Ardelio Galletti

Power consumption is a significant challenge in the sustainability of computational science. The growing energy demands of increasingly complex simulations and algorithms lead to substantial resource use, which conflicts with global sustainability goals. This paper investigates the energy efficiency of different parallel implementations of a Cellular Potts model, which models cellular behavior through Hamiltonian energy minimization techniques, leveraging modern GPU architectures. By evaluating alternative solvers, it demonstrates that specific methods can significantly enhance computational efficiency and reduce energy use compared to traditional approaches. The results confirm notable improvements in execution time and energy consumption. In particular, the experiments show a reduction in terms of power of up to 53%, providing a pathway towards more sustainable high-performance computing practices for complex biological simulations.

Nowadays, the Machine Learning (ML) approach is needful to many research fields. Among these, the Environmental Science (ES) which involves a large amount of data to be processed and collected. On the other hand, in order to provide a reliable output, those data information must be assimilated. Since this process requires a large execution time when the input dataset is very huge, here we propose a parallel GPU algorithm based on a curve fitting method, to filter the starting dataset, by exploiting the computational power of the CUDA tool. The innovative aspect of the proposed procedure can be used in several application fields. Our experiments show the achieved results in terms of performance.

Machine Learning algorithms try to provide an adequate forecast for predicting and understanding a multitude of phenomena. However, due to the chaotic nature of real systems, it is very difficult to predict data: a small perturbation from initial state can generate serious errors. Data Assimilation is used to estimate the best initial state of a system in order to predict carefully the future states. Therefore, an accurate and fast Data Assimilation can be considered a fundamental step for the entire Machine Learning process. Here, we deal with the Gaussian convolution operation which is a central step of the Data Assimilation approach and, in general, in several data analysis procedures. In particular, we propose a parallel algorithm, based on the use of Recursive Filters to approximate the Gaussian convolution in a very fast way. Tests and experiments confirm the efficiency of the proposed implementation.

While the everything as a sensor is a typical data gathering pattern in the Internet of Things (IoT) applications in contexts such as smart cities, smart factories, and precision agriculture, among others, the use of the same technique in the coastal marine environment is still not explored at full potential. Nevertheless, when it comes to maritime scenarios, the application of IoT and networks of distributed sensors and actuators are still limited, even though the development of marine electronics and extreme network technologies are present for decades also in this area. In this paper, we first introduce the concept of the Internet of Floating Things (IoFT), which extends the IoT to the maritime scenario. Next, we present our latest implementation of the DYNAMO (Distributed leisure Yachts sensor Network for Atmosphere and Marine Observations) system, a framework for coastal data collection from sensors and devices deployed in marine equipment. To demonstrate the importance of IoFT data collection in the real-world environmental science context, we consider a scientific workflow for coastal water quality. The selected application focuses on predicting the spatial and temporal pattern of sea pollutants and their possible presence and time of persistence in the proximity of mussel farm areas in the Bay of Pozzuoli in Italy. The pollutants are simple Lagrangian particles, so the ocean dynamics play an important role in the simulation. Our results show that integrating crowdsourced bathymetry data in the workflow numerical model setup improves the accuracy of the final results, allowing for a more detailed spatial distribution pattern of the sea current driving the Lagrangian tracers.