Volume 37 (2018), number 1 pp. 45–59
CPU–GPU Parallel Framework for Real-Time Interactive Cutting
of Adaptive Octree-Based Deformable Objects
Shiyu Jia , Weizhong Zhang, Xiaokang Yu and Zhenkuan Pan
College of Computer Science and Technology, Qingdao University, Qingdao, Shandong 266071, P. R. China
email@example.com, firstname.lastname@example.org, email@example.com, firstname.lastname@example.org
A software framework taking advantage of parallel processing capabilities of CPUs and GPUs is designed for the real-time
interactive cutting simulation of deformable objects. Deformable objects are modelled as voxels connected by links. The voxels
are embedded in an octree mesh used for deformation. Cutting is performed by disconnecting links swept by the cutting tool and
then adaptively reﬁning octree elements near the cutting tool trajectory. A surface mesh used for visual display is reconstructed
from disconnected links using the dual contour method. Spatial hashing of the octree mesh and topology-aware interpolation of
distance ﬁeld are used for collision. Our framework uses a novel GPU implementation for inter-object collision and object self
collision, while tool-object collision, cutting and deformation are assigned to CPU, using multiple threads whenever possible.
A novel method that splits cutting operations into four independent tasks running in parallel is designed. Our framework
also performs data transfers between CPU and GPU simultaneously with other tasks to reduce their impact on performances.
Simulation tests show that when compared to three-threaded CPU implementations, our GPU accelerated collision is 53–160%
faster; and the overall simulation frame rate is 47–98% faster.
Keywords: deformable object, physics-based modelling, interactive cutting, adaptive octree mesh, GPU acceleration, multi-
ACM CCS: Computing methodologies—Massively parallel and high-performance simulations, Physical simulation
Simulating real-time interactive cutting of deformable objects has
important applications in surgical simulation, virtual sculpting and
video games. The ﬁnite element method (FEM) is now widely used
to simulate deformation. As for cutting, traditional methods split
each element swept by the cutting tool into several smaller ele-
ments. However, very small or degenerated elements can be created
that negatively affect the numerical stability of deformation. Cut-
ting methods based on adaptive octree mesh solve this problem
by embedding a surface mesh or voxel representation in an octree
mesh, the former is used for rendering and collision, while the latter
is used for deformation. During cutting, octree elements near the
cutting tool trajectory are adaptively reﬁned. Since the undeformed
shape of each octree element is a perfect cube and the size of an
octree element cannot be smaller than a voxel, the problem with
very small or degenerated elements is avoided.
Compared to traditional element splitting methods, the compu-
tational load of deformation, collision and cutting associated with
adaptive octree-based methods is much higher. As massively parallel
processors, GPUs have been widely utilized to accelerate deforma-
tion and collision of deformable objects. However, a GPU cannot
allocate memory on its own, and its performance is greatly reduced
when parallel tasks take different branches. Data transfers between
CPU and GPU can also become performance bottlenecks. Cutting
operation especially presents difﬁculty for GPUs due to constant
irregular changing of topology and the need to change buffer sizes
and update their data in GPU memory. To the best of our knowledge,
GPUs have never been used for general-purpose computation in a
deformable simulator supporting an adaptive octree-based cutting
method. The objective of this work is to design a parallel framework
utilizing both CPUs and GPUs for such simulator. CPU is also used
because some tasks cannot be efﬁciently implemented on GPU. Our
main contributions are:
(1) A novel multi-stage reduction algorithm for GPU implemen-
tation of collision processing.
(2) A novel CPU parallel implementation of cutting operation.
2017 The Authors
Computer Graphics Forum
2017 The Eurographics Association and
John Wiley & Sons Ltd. Published by John Wiley & Sons Ltd.