arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2303.05735cs.ARcs.GRcs.LG
收藏

神经图形的硬件加速

Hardware Acceleration of Neural Graphics

UIUC(伊利诺伊大学厄巴纳-香槟分校);Intel Corporation(英特尔公司)

Muhammad Husnain Mubarik, Ramakrishna Kanungo, Tobias Zirr, Rakesh Kumar

AI总结:

本文针对神经图形在现有 GPU 上性能不足的问题,提出专用硬件架构 NGPC 加速输入编码与 MLP 内核,并结合 Vulkan 内核融合,实现最高 58 倍端到端性能提升,支持高分辨率实时神经渲染。

摘要:

驱动传统计算机图形的渲染和逆渲染算法最近已被神经表示(NR)所取代。NR 最近被用于学习场景的几何和材质属性,并利用这些信息合成照片级真实感图像,从而有望以可扩展的质量和可预测的性能替代传统渲染算法。在这项工作中,我们提出一个问题:神经图形(NG)是否需要硬件支持?我们研究了具有代表性的 NG 应用,结果表明,如果我们希望在当前 GPU 上以 60FPS 渲染 4k 分辨率,所需性能存在 1.5 倍至 55 倍的差距。对于 AR/VR 应用,所需性能与所需系统功耗之间的差距更大,达到 2 至 4 个数量级(OOM)。我们发现输入编码和 MLP 内核是性能瓶颈,对于多分辨率哈希网格、多分辨率密集网格和低分辨率密集网格编码,它们分别消耗应用时间的 72%、60% 和 59%。我们提出了一种 NG 处理集群(NGPC),这是一种可扩展且灵活的硬件架构,通过专用引擎直接加速输入编码和 MLP 内核,并支持广泛的 NG 应用。我们还通过在 Vulkan 中将其余内核融合在一起进行加速,与未融合的预处理和后处理内核实现相比,在内核级别实现了 9.94 倍的性能提升。我们的结果表明,对于多分辨率哈希网格编码,NGPC 在四个 NG 应用中平均实现了高达 58 倍的端到端应用级性能提升;对于缩放因子 8、16、32 和 64,性能收益分别为 12 倍、20 倍、33 倍和 39 倍。我们的结果表明,借助多分辨率哈希网格编码,NGPC 能够使 NeRF 以 30FPS 渲染 4k 分辨率,并使我们所有其他 NG 应用以 120FPS 渲染 8k 分辨率。

英文摘要:

Rendering and inverse-rendering algorithms that drive conventional computer graphics have recently been superseded by neural representations (NR). NRs have recently been used to learn the geometric and the material properties of the scenes and use the information to synthesize photorealistic imagery, thereby promising a replacement for traditional rendering algorithms with scalable quality and predictable performance. In this work we ask the question: Does neural graphics (NG) need hardware support? We studied representative NG applications showing that, if we want to render 4k res. at 60FPS there is a gap of 1.5X-55X in the desired performance on current GPUs. For AR/VR applications, there is an even larger gap of 2-4 OOM between the desired performance and the required system power. We identify that the input encoding and the MLP kernels are the performance bottlenecks, consuming 72%,60% and 59% of application time for multi res. hashgrid, multi res. densegrid and low res. densegrid encodings, respectively. We propose a NG processing cluster, a scalable and flexible hardware architecture that directly accelerates the input encoding and MLP kernels through dedicated engines and supports a wide range of NG applications. We also accelerate the rest of the kernels by fusing them together in Vulkan, which leads to 9.94X kernel-level performance improvement compared to un-fused implementation of the pre-processing and the post-processing kernels. Our results show that, NGPC gives up to 58X end-to-end application-level performance improvement, for multi res. hashgrid encoding on average across the four NG applications, the performance benefits are 12X,20X,33X and 39X for the scaling factor of 8,16,32 and 64, respectively. Our results show that with multi res. hashgrid encoding, NGPC enables the rendering of 4k res. at 30FPS for NeRF and 8k res. at 120FPS for all our other NG applications.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