The FF4EuroHPC project is glad to announce the results of its second call for experiment proposals, referred to as Open Call-2. The call opened on June 22, 2021 and closed on September 29 the same year. A total of 70 experiment proposals were received and evaluated in a two-stage, consensus-based review effort involving 38 expert evaluators and 7 moderators. Each proposal was independently evaluated by two experts, and a consensus review was then written with the help of a moderator.

Of the 70 proposals, 26 were selected based on their impact, the soundness of the technical concept, the quality of the proposing consortium and the deployment of requested resources.

Country Statistics

The 70 proposals scored involve a total of 184 organisations from 27 countries, with Italy and Spain leading the field in the number of proposals with participation from their countries.

 

The 26 selected experiments comprise 79 partners from 22 countries, with organisations from Italy participating in 10 experiments (or 38% of the total). 

An important objective for the FF4EuroHPC Open Call-2 was to broaden the geographical base of selected experiments – in the preceding Open Call-1, 9 countries (Croatia, France, Germany, Greece, Italy, Montenegro, Serbia, Spain and the United Kingdom) accounted for all partners in the 16 selected experiments. For each country without an organisation participating in an Open Call-1 funded experiment, the selection process in Open Call-2 first picked the best proposal involving participants from that country and rated better than 17.0 points (of a maximum of 25), and then proceeded to select the highest-ranking experiments from the remaining set until the available funding volume of € 5 Million was expended. A total of 15 selected experiments (or 58% of the total) involve participants from 15 countries without organisations in Open Call-1 experiments (Austria, Bulgaria, Denmark, Estonia, Iceland, Latvia, Netherlands, North Macedonia, Liechtenstein, Poland, Portugal, Slovenia, Sweden, Switzerland, Turkey) – the specific geographical objective was achieved with flying colours. It is notable that all selected experiments involve organisations from at least one of the EuroHPC Joint Undertaking member countries (as listed here). 

Each FF4EuroHPC experiment is limited to funding of € 200,000 aggregated over all its partners; the 26 selected experiments require a total of € 4,968,270. Funding rules are equivalent to those of Horizon 2020 R&I projects, and the University of Stuttgart handles experiment finances and management.

 

Small and Medium Enterprise Participation

Of the 79 organisations participating in the selected experiments, 47(or 59%) are Small and Medium-Sized Enterprises (SMEs). This is significant since the main objective of the FF4EuroHPC project is to demonstrate the business value achievable by the use of HPC resources or services for SMEs. Non-SME organisations do include academic/public organizations such as HPC centres or universities as well as medium/large companies.

End-User Domains

Each of the selected FF4EuroHPC Open Call-2 experiment targets a specific economic end-user sector, and is designed to accelerate innovation and create business value in that domain. The Manufacturing domain (according to the commonly accepted NACE2 classification) accounts for 19 (or 73%) of the 26 selected experiments. 

The NACE2 Manufacturing domain contains a large number of different industrial activities. The selected experiments cover a wide spread of different activities, from the manufacturing of medical instruments to special-purpose machines. 

Computational Disciplines

Orthogonal to the industrial application area, the selected proposals rely on a total of 13 different HPC and computational disciplines, as depicted in the word cloud. Computational Fluid Dynamics (CFD) is leading with 12 projects, followed by the training of complex machine learning (ML) neural networks (5 projects), design space Optimisation and Finite-Elements techniques for structural simulation (4 projects each). A total of 6 projects (23%) use Machine Learning techniques. Each project has a primary and an optional secondary computational discipline, so the numbers do not add up to the total of 26 experiments. 

Experiments will start on 1st March and will be running for a maximum of 15 months. When the experiment is successfully concluded, it is resulting in a success story, inspiring the Industry community.