DAVIDE (Development of an Added Value Infrastructure Designed in Europe) is an energy-aware Petaflops Class High Performance Cluster based on IBM Power Architecture and coupled with NVIDIA Tesla Pascal GPUs with NVLink. The innovative design of DAVIDE has been developed by E4 Computer Engineering for PRACE, with the aim of providing a leading edge HPC cluster showing higher performance, reduced power consumption and ease of use. A key technology of the architecture, developed in collaboration with the University of Bologna, is an innovative infrastructure for measuring, monitoring and capping the power consumption of the node and of the whole system.
With a focus on energy efficiency, in this workshop we describe in more detail the DAVIDE architecture and uses cases in many different scientific fields based on the users' experiences. These examples will demonstrate to the users how to exploit this system taking full advantage from it.
By the end of the course, participants will be expected to:
- have a good understanding of the technologies employed in the DAVIDE HPC supercomputer
- recognize the types of problem and application particularly suited to the cluster
- understand how the energy monitoring system can be used to improve the energy efficiency of applications
Researchers who may wish to use the cluster in the future for their research or simply those interested in the evolving field of energy monitoring and efficiency in HPC.