com/paper/2406.11836/) (2024)
- [SlimmeRF: Slimmable Radiance Fields](https://mlsanity.com/paper/2312.10034/) (2023)
- [Unifying Appearance Codes and Bilateral Grids for Driving Scene Gaussian Splatting](https://mlsanity.com/paper/2506.05280/) (2025)
- [Balanced 3DGS: Gaussian-wise Parallelism Rendering with Fine-Grained Tiling](https://mlsanity.com/paper/2412.17378/) (2024)
- [OccluGaussian: Occlusion-Aware Gaussian Splatting for Large Scene Reconstruction and Rendering](https://mlsanity.com/paper/2503.16177/) (2025)
- [GS-Cache: A GS-Cache Inference Framework for Large-scale Gaussian Splatting Models](https://mlsanity.com/paper/2502.14938/) (2025)
- [GS-Scale: Unlocking Large-Scale 3D Gaussian Splatting Training via Host Offloading](https://mlsanity.com/paper/2509.15645/) (2025)
- [ContraGS: Codebook-Condensed and Trainable Gaussian Splatting for Fast, Memory-Efficient Reconstruction](https://mlsanity.com/paper/2509.03775/) (2025)

## Built on by (most cited first, 2 of 2)

- [One Video, One World: Turning Monocular Video into Physical 4D Scenes](https://mlsanity.com/paper/2606.31388/) (2026)
- [From Splats to Silicon: Rethinking Computational Efficiency of 3DGS](https://mlsanity.com/paper/2609.06157/) (2026)

## Abstract (arXiv)

Training 3D Gaussian Splatting (3DGS) at billion-primitive scale is fundamentally memory-bound: each Gaussian primitive carries a large attribute vector, and the aggregate parameter table quickly exceeds GPU capacity, limiting prior systems to tens of millions of Gaussians on commodity single-GPU hardware. We observe that 3DGS training is inherently sparse and trajectory-conditioned: each iteration activates only the Gaussians visible from the current camera batch, so GPU memory can serve as a working-set cache rather than a persistent parameter store. Building on this insight, we introduce TideGS, an out-of-core training framework that manages parameters across an SSD-CPU-GPU hierarchy via three synergistic techniques: block-virtualized geometry for SSD-aligned spatial locality, a hierarchical asynchronous pipeline to overlap I/O with computation, and trajectory-adaptive differential streaming that transfers only incremental working-set deltas between iterations. Experiments show that TideGS enables training with over one billion Gaussians on a single 24 GB GPU while achieving the best reconstruction quality among evaluated single-GPU baselines on large-scale scenes, scaling beyond prior out-of-core baselines (e.g., approximately 100M Gaussians) and standard in-memory training (e.g., approximately 11M Gaussians).