# On Scaling Up 3D Gaussian Splatting Training

arXiv: https://arxiv.org/abs/2406.18533 · HTML: https://mlsanity.com/paper/2406.18533/

## Key facts

- Cited by 49 papers in the map (data through 14 Oct 2026).
- Builds on 26 of its 37 references that are in the map.
- First posted on arXiv 26 Jun 2024.

## Builds on (most cited first, 26 of 26)

- [3D Gaussian Splatting for Real-Time Radiance Field Rendering](https://mlsanity.com/paper/2308.04079/) (2023)
- [Mip-NeRF 360: Unbounded Anti-Aliased Neural Radiance Fields](https://mlsanity.com/paper/2111.12077/) (2021)
- [Megatron-LM: Training Multi-Billion Parameter Language Models Using Model Parallelism](https://mlsanity.com/paper/1909.08053/) (2019)
- [Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour](https://mlsanity.com/paper/1706.02677/) (2017)
- [On Large-Batch Training for Deep Learning: Generalization Gap and Sharp Minima](https://mlsanity.com/paper/1609.04836/) (2016)
- [ZeRO: Memory Optimizations Toward Training Trillion Parameter Models](https://mlsanity.com/paper/1910.02054/) (2019)
- [Scaffold-GS: Structured 3D Gaussians for View-Adaptive Rendering](https://mlsanity.com/paper/2312.00109/) (2023)
- [Large Batch Optimization for Deep Learning: Training BERT in 76 minutes](https://mlsanity.com/paper/1904.00962/) (2019)
- [Efficient Large-Scale Language Model Training on GPU Clusters Using Megatron-LM](https://mlsanity.com/paper/2104.04473/) (2021)
- [Mega-NeRF: Scalable Construction of Large-Scale NeRFs for Virtual Fly-Throughs](https://mlsanity.com/paper/2112.10703/) (2021)
- [Octree-GS: Towards Consistent Real-time Rendering with LOD-Structured 3D Gaussians](https://mlsanity.com/paper/2403.17898/) (2024)
- [BungeeNeRF: Progressive Neural Radiance Field for Extreme Multi-scale Scene Rendering](https://mlsanity.com/paper/2112.05504/) (2021)
- [A Hierarchical 3D Gaussian Representation for Real-Time Rendering of Very Large Datasets](https://mlsanity.com/paper/2406.12080/) (2024)
- [VastGaussian: Vast 3D Gaussians for Large Scene Reconstruction](https://mlsanity.com/paper/2402.17427/) (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)
- [GSPMD: General and Scalable Parallelization for ML Computation Graphs](https://mlsanity.com/paper/2105.04663/) (2021)
- [On the SDEs and Scaling Rules for Adaptive Gradient Algorithms](https://mlsanity.com/paper/2205.10287/) (2022)
- [On the Validity of Modeling SGD with Stochastic Differential Equations (SDEs)](https://mlsanity.com/paper/2102.12470/) (2021)
- [Pollux: Co-adaptive Cluster Scheduling for Goodput-Optimized Deep Learning](https://mlsanity.com/paper/2008.12260/) (2020)
- [DOGS: Distributed-Oriented Gaussian Splatting for Large-Scale 3D Reconstruction Via Gaussian Consensus](https://mlsanity.com/paper/2405.13943/) (2024)
- [Learning Rates as a Function of Batch Size: A Random Matrix Theory Approach to Neural Network Training](https://mlsanity.com/paper/2006.09092/) (2020)
- [Supporting Very Large Models using Automatic Dataflow Graph Partitioning](https://mlsanity.com/paper/1807.08887/) (2018)
- [How to Scale Your EMA](https://mlsanity.com/paper/2307.13813/) (2023)
- [Surge Phenomenon in Optimal Learning Rate and Batch Size Scaling](https://mlsanity.com/paper/2405.14578/) (2024)
- [RetinaGS: Scalable Training for Dense Scene Rendering with Billion-Scale 3D Gaussians](https://mlsanity.com/paper/2406.11836/) (2024)

