基于CUDA的实时手术模拟。

CUDA-based real time surgery simulation.

作者信息

Liu Youquan, De Suvranu

机构信息

Department of Mechanical Aerospace and Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY, USA.

出版信息

Stud Health Technol Inform. 2008;132:260-2.

PMID:18391300

Abstract

In this paper we present a general software platform that enables real time surgery simulation on the newly available compute unified device architecture (CUDA)from NVIDIA. CUDA-enabled GPUs harness the power of 128 processors which allow data parallel computations. Compared to the previous GPGPU, it is significantly more flexible with a C language interface. We report implementation of both collision detection and consequent deformation computation algorithms. Our test results indicate that the CUDA enables a twenty times speedup for collision detection and about fifteen times speedup for deformation computation on an Intel Core 2 Quad 2.66 GHz machine with GeForce 8800 GTX.

摘要

在本文中，我们展示了一个通用软件平台，该平台能够在NVIDIA新推出的计算统一设备架构（CUDA）上进行实时手术模拟。支持CUDA的图形处理器利用128个处理器的能力，实现数据并行计算。与之前的通用图形处理器相比，它通过C语言接口具有更高的灵活性。我们报告了碰撞检测算法和后续变形计算算法的实现。我们的测试结果表明，在配备GeForce 8800 GTX的英特尔酷睿2四核2.66 GHz机器上，CUDA可使碰撞检测加速20倍，变形计算加速约15倍。