This paper shows simulations aspects of a Scenario-based run-time task mapping application. The application falls into the category of hydrologic prediction based on meteorological forecast. These applications demand computational resources, which depend on the scenario. In our case the study is focused partly on the rainfall-runoff model, the uncertainties that have to be computed with time constraints and with a minimum requirement of quality (i.e. precision). The main aim of this paper is to detect the simulation aspects and the trade-offs (such as power vs. time) which give a runtime manager running a safety-critical system. It shows two scenarios, the first a multi-core machine where several instances of the model have to compete for resources. And the second one presents the range of High Performance Computing resources needed to compute such model that can vary significantly depending on the scenario.
|Titolo:||Simulation of a runoff model running with multi-criteria in a cluster system|
|Data di pubblicazione:||2015|
|Appare nelle tipologie:||04.1 Contributo in Atti di convegno|
File in questo prodotto:
|CSCS15.pdf||pre print||Publisher’s version||Accesso riservato|