New model enables anyone to run powerful simulations, complex calculations easily
Imagine a world in which complex calculations that could help improve someone’s health, clean up our planet or test an entrepreneur’s big idea don’t require specific skills.
That day may not be far away thanks to the EU-funded MSO4SC project, which is putting high-performance computing, cloud resources, mathematical simulation and other cutting-edge technologies directly in the hands of end users – from doctors, scientists and engineers to small businesses and policymakers.
The project’s e-infrastructure is designed to enable anyone to run powerful simulations and complex calculations easily, quickly and cheaply without specific IT skills or knowledge.
The solution developed by MSO4SC’s multi-disciplinary team brings together – via a simple user interface – all the tools, services, models and resources needed to run complex calculations for applications such as forecasting, prototyping or risk analysis.
‘A doctor wanting to run a simulation to assist in diagnosing a patient does not need or want to know about IT, mathematical models or simulation processes,’ says MSO4SC project coordinator Francisco Javier Nieto of Atos Research and Innovation in Spain. ‘They just want to introduce some parameters, click a button and see the results that will be useful for making their diagnosis – and, of course, they don’t want to pay high prices for doing this or wait days to get the results.’
Solutions for healthcare, energy and the environment
Test applications developed by the project partners highlight the many challenges that can be addressed and the types of innovation supported through the MSO4SC infrastructure.
In the healthcare sector, for example, doctors would be able to run novel diagnostic procedures for the early detection and clinical care of neurodegenerative diseases based on imaging patients’ eyes, while pharmaceutical firms could more effectively model chemical processes to develop novel forms of pain relief.
In the energy industry, engineers developing wind turbine blades would be able to more easily simulate the technical and structural performance of prototype designs, or improve the efficiency and safety of oil extraction.
Environmental engineers and policymakers would be able to run simulations to better forecast urban air quality or to improve the filtration of harmful substances from drinking water.
‘Our test applications cover just a few of the potential uses,’ Nieto says. ‘Ultimately, simplifying access to these tools will drive innovation, leading to better designs for generating renewable energy, new ways to diagnose disease, improved air quality and environmental protection, and enable future technologies across many sectors.’
Instead of the weeks or months needed at present to learn a mathematical modeling system, the MSO4SC tools enable developers to create a new modeling workflow from scratch in around 30 hours. Setting up a new simulation and inputting datasets on an existing application takes just minutes.
Combining high-performance and cloud computing
Significantly, running modeling processes is not only simple but cheap and fast as well, thanks to groundbreaking technological innovations achieved by the project.
Notably, the partners developed a solution to combine and optimize high-performance computing infrastructure from the Partnership for Advanced Computing in Europe (PRACE) with cloud resources from the European Open Science Cloud (EOSC).
Nieto says project researchers wanted to find a solution to make use of both kinds of systems in an optimal way, exploiting the best features of each one.
‘According to our experiments, combining both HPC and cloud in the right way generates performance improvements of 20 to 50 percent across different metrics compared to using each system individually,’ he adds. ‘Taking into account that running a simulation can take hours or days depending on the volume of data and the complexity of the calculations, the improvement is significant.’
The innovative solutions developed by MSO4SC have led to partnerships with other projects and attracted extensive interest from third-party developers, supported by the project’s involvement in the EU-MATHS-IN network for application-driven mathematical research.
The consortium is planning to maintain and grow the platform, building a community around the e-infrastructure that it hopes will continue to drive innovation and expand access to mathematical modeling and simulation in Europe and beyond.
No hay comentarios:
Publicar un comentario