#	Pagina
attuale pagina	/open-h2020/projects/198333/results.html

Report

Teaser, summary, work performed and final results

Periodic Reporting for period 3 - E-CAM (An e-infrastructure for software, training and consultancy in simulation and modelling)

Teaser

Summary

The overall objective of E-CAM is to create, develop and sustain a European infrastructure for computational science applied to simulation and modelling of materials and of biological processes of industrial and societal interest. To achieve its objective, E-CAM uses the following three complementary instruments: (1) development, testing and dissemination of software targeted at end-user needs; (2) advanced training of current and future academic and industrial researchers in this area; and (3) multidisciplinary, coordinated, top level applied discussions to support industrial end-users (both large multinationals and SMEs) in their use of simulation and modelling.

E-CAMâ€™s activities aim at stimulating, supporting and enhancing simulation-based research and development (R&D) in industries. The development of this infrastructure also has an impact on academic research by creating a stable platform for scientific software development and optimization for high performance computers; by establishing strong links with industry; and by training researchers in different stages of their careers in aspects of computational science that will enhance their domain expertise.

E-CAM is built around the scientific community of CECAM (Centre EuropÃ©en de Calcul Atomique et MolÃ©culaire), and the computational and hardware expertise of PRACE (Partnership for Advanced Computing in Europe). We are a partnership of 16 CECAM Nodes, 3 PRACE centres, 11 industrial partners and one Centre for Industrial Computing (Hartree).

Our approach is focused on the following four scientific areas, critical for high-performance simulations relevant to key societal and industrial challenges: classical molecular dynamics, electronic structure calculations, quantum dynamics and meso- and multi-scale modelling. E-CAM develops new scientific ideas to algorithm development, optimisation, and parallelisation in these four respective areas, and the delivery of related training.

The goals of E-CAM are pursued via a set of coordinated actions and networking. Its main tasks are as follows:
-Software development targeted at specific end-users needs, and including testing, scaling and portability;
-Development and maintenance of the E-CAM repository, an open source repository of software modules and packages for simulations in academy, material and life science, engineering;
-Extended software development workshops (ESDWs) for production of modules for the repository based on input from the community and the industrial partners;
-Scoping workshops (SCOWs) for discussion and problem definition with industrial end-users;
-State-of-the-art workshops (SAWs) for monitoring developments, identifying new directions and transferring knowledge to industry;
-Support for academic and industrial research via a set of pilot projects sustained by E-CAM funded Postdoctoral Research Associates (PDRAs).

Work performed

All software and related documentation is stored on our open-source repository that is easily available for the industrial and academic communities. Software is certified according to best-practice programming, documentation and testing. Since the beginning of the project we have certified 118 software modules and have 45 which are work in progress.
We have disseminated our software through deliverables, our project website, our quarterly newsletters, and at E-CAM workshops and in international conferences. 19 scientific papers related to the software produced in E-CAM were published so far in international journals.
Access to leading-edge discussions on software and computer applications has also been provided through our programme of events published every year on our website at www.e-cam2020.eu/calendar/. Since the beginning of the project we held 7 SAWs, 6 SCOWs, 1 extreme-scale workshop and 13 ESDWs.
The outcomes of our SAWs and SCOWs has been shared with our industrial partners through the workshop scientific reports, and taken into account in future project activities. Support for academic and industrial research has also continued through our pilot projects, and E-CAM has supported the transfer of 43 certified software modules to industrial users, for further testing and application in the end user environment.
The participants to our ESDWs were trained in advanced computational methods and good practices for code development, documentation and maintenance. So far E-CAM has trained 230 scientists in different stages of their careers.
Our online training infrastructure continues to deliver content-specific training modules, focused on tools that can assist the participants of our ESDWs and any other interested groups to develop software for extreme-scale hardware. The content of the portal has increased to 248 items (videos and documents), viewed over 14\'000 times.

Final results

\"Here we highlight significant achievements within the project so far and future plans:
- The development of a load-balancing library applicable to all particle-based application codes, which is already being integrated into applications such as HemeLB and DL_MESO
- The development of a high throughput computing (HTC) library, that builds on top of the scalable analytics framework Dask and enables it to resiliently manage multi-node and multi architecture environments. This offers exciting possibilities in the areas of interactive supercomputing and burst supercomputing
- Significant progress was made in the multi-GPU version of DL_MESO
- The release of the v1.0.0 of OpenPathSampling, one of the key application codes of WP1, and the publication of two associated papers acknowledging E-CAM
- The development of PANNA, a machine-learning package that provides a comprehensive toolkit for creating neural network models for atomistic systems. Machine learning models are of particular interest to industrial applications where large amounts of data is generated during high throughput search for new materials, pharmaceuticals, etc.
- Software developed in the context of an E-CAM pilot project in collaboration with IBM is being used by the industrialist to engineer pulses for the experimental construction of superconducting qubits
- E-CAM organised the HPC sessions at CarpentryCon2018. Based on this activity, E-CAM presented a paper at the ISC 2019 conference during the workshop \"\"HPC Education and Training for Emerging Technologies\"\"
- Collaboration with the MaX CoE leverages the SCDM algorithm, implementing it in a fully-integrated framework that combines the AiiDA workflow management platform, Quantum ESPRESSO and Wannier90 to achieve user-intervention-free construction of MLWFs. This work paves the way for HTC of advanced materials properties with Wannier functions
- Protein based biosensor invented in the context of an E-CAM pilot project (and under IP protection) has been synthesised in vitro and cloned into a gateway entry vector â€“ the last stage of performance characterisation is expected to be completed soon.
\"