Artificial intelligence and Big Data are going to revolutionize the HPC domain, by offering new tools for analyzing the immense amount of data generated by scientific experiments and modern simulation tools.
To this end, ACROSS aims at developing an innovative execution platform for serving emerging workflows that mix large and complex numerical simulations along with Big Data analytics, machine learning and deep learning techniques. ACROSS will co-design this innovative platform revolving around a large set of advanced hardware and software technologies towards the Exascale level of performance. As such, hardware acceleration technologies will be of primary interest for ACROSS to get the maximum performance while consuming the least amount of energy. In detail, ACROSS will move beyond traditional GPUs, integrating and exploiting FPGAs and Neuromorphic hardware.
Looking forward, ACROSS will envision the adoption of the future European Processor Initiative (EPI) outcomes, also to promote European first-class technology. Energy-efficiency, efficient usage of computing resources and application performance will be achieved through the design and implementation of an innovative workflow and resource management system. As such, a multi-level approach will be used to define the workflows, manage the resource allocation beyond the traditional queuing systems by using a machine learning based approach, efficiently execute workload tasks on the allocated heterogeneous resources. Three pilots have been selected to demonstrate the capabilities of the ACROSS platform in three relevant industrial and scientific domains: (1) aeronautics pilot; (2) weather, climate, hydrological and farming pilot; and (3) energy and carbon sequestration pilot.
The ACROSS platform will be backed by super-computing and Cloud resources provided by European HPC centres (IT4Innovations and CINECA), including a new generation of super-computers.