Periodic Reporting for period 2 - EuroLab-4-HPC (EuroLab-4-HPC: Foundations of a European Research Center of Excellence in High Performance Computing Systems)
Reporting period: 2016-09-01 to 2017-08-31
• Obj. 1. To create a portfolio of assets, including HPC platforms, technologies and align the agendas of the best research groups across the HPC stack on a roadmap for HPC system research.
• Obj. 2. To dene a curriculum for HPC systems and best education/training practices to foster future European technology leaders inHPC systems.
• Obj. 3. To accelerate uptake of new HPC system technologies in the commercial arena.
• Obj. 4. To build a community where the objectives and activities of EuroLab-4-HPC are communicated and disseminated to all HPC stakeholder groups from technology via business to venture capital.
• Obj. 5. The fifth objective is sustainability. The project will explore and identify the most appropriate legal and governance structure and business models to build a non-profit Centre of Excellence in HPC.
• WP1 has coordinated the project. It has organized monthly steering committee meetings physically or virtually. In addition, in the first year, it set up the infrastructure for the two project grant instrument calls (cross-site visit projects and business prototyping projects) and overlooked that all milestones were met and that project results were disseminated to the general public as well as to the European Commission in a timely fashion.
• WP2 has delivered a roadmap for research on HPC systems targeting 7-15 years ahead. In the second year, a final version of the roadmap has been submitted as deliverable D2.2. This contributes to Objective 1. Moreover, WP2 has carried out call for cross-site proposals. At the closing of the call, 14 proposals had been submitted out of which 10 were selected for funding. These accomplishments directly contribute to Objective 1. WP2 also mentored ten projects for business prototyping. Five of them later submitted a BPP proposal of which two were selected for funding.
• WP3 has validated a HPC curriculum and best practices for online learning. In the first year WP3 made an inventory of courses relevant for an HPC curriculum in HPC systems. Furthermore, WP3 also made an inventory of best practices in on-line training that is reported in D3.1. In the second year, WP3 has focused on validating both the proposed HPC curriculum and best practices for on-line learning (reported in D3.2).
• WP4 has piloted business prototyping as a method for accelerating innovation. In the first year, WP4 carried out a first selection of BPPs. 12 proposals were submitted. In the end, 3 proposals were accepted and successfully concluded. In the second year, the activities focused on analysing how to improve the BPP process. In the improved process, the number of submissions was raised to 20 and three BPP projects were granted and carried out.
• WP5 has organized a large number of dissemination and engagement events. In the second year WP5 organized in total 9 events collocated with HiPEAC and EXDCI conferences: Organization of a workshop on open-source software was organized as part of HiPEAC computing systems week in Dublin; organization of a workshop on disruptive technologies at the HiPEAC 2017 conference; participation in three workshops at HiPEAC2017; participation in a panel session on open-source software in the HiPEAC CSW in Zagreb; organization of a roadmap workshop in the HiPEAC CSW in Zagreb; organization of workshop in the HPC Summit 2017 (in Barcelona); and finally, organization of joint event at the ACM Europe Conference in 2017 Barcelona.
• WP6 has established a model for self-sustainability by looking into organizational issues and business models. In the first year, WP6 has developed a strategic plan for the agenda of a future Centre of Excellence in HPC systems and has also carried out a survey with interviews of stakeholders to identify successful models for a CoE. In the second year, the business model has been validated through a survey including major stakeholders in the HPC segment from academia to industry. The plan has been adjusted to conclude on a sustainability model for funding of EuroLab-4-HPC activities.
• [I1] Strengthened European research and industrial leadership in the supply, operation and use of HPC systems;
• [I2] Contribution to the realisation of the ETP4HPC Strategic Research Agenda;
• [I3] Development of competitive European technology for building and exploiting a wide range of next-generation extreme performance computing systems;
• [I4] Reinforced cooperation in international endeavours on HPC software and systems towards exascale;
• [I5] Structuring the efforts of stakeholders for implementing the European HPC strategy;
• [I6] European Excellence in Exascale Computing systems.
The outcome of EuroLab-4-HPC has been successful in meeting its objectives. We argue that its achievements are expected to match the expected impacts described in the call as follows.
With respect to I1, EuroLab-4-HPC has managed to link 2 communities together: the HiPEAC community focusing on high-performance and embedded computing with the HPC community through roadmap, curriculum & innovation activities that have been disseminated in global events joining these communities. On top of this, roadmap activities have targeted cross-cutting issues linking systems to applications and linking the ETP4HPC agenda with long-term challenges.
Concerning I2, EuroLab-4-HPC has developed a long-term HPC roadmap in collaboration with the ETP4HPC SRA (with an Exascale goal) that goes beyond Exascale computing and reviews future technological challenges and opportunities.
As for I3, EuroLab-4-HPC has made several contributions. First, it has fuelled developments by collaborations in HPC research projects and innovation projects through grants with good outcomes. Second, it has linked the top academic and industrial research groups in HiPEAC with the HPC community through several engagement events. Third, it has fuelled collaborations between user, supply, research and VC communities through its innovation work package.
Regarding I4, EuroLab-4-HPC has made several contributions. First, it has put together a roadmap based on the frontier of research in HPC. We expect this roadmap will have significant impact on international research efforts. Second, through its collaboration with key European HPC stakeholders (PRACE, ETP4HPC & EXDCI) in engagement events, EuroLab-4-HPC has now joined excellence in research with users, suppliers.
As for I5, EuroLab-4-HPC has carried out several measures: First, it has united the premiere research community in HPC (HiPEAC) with the European HPC community. Second, it has linked it to the HPC user/supply community in Europe. Third, it has engaged key stakeholders in the HPC innovation network to participate in EuroLab-4-HPC events and initiatives (the BPP process in particular).
Finally, as for I6, EuroLab-4-HPC has united Europe’s best academic institutions to deliver on the Exascale challenges through driving these institutions towards alignment to a long-term HPC systems research agenda.