The Roots of this Project

The reasons why my colleague and I have decided to choose project 2109 ‘Benchmarking HEP workloads on HPC facilities’ are several. Unravelling the deeper reasons behind the project, we find a very interesting collaboration that emerged in 2020 between CERN, PRACE and other organisations with the aim of joining forces to improve the HPC world. That is why, in this post, we will describe in more detail where this collaboration comes from and what are its objectives in case it could be of interest to readers.

As is common knowledge in the world of supercomputers, as science advances the technological resources needed to sustain the necessary computing power are increasing. This is why, in order to make the most of the different existing heterogeneous architectures, a collaboration has arisen between different pioneering entities in high-performance computing (HPC). The members of the collaboration are CERN, the European Organization for Nuclear Research; SKAO, the organisation leading the development of the Square Kilometre Array radio-telescope; GÉANT, the pan-European network and services provider for research and education; and PRACE, the Partnership for Advanced Computing in Europe.

This is why this collaboration arises from the need to join forces in order to obtain better results from HPC calculations. This is where the project that my colleague and I are carrying out comes from. During the first 18 months since the agreement was signed between the different partners, the objective has been to develop a benchmarking test suite and a series of common pilot ‘demonstrator’ systems.

The purpose of developing this Benchmark Suite is to determine the different performances that supercomputers within the reach of the four participating entities can have for the development of high-energy physics and astronomy calculations. Regarding the project related to the SoHPC that my colleague Miguel and I are carrying out, we are currently trying to implement the Unified European Application Benchmark Suite (UEABS) used by PRACE in the Benchmark Suite developed by CERN. UEABS is a set of currently 13 application codes taken from the pre-existing PRACE and DEISA application benchmark suites, and extended with the PRACE Accelerator Benchmark Suite.

Not only that, but one of the main aims of this collaboration is the development of training programmes so that researchers will be able to make the most of the different existing HPC architectures. So the roots of this project lie not only in scientific study, but also in the possible training of people so that science can continue to advance.

My colleague and I are very proud and excited to be able to contribute to a project that is the result of a collaboration between four pioneering European organisations with computational resources available all over the world. That is why, with this post, we encourage you to find out more about it through the news published on 20 July 2020 on the official CERN website, whose link you can find here.

See you soon in the following post!

María Menéndez Herrero, Miguel de Oliveira Guerreiro