Multiscale Computing: From the Desktop to the Exascale

Aim and Description

Science is at its most powerful when it can not only convincingly explain the processes at work in natural phenomena but is able to predict what will occur before it does so. Predictions of real world events, such as weather forecasting, when a cement will set, the occurrence of an earthquake or what intervention to perform to cure a medical condition, all require the bringing together of substantial quantities of data together with the performance of many high-fidelity simulations before the event in question occurs. Such forms of calculation are among the most demanding known in computational science, as they need to be done rapidly, accurately, precisely and reliably, including quantification of the uncertainties associated with them.

These processes are also multiscale in nature, as their accuracy and reliability depend on the correct representation of processes taking place on several length and time scales. Only now, as we move toward the exascale era in high performance computing (HPC) can we expect to be able to tackle such problems effectively and, eventually, in a routine manner. This workshop will bring together a dedicated group of computational scientists to identify generic methodologies, algorithms and languages for multiscale computing, as well as software environments that support multiscale computing, cutting through specific scientific domains.

The goal of this workshop is to bring together a group of computational scientists and scientists from several different scientific disciplines to further explore generic methods, algorithms, specification and modelling languages, and software environments for multiscale computing. We will build upon the current understanding of multiscale modelling and focus on multiscale computing on state-of-the-art computing resources, with focus on high end HPC/Cloud ecosystems. The workshop will foster topical discussion, both in breakout sessions and plenary sessions, aiming at unifying ideas and setting future research agendas.

The workshop has a length of 5 days. We will structure the discussions around the following five topics:

Setting the scene, examples of state-of-the-art applications Frameworks for Multiscale Modelling and Simulation Multiscale Computing in general High Performance Multiscale Computing & Distributed Multiscale Computing Multiscale Computing in the Cloud.

Moreover, a number of sessions will be dedicated on reporting state-of-the-art applications of multiscale modeling and simulation.

The workshop will be considered a success if we are able to find a common methodology for multiscale computing, cutting through scientific disciplines, if we are able to further strengthen our community of scientists working on multiscale computing (proven by follow-up workshops or conferences, joint papers, special issues of journals), and if we are able to conceive a white paper on multiscale computing, that should form the basis for new programs, either at the national level, or at the European (Horizon 2020) level. As the European commission has embraced HPC as a strategically important field with high innovation capacity for Europe, we are in a position influence the European agenda and the outcome of the workshop will be taken up by participants to do so. As another tangible output of the workshop we will also aim at a special issue on multiscale computing in the Philosophical Transactions of the Royal Society A.