A multi‐GPU finite element computation and hybrid collision handling process framework for brain deformation simulation. (24th July 2018)

• Record Type: Journal Article
• Authors: Tian, Ye; Hu, Yong; Shen, Xukun
• Editors: Daniel, Thalmann
• Is Part Of: Computer animation and virtual worlds. Volume 30:Number 1(2019)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2018-07-24
• Subjects: nonlinear finite element method -- parallel algorithms -- parallel collision detection -- real time; Computer animation -- Periodicals; Visualization -- Periodicals; 006.6
• DOI: 10.1002/cav.1846
• Languages: English
• ISSNs: 1546-4261

Abstract

Abstract: This paper offers a fast multi‐graphics processing unit (GPU) parallel simulation framework to the problem of real‐time and nonlinear finite element computation of brain deformation. A load balancing strategy is proposed to ensure the efficient distribution of nonlinear finite element computation on multi‐GPU. A data storage structure is designed to minimize the amount of data transfer and make full use of the overlay technique of GPU to reduce the transferring latency between multi‐GPUs. We further present a fast central processing unit (CPU)–GPU parallel continuous collision detection and response method, which not only can deal with the collision between the brain and skull but also can handle the self‐collision of the brain. Our method can make full use of CPU and GPU to implement a parallel computation about deformation and collision detection. Our experimental results show that our method is able to handle a brain geometric model with high detail gyrus composed of more than 40, 000 tetrahedron elements. This can facilitate the fidelity of the current virtual brain surgery simulator. We evaluate our approach qualitatively and quantitatively and compare it with related works. Abstract : This paper presents a multi‐GPU simulation framework for brain deformation capable of solving continuous collision detection efficiently for geometric models of high complexity. We propose a load balancing strategy to distribute nonlinear finite element computation and a data storage structure to minimize the data transfer on multi‐GPU, as well as two hybrid CPU/GPU algorithms for parallel continuous collision detection and resolution. Besides, brain region partition, memory data coherence, and GPU streaming, among others, are also considered in order to optimize performance.