The strategic research agenda (SRA)

The current SRA 3 is the third edition of the SRA and the second full edition - SRA 2 in 2013 being an incremental update on SRA 1 in 2013). The final version was published in the second half of December 2017.

In 2019, we are working on the fourth edition of the SRA, which will be published at the end of Jan 2020.

What is the SRA?

The ETP4HPC Strategic Research Agenda (SRA) is our key deliverable. It is a document that outlines a roadmap for the achievement of exascale capabilities by the European High-Performance Computing (HPC) ecosystem. The milestones set in the SRA help the European Commission - now the EuroHPC Joint Undertaking - define the contents of the HPC Technology R&D Work Programmes as part of the European Research programme - Horizon Europe (or FP9). Four ETP4HPC representatives and three observers are present on the Research and Innovation Advisory Group of EuroHPC, which is chaired by the ETP4HPC chairman.

Our work on the SRA is supported by EXDCI-2 and involves many of our academic and industrial members as well as organisations from the related areas of Big Data and IoT (BDVA, HiPEAC...).

The structure of SRA 4 will differ from those of the previous SRAs. We have kept the concept of vertical areas (see figure), corresponding to the different elements of the supercomputing stack or the use of HPC, and each addressed by a dedicated Working Group:

System Architecture
System Hardware Components
System Software & Management
Programming Environment
I/O & Storage
Mathematics & Algorithms
Application co-design
Centre-to-Edge Framework

In addition, we are introducing a number (a maximum of 10) of ‘horizontal’ Research Clusters, addressing the most important research challenges for Work Programmes 2021-2022 and beyond. The final set of Clusters is not yet defined (although we have included some examples below) and we appreciate your help in defining them. Besides describing the technical content of the Clusters, the following parameters need to be detailed in order to assess their relevance:

Relevance & Impact (why do this?) - In general, why is this Cluster’s contribution important for the future work programmes?
Maturity (time to market) - How long will it take approximately for the technology, features and functions addressed by the cluster to turn into commercially available products or services?
Hurdles to overcome - What are the main R&D hurdles presented by the Cluster? These could be of e.g. technological, skills, infrastructure, financial investment, risk or legal nature.
Driving competencies in Europe - Where are the competences necessary for a successful execution of the R&D associated with the Cluster? In Europe and outside of Europe?
Cost of research to achieve significant uptake - What financial investment is necessary to produce commercially usable results?

This is the current SRA 4 Model:

A bit of history

ETP4HPC maintains the SRA on a continuous basis. The first SRA was issued in 2013 and we published an update (SRA 2) in November 2015. Our latest SRA 3 was published in Nov 2017.

The SRA is the work of experts associated with ETP4HPC. We also welcome comments from other stakeholders and experts. In order to facilitate this process we open Public Calls for Comments.

The SRA reflects the needs of the European HPC Technology providers and other members of the European HPC ecosystem, i.e.: PRACE (academic users), industrial HPC end users, Independent Software Vendors. We have also used the conclusions made available by other initiatives such as HiPEAC, EESI and PlanetHPC. We also talk to the Big Data community (represented by the Big Data Value Association).

The roadmap is developed through a series of internal and external workshops. A team of ETP4HPC experts supervise and manage the process, whilst the technical areas are developed by Working Groups led by members of ETP4HPC. All members of ETP4HPC can participate in the development of the SRA.

Example of SRA 2 (2015) Milestones: