Open Access

ABSTRACT

PDSL and SDSL Parallel Visualization Algorithms for Large-scale Finite Element Analysis Data in Distributed Parallel Computing Environment

Jin Yeon Cho¹, Yun Hyuk Choi², You Me Song³, Chang Sik Kim⁴

1 Associate Professor, Department of Aerospace Engineering, INHA University, Incheon 402-751, Korea
2 Research Assistant, Department of Aerospace Engineering, INHA University, Incheon 402-751, Korea
3 Research Assistant, Department of Aerospace Engineering, INHA University, Currently Hyundai Motors
4 Research Assistant, Department of Aerospace Engineering, INHA University, Currently Samsung Electronics

The International Conference on Computational & Experimental Engineering and Sciences 2007, 4(3), 151-158. https://doi.org/10.3970/icces.2007.004.151

Abstract

In this work, PDSL(pre-detection sort last) and SDSL(strip-wise decomposition sort last) parallel visualization algorithms are proposed for efficient visualization of massive data generated from large-scale parallel finite element analysis through investigating the characteristics of distributed parallel finite element analysis procedure. The proposed parallel visualization algorithms are based on the sort last approach, and designed to be highly compatible with the characteristics of domain-wise computation in parallel finite element analysis. To investigate the performances of proposed algorithms, in-house software is developed by applying the binary tree network communication pattern along with the proposed sorting algorithms, and benchmarking test for parallel visualization is carried out.

---