Robust polyhedral Minkowski sums with GPU implementation

Min Ho Kyung, Elisha Sacks, Victor Milenkovic

Research output: Contribution to journalArticle

7 Citations (Scopus)

Abstract

Abstract We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and GPU implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bound. We test the programs on 45 inputs with an error bound of 10-8. The CPU program outperforms prior work, including non-robust programs. The GPU program using 2688 CUDA cores exhibits a median speedup factor of 36, which increases to 68 on the 6 hardest tests. For example, it computes a Minkowski sum with a million features in 20 seconds.

Original languageEnglish (US)
Article number2332
Pages (from-to)48-57
Number of pages10
JournalCAD Computer Aided Design
Volume67-68
DOIs
StatePublished - Jun 4 2015

Fingerprint

Program processors
Convolution
Graphics processing unit

Keywords

  • GPU algorithms
  • Minkowski sums
  • Robust computational geometry

ASJC Scopus subject areas

  • Computer Graphics and Computer-Aided Design
  • Computer Science Applications
  • Industrial and Manufacturing Engineering

Cite this

Robust polyhedral Minkowski sums with GPU implementation. / Kyung, Min Ho; Sacks, Elisha; Milenkovic, Victor.

In: CAD Computer Aided Design, Vol. 67-68, 2332, 04.06.2015, p. 48-57.

Research output: Contribution to journalArticle

Kyung, Min Ho ; Sacks, Elisha ; Milenkovic, Victor. / Robust polyhedral Minkowski sums with GPU implementation. In: CAD Computer Aided Design. 2015 ; Vol. 67-68. pp. 48-57.
@article{82a2ed5ea7d04bd59fd225dc4521299e,
title = "Robust polyhedral Minkowski sums with GPU implementation",
abstract = "Abstract We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and GPU implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bound. We test the programs on 45 inputs with an error bound of 10-8. The CPU program outperforms prior work, including non-robust programs. The GPU program using 2688 CUDA cores exhibits a median speedup factor of 36, which increases to 68 on the 6 hardest tests. For example, it computes a Minkowski sum with a million features in 20 seconds.",
keywords = "GPU algorithms, Minkowski sums, Robust computational geometry",
author = "Kyung, {Min Ho} and Elisha Sacks and Victor Milenkovic",
year = "2015",
month = "6",
day = "4",
doi = "10.1016/j.cad.2015.04.012",
language = "English (US)",
volume = "67-68",
pages = "48--57",
journal = "CAD Computer Aided Design",
issn = "0010-4485",
publisher = "Elsevier Limited",

}

TY - JOUR

T1 - Robust polyhedral Minkowski sums with GPU implementation

AU - Kyung, Min Ho

AU - Sacks, Elisha

AU - Milenkovic, Victor

PY - 2015/6/4

Y1 - 2015/6/4

N2 - Abstract We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and GPU implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bound. We test the programs on 45 inputs with an error bound of 10-8. The CPU program outperforms prior work, including non-robust programs. The GPU program using 2688 CUDA cores exhibits a median speedup factor of 36, which increases to 68 on the 6 hardest tests. For example, it computes a Minkowski sum with a million features in 20 seconds.

AB - Abstract We present a Minkowski sum algorithm for polyhedra based on convolution. We develop robust CPU and GPU implementations, using our ACP strategy to eliminate degeneracy and to enforce a user-specified backward error bound. We test the programs on 45 inputs with an error bound of 10-8. The CPU program outperforms prior work, including non-robust programs. The GPU program using 2688 CUDA cores exhibits a median speedup factor of 36, which increases to 68 on the 6 hardest tests. For example, it computes a Minkowski sum with a million features in 20 seconds.

KW - GPU algorithms

KW - Minkowski sums

KW - Robust computational geometry

UR - http://www.scopus.com/inward/record.url?scp=84930677302&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84930677302&partnerID=8YFLogxK

U2 - 10.1016/j.cad.2015.04.012

DO - 10.1016/j.cad.2015.04.012

M3 - Article

AN - SCOPUS:84930677302

VL - 67-68

SP - 48

EP - 57

JO - CAD Computer Aided Design

JF - CAD Computer Aided Design

SN - 0010-4485

M1 - 2332

ER -