Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Zhao, Hexu, Weng, Haoyang, Lu, Daohan, Li, Ang, Li, Jinyang, Panda, Aurojit, Xie, Saining
Format: Preprint
Veröffentlicht: 2024
Schlagworte:
Online-Zugang:https://arxiv.org/abs/2406.18533
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866909890799206400
author Zhao, Hexu
Weng, Haoyang
Lu, Daohan
Li, Ang
Li, Jinyang
Panda, Aurojit
Xie, Saining
author_facet Zhao, Hexu
Weng, Haoyang
Lu, Daohan
Li, Ang
Li, Jinyang
Panda, Aurojit
Xie, Saining
contents 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
format Preprint
id arxiv_https___arxiv_org_abs_2406_18533
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle On Scaling Up 3D Gaussian Splatting Training
Zhao, Hexu
Weng, Haoyang
Lu, Daohan
Li, Ang
Li, Jinyang
Panda, Aurojit
Xie, Saining
Computer Vision and Pattern Recognition
I.4.5
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
title On Scaling Up 3D Gaussian Splatting Training
topic Computer Vision and Pattern Recognition
I.4.5
url https://arxiv.org/abs/2406.18533