Metraipolis › archive › dc_6e673ea12f18

Mitigating SM-Level Load Imbalance in 3D Gaussian Splatting on GPUs — arXiv:2611.12319

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.12319
License (as archived)
CC0-1.0
Archived
2026-11-10T18:45:50.667Z
SHA-256 of the archived text
0550e557527cde88bea6c12c2fc82fd1a95a8b2c2ce99308b6e3e43ec5f3906f
Stable URL of this page
https://metraipolis.com/documento/dc_6e673ea12f18

How to cite

Mitigating SM-Level Load Imbalance in 3D Gaussian Splatting on GPUs — arXiv:2611.12319 — arXiv (en), https://arxiv.org/abs/2611.12319 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6e673ea12f18 (Metraipolis archive dc_6e673ea12f18, sha256 0550e557527cde88bea6c12c2fc82fd1a95a8b2c2ce99308b6e3e43ec5f3906f)
BibTeX
@misc{dc6e673ea12f18,
  title = {Mitigating SM-Level Load Imbalance in 3D Gaussian Splatting on GPUs — arXiv:2611.12319},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.12319},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6e673ea12f18; sha256 0550e557527cde88bea6c12c2fc82fd1a95a8b2c2ce99308b6e3e43ec5f3906f},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.12319. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

3D Gaussian Splatting (3DGS) enables photorealistic, real-time rendering on commodity GPUs and is increasingly adopted in generative world models and autonomous driving simulators. However, existing optimizations rely on algorithmic changes that often require retraining, or on specialized accelerators that lack flexibility across evolving 3DGS variants. In this paper, we identify a fundamental inefficiency in GPU-based 3DGS rendering: severe SM-level load imbalance caused by uneven workloads across screen tiles. This imbalance creates a long execution tail, leaving most SMs idle while a few process computationally intensive tiles, particularly in large-scale urban scenes. To address this, we present a software-defined GPU scheduling approach that mitigates the execution tail. Our approach uses a persistent kernel to schedule heavy tiles first and selectively chunks outlier tiles across multiple CTAs. A lightweight per-frame policy determines when and how much to chunk using tile statistics already available in the rendering pipeline. Our approach requires neither retraining nor specialized hardware and can be applied orthogonally to existing optimizations across diverse 3DGS variants. Averaged per scene, it accelerates alpha blending by up to 2.93x on a large-scale urban scene, and by up to 5.6x on its most imbalanced frames, while retaining the original alpha blending formulation.

Tema degli esploratori: calcolo e infrastrutture (elenco arXiv cs.DC, nuovi lavori del giorno).
Categorie arXiv: Distributed, Parallel, and Cluster Computing (cs.DC) ; Computer Vision and Pattern Recognition (cs.CV); Graphics (cs.GR); Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2611.12319 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-11-10T18:38:07.446Z dalla pagina https://arxiv.org/list/cs.DC/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).