Parallelization, Spatial Decomposition and Load Balancing of a Single Tree Level Forest Dynamics Simulator
Ključne riječi:
Forest dynamics simulation, individual tree, high-performance computing, parallel programmingSažetak
SPATE-HPC is a single tree level forest dynamics simulator capable of simulating very large forest areas. The size and shape of the simulation area as well as the number of trees are only restricted by the amount of memory available and subareas can be used for a tighter specification of the simulation.In this article we describe the parallelization of SPATE-HPC and its major computational challenges. We describe the domain decomposition methods and the load balancing strategy we have used to ensure good performance and scalability of huge simulations. We also describe how we have verified the correctness of the parallel implementation. Additionally we present performance measurement results for the simulator by running a fixed 32,300 ha simulation on 32-2,048 processors and a simulation with 1,000 ha/core on 32-2,048 cores. The results show that it is possible to simulate a forest area of more than 1,000,000 ha with several billion trees. MCFNS 2(1):30-40.
Reference
Beaumont, O., V. Boudet, F. Rastello, and Y. Robert, 2000. Matrix-matrix multiplication on heterogeneous platforms. In 2000 International Conference on Parallel Processing (ICPP00), p. 289.
Buckland, M., 2004. Programming game AI by example. Jones and Bartlett Publishers.
Coligny, F. D., et al., 2003. Capsis: Computer-aided projection for strategies in silviculture: Advantages of a shared forest-modelling platform. In Modelling forest systems. Reality, models and parameters estimation - the forestry scenario, pp. 320-323. CABI Publishing, Sesimbra, Portugal. Workshop on Reality, Models and Parameter Estimation, the Forestry Scenario.
Cybenko, G., 1989. Dynamic load balancing for distributed memory multiprocessors. Journal of Parallel and Distributed Computing 7(2):279-301.
Finkel, R. A., and J. L. Bentley, 1974. Quad trees a data structure for retrieval on composite keys. Acta Informatica 4(1):1-9.
Gropp, W., E. Lusk, and A. Skjellum, 1999. Using MPI: Portable Parallel Programming with the Message-Passing Interface. MIT Press.
Harja, D., and G. Vincent, 2008. Sexi-fs: Spatially explicit individual-based forest simulator - user guide and software. World Agroforestry Centre ICRAF, Bogor, Indonesia. http://www.worldagroforestry.org/.../SExI/ (last access date: Feb. 17, 2010).
Lin, C., 2003. Generating Forest Stands with Spatio-Temporal Dependencies. Ph.D. thesis, University of Joensuu.
Mascagni, M., and A. Srinivasan, 2000. Algorithm 806: Sprng: A scalable library for pseudorandom number generation. ACM Transactions on Mathematical Software 26:436-461.
Porte, A., and H. Bartelink, 2002. Modelling mixed forest growth: a review of models for forest management. Ecological Modelling 150(1-2):141-188.
Pretzsch, H., P. Biber, and J. Dursky, 2002. The single tree-based stand simulator silva: construction, application and evaluation. Forest Ecology and Management 162:3-21.
Warren, M., and J. Salmon, 1993. A parallel hashed oct-tree n-body algorithm. In Proceedings of the 1993 ACM/IEEE conference on Supercomputing, pp. 12-21. Portland, Oregon, United States.
Preuzimanja
Objavljeno
Broj časopisa
Rubrika
Licenca
Authors who publish with this journal agree to the following terms:- Authors grant the journal right of first publication, with the work simultaneously licensed under a Creative Commons Attribution License that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) following the publication by the journal, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See The Effect of Open Access).
