Cutting-Edge ‘Alps’ Supercomputer Unveiled

The research infrastructure in Switzerland has received a significant upgrade with the introduction of 'Alps,' a supercomputer powered by a Cray Supercomputer EX from Hewlett-Packard Enterprise (HPE) and equipped with 10,752 NVIDIA Grace Hopper superchips. This impressive setup places 'Alps' among the fastest computers globally, ranking 6th in the Top500 list of supercomputers as of June 2024 during its initial expansion phase. Upon full expansion, 'Alps' is projected to achieve a performance level approaching half an exaflop, equivalent to one billion billion floating point operations per second.

In a statement, Federal Councillor Guy Parmelin expressed that 'Alps' symbolizes a future driven by knowledge and progress. He highlighted the supercomputer as a tribute to pioneers exploring uncharted scientific territories and a pinnacle of cutting-edge technology. The collaboration between the Confederation, cantons, and communes was instrumental in bringing this advanced research infrastructure to fruition.

ETH Vice President for Research Christian Wolfrum underscored the enduring commitment of the Confederation and the ETH Domain to the development of 'Alps.' He emphasized the significance of the longstanding partnership between CSCS and industry in the installation of the supercomputer, stating, "'Alps' is not merely a stroke of luck but a result of the extensive collaboration between CSCS and industry."

'Alps' has been meticulously designed to cater to the extreme data and computing demands of scientific research. Wolfrum noted that the supercomputer empowers Swiss scientists to leverage artificial intelligence (AI) to its full potential, marking a significant milestone in advancing research capabilities in the country.

The supercomputer serves as a cornerstone of the Swiss AI initiative, spearheaded by ETH Zurich and EPFL, aimed at positioning Switzerland as a global leader in the development and deployment of transparent and trustworthy AI solutions. Andreas Krause, head of the AI Centre at ETH Zurich, highlighted the role of 'Alps' in training complex AI models for critical applications such as medicine and climate research, while also driving methodological advancements in AI transparency, reliability, and sustainability.

At the heart of the research infrastructure at CSCS, 'Alps' plays a pivotal role in enabling the creation of versatile software-defined clusters tailored to specific user community needs. Thomas Schulthess, Director of the CSCS, emphasized the cloud-native architecture of 'Alps,' ensuring the confidentiality of user data. Institutions like the Paul Scherrer Institute are leveraging this innovative platform to tackle scientific challenges and inspire breakthroughs in high-performance computing and extreme data analytics.

While the hardware forms a crucial component of 'Alps,' the software development aspect is equally vital. Engineers at CSCS collaborate closely with researchers to develop tools and software that bridge scientific inquiries with computational language, driving advancements in Swiss scientific research with tangible societal benefits. For instance, 'Alps' has enabled MeteoSwiss to adopt a higher-resolution weather forecasting model, better capturing Switzerland's intricate topographical features of mountains and valleys.