Burn probability maps (BPMs) are among the most effective tools to support strategic wildfire and fuels management. In such maps, an estimate of the probability to be burned by a wildfire is assigned to each point of a raster landscape. A typical approach to build BPMs is based on the explicit propagation of thousands of fires using accurate simulation models. However, given the high number of required simulations, for a large area such a processing usually requires high performance computing. In this paper, we propose a multi-GPU approach for accelerating the process of BPM building. The paper illustrates some alternative implementation strategies and discusses the achieved speedups on a real landscape.
A Multi-GPU Approach to Fast Wildfire Hazard Mapping / D’Ambrosio, Donato; Di Gregorio, Salvatore; Filippone, Giuseppe; Rongo, Rocco; Spataro, William; Trunfio, Giuseppe, Andrea. - 256:(2014), pp. 183-195. [10.1007/978-3-319-03581-9_13]
A Multi-GPU Approach to Fast Wildfire Hazard Mapping
TRUNFIO, Giuseppe, Andrea
2014-01-01
Abstract
Burn probability maps (BPMs) are among the most effective tools to support strategic wildfire and fuels management. In such maps, an estimate of the probability to be burned by a wildfire is assigned to each point of a raster landscape. A typical approach to build BPMs is based on the explicit propagation of thousands of fires using accurate simulation models. However, given the high number of required simulations, for a large area such a processing usually requires high performance computing. In this paper, we propose a multi-GPU approach for accelerating the process of BPM building. The paper illustrates some alternative implementation strategies and discusses the achieved speedups on a real landscape.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.