Improvement of the Embarrassingly Parallel Search for Data Centers
We propose an adaptation of the Embarrassingly Parallel Search (EPS) method for data centers. EPS is a simple but efficient method for parallel solving of CSPs. EPS decomposes the problem in many distinct subproblems which are then solved independently by workers. EPS performed well on multi-cores machines (40), but some issues arise when using more cores in a datacenter. Here, we identify the decomposition as the cause of the degradation and propose a parallel decomposition to address this issue. Thanks to it, EPS gives almost linear speedup and outperforms work stealing by orders of magnitude using the Gecode solver.
Unable to display preview. Download preview PDF.
- 1.Gecode 4.0.0 (2012), http://www.gecode.org/
- 5.Jaffar, J., Santosa, A.E., Yap, R.H.C., Zhu, K.Q.: Scalable Distributed Depth-First Search with Greedy Work Stealing. In: ICTAI, pp. 98–103. IEEE Computer Society (2004)Google Scholar
- 6.Kjellerstrand, H. (2014), http://www.hakank.org/
- 9.MiniZinc (2012), http://www.g12.csse.unimelb.edu.au/minizinc/
- 10.Nielsen, M.: Parallel search in gecode. PhD thesis, Masters thesis, KTH Royal Institute of Technology (2006)Google Scholar
- 11.Pedro, V., Abreu, S.: Distributed Work Stealing for Constraint Solving. CoRR, abs/1009.3800:1–18 (2010)Google Scholar
- 14.Schulte, C.: Parallel Search Made Simple. In: Proceedings of TRICS: Techniques for Implementing Constraint programming Systems, a post-conference Workshop of CP 2000, pp. 41–57 (2000)Google Scholar