## Built on by (most cited first, 30 of 49)

- [FlashGS: Efficient 3D Gaussian Splatting for Large-scale and High-resolution Rendering](https://mlsanity.com/paper/2408.07967/) (2024)
- [CityGaussianV2: Efficient and Geometrically Accurate Reconstruction for Large-Scale Scenes](https://mlsanity.com/paper/2411.00771/) (2024)
- [3DGS-LM: Faster Gaussian-Splatting Optimization with Levenberg-Marquardt](https://mlsanity.com/paper/2409.12892/) (2024)
- [Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenes](https://mlsanity.com/paper/2412.01745/) (2024)
- [CityGS-X: A Scalable Architecture for Efficient and Geometrically Accurate Large-Scale Scene Reconstruction](https://mlsanity.com/paper/2503.23044/) (2025)
- [MVGS: Multi-view Regulated Gaussian Splatting for Novel View Synthesis](https://mlsanity.com/paper/2410.02103/) (2024)
- [FlexGaussian: Flexible and Cost-Effective Training-Free Compression for 3D Gaussian Splatting](https://mlsanity.com/paper/2507.06671/) (2025)
- [OccluGaussian: Occlusion-Aware Gaussian Splatting for Large Scene Reconstruction and Rendering](https://mlsanity.com/paper/2503.16177/) (2025)
- [3DGEER: 3D Gaussian Rendering Made Exact and Efficient for Generic Cameras](https://mlsanity.com/paper/2505.24053/) (2025)
- [Robust and Efficient 3D Gaussian Splatting for Urban Scene Reconstruction](https://mlsanity.com/paper/2507.23006/) (2025)
- [Unifying Color and Lightness Correction with View-Adaptive Curve Adjustment for Robust 3D Novel View Synthesis](https://mlsanity.com/paper/2602.18322/) (2026)
- [Anti-Aliased 2D Gaussian Splatting](https://mlsanity.com/paper/2506.11252/) (2025)
- [CityGo: Lightweight Urban Modeling and Rendering with Proxy Buildings and Residual Gaussians](https://mlsanity.com/paper/2505.21041/) (2025)
- [Proxy-GS: Unified Occlusion Priors for Training and Inference in Structured 3D Gaussian Splatting](https://mlsanity.com/paper/2509.24421/) (2025)
- [From Orbit to Ground: Generative City Photogrammetry from Extreme Off-Nadir Satellite Images](https://mlsanity.com/paper/2512.07527/) (2025)
- [Optimized 3D Gaussian Splatting using Coarse-to-Fine Image Frequency Modulation](https://mlsanity.com/paper/2503.14475/) (2025)
- [SLAG: Scalable Language-Augmented Gaussian Splatting](https://mlsanity.com/paper/2505.08124/) (2025)
- [Differentiable Light Transport with Gaussian Surfels via Adapted Radiosity for Efficient Relighting and Geometry Reconstruction](https://mlsanity.com/paper/2509.18497/) (2025)
- [Industrial-Grade Sensor Simulation via Gaussian Splatting: A Modular Framework for Scalable Editing and Full-Stack Validation](https://mlsanity.com/paper/2503.11731/) (2025)
- [M3: 3D-Spatial MultiModal Memory](https://mlsanity.com/paper/2503.16413/) (2025)
- [TextureSplat: Per-Primitive Texture Mapping for Reflective Gaussian Splatting](https://mlsanity.com/paper/2506.13348/) (2025)
- [CLM: Removing the GPU Memory Barrier for 3D Gaussian Splatting](https://mlsanity.com/paper/2511.04951/) (2025)
- [Time-resolved dynamic CBCT reconstruction using prior-model-free spatiotemporal Gaussian representation (PMF-STGR)](https://mlsanity.com/paper/2503.22139/) (2025)
- [A LoD of Gaussians: Unified Training and Rendering for Ultra-Large Scale Reconstruction with External Memory](https://mlsanity.com/paper/2507.01110/) (2025)
- [LoBE-GS: Load-Balanced and Efficient 3D Gaussian Splatting for Large-Scale Scene Reconstruction](https://mlsanity.com/paper/2510.01767/) (2025)
- [BalanceGS: Algorithm-System Co-design for Efficient 3D Gaussian Splatting Training on GPU](https://mlsanity.com/paper/2510.14564/) (2025)
- [MetroGS: Efficient and Stable Reconstruction of Geometrically Accurate High-Fidelity Large-Scale Scenes](https://mlsanity.com/paper/2511.19172/) (2025)
- [Towards Next-Generation SLAM: A Survey on 3DGS-SLAM Focusing on Performance, Robustness, and Future Directions](https://mlsanity.com/paper/2602.04251/) (2026)
- [BlitzGS: City-Scale Gaussian Splatting at Lightning Speed](https://mlsanity.com/paper/2605.13794/) (2026)
- [TideGS: Scalable Training of Over One Billion 3D Gaussian Splatting Primitives via Out-of-Core Optimization](https://mlsanity.com/paper/2605.20150/) (2026)

## Abstract (arXiv)

3D Gaussian Splatting (3DGS) is increasingly popular for 3D reconstruction due to its superior visual quality and rendering speed. However, 3DGS training currently occurs on a single GPU, limiting its ability to handle high-resolution and large-scale 3D reconstruction tasks due to memory constraints. We introduce Grendel, a distributed system designed to partition 3DGS parameters and parallelize computation across multiple GPUs. As each Gaussian affects a small, dynamic subset of rendered pixels, Grendel employs sparse all-to-all communication to transfer the necessary Gaussians to pixel partitions and performs dynamic load balancing. Unlike existing 3DGS systems that train using one camera view image at a time, Grendel supports batched training with multiple views. We explore various optimization hyperparameter scaling strategies and find that a simple sqrt(batch size) scaling rule is highly effective. Evaluations using large-scale, high-resolution scenes show that Grendel enhances rendering quality by scaling up 3DGS parameters across multiple GPUs. On the Rubble dataset, we achieve a test PSNR of 27.28 by distributing 40.4 million Gaussians across 16 GPUs, compared to a PSNR of 26.28 using 11.2 million Gaussians on a single GPU. Grendel is an open-source project available at: https://github.com/nyu-systems/Grendel-GS