Potamoi: Accelerating Neural Rendering via a Unified Streaming Architecture

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Feng, Yu, Lin, Weikai, Liu, Zihan, Leng, Jingwen, Guo, Minyi, Zhao, Han, Hou, Xiaofeng, Zhao, Jieru, Zhu, Yuhao
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866929457582833664
author Feng, Yu
Lin, Weikai
Liu, Zihan
Leng, Jingwen
Guo, Minyi
Zhao, Han
Hou, Xiaofeng
Zhao, Jieru
Zhu, Yuhao
author_facet Feng, Yu
Lin, Weikai
Liu, Zihan
Leng, Jingwen
Guo, Minyi
Zhao, Han
Hou, Xiaofeng
Zhao, Jieru
Zhu, Yuhao
contents Neural Radiance Field (NeRF) has emerged as a promising alternative for photorealistic rendering. Despite recent algorithmic advancements, achieving real-time performance on today's resource-constrained devices remains challenging. In this paper, we identify the primary bottlenecks in current NeRF algorithms and introduce a unified algorithm-architecture co-design, Potamoi, designed to accommodate various NeRF algorithms. Specifically, we introduce a runtime system featuring a plug-and-play algorithm, SpaRW, which significantly reduces the per-frame computational workload and alleviates compute inefficiencies. Furthermore, our unified streaming pipeline coupled with customized hardware support effectively tames both SRAM and DRAM inefficiencies by minimizing repetitive DRAM access and completely eliminating SRAM bank conflicts. When evaluated against a baseline utilizing a dedicated DNN accelerator, our framework demonstrates a speed-up and energy reduction of 53.1$\times$ and 67.7$\times$, respectively, all while maintaining high visual quality with less than a 1.0 dB reduction in peak signal-to-noise ratio.
format Preprint
id arxiv_https___arxiv_org_abs_2408_06608
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Potamoi: Accelerating Neural Rendering via a Unified Streaming Architecture
Feng, Yu
Lin, Weikai
Liu, Zihan
Leng, Jingwen
Guo, Minyi
Zhao, Han
Hou, Xiaofeng
Zhao, Jieru
Zhu, Yuhao
Hardware Architecture
Graphics
Neural Radiance Field (NeRF) has emerged as a promising alternative for photorealistic rendering. Despite recent algorithmic advancements, achieving real-time performance on today's resource-constrained devices remains challenging. In this paper, we identify the primary bottlenecks in current NeRF algorithms and introduce a unified algorithm-architecture co-design, Potamoi, designed to accommodate various NeRF algorithms. Specifically, we introduce a runtime system featuring a plug-and-play algorithm, SpaRW, which significantly reduces the per-frame computational workload and alleviates compute inefficiencies. Furthermore, our unified streaming pipeline coupled with customized hardware support effectively tames both SRAM and DRAM inefficiencies by minimizing repetitive DRAM access and completely eliminating SRAM bank conflicts. When evaluated against a baseline utilizing a dedicated DNN accelerator, our framework demonstrates a speed-up and energy reduction of 53.1$\times$ and 67.7$\times$, respectively, all while maintaining high visual quality with less than a 1.0 dB reduction in peak signal-to-noise ratio.
title Potamoi: Accelerating Neural Rendering via a Unified Streaming Architecture
topic Hardware Architecture
Graphics
url https://arxiv.org/abs/2408.06608