Starts 1 Mar 2024 14:30
Ends 1 Mar 2024 17:30
Central European Time
Search
Search in Conferences:
MHPC DAY 2024 and 10 years of MHPC
ICTP
Leonardo Building - Budinich Lecture Hall
Strada Costiera 11 34151 Trieste Italy
PROGRAMME
Welcome & Institutional Greetings
Lectio Magistralis by Prof. Dr. Thomas C. Schulthess
“Alps”: A research infrastructure for science with extreme-scale data and computing needs
In the spring of 2021, we unveiled plans for the "Alps" infrastructure, aiming to establish it among the leading AI-capable supercomputers. Now, in spring 2024, we are proud to announce the integration of NVIDIA's cutting-edge Grace Hopper (GH200) superchips into our Cray Shasta EX infrastructure, provided by HPE. This significant upgrade propels "Alps" into the forefront of computational technology as a versatile, cloud-native supercomputing platform.
"Alps" distinguishes itself with a cloud-native architecture that positions it as both a pre- and post-exascale supercomputer. It achieves the pre-exascale distinction by delivering several hundred petaflops of computing power in 64-bit precision, while it reaches into the exascale with many exaflops in 32- and 16-bit precision. Its post-exascale status stems from an inherently flexible design, which not only supports but enhances the specific requirements of various scientific domains. This includes accommodating their unique workflows and data needs, showcasing its versatility and forward-thinking engineering.
Currently, "Alps" facilitates traditional high-performance and high-throughput computing (HPC and HTC) workloads, as well as specialized applications such as numerical weather prediction (NWP) for MeteoSwiss's operational forecasting. The introduction of the GH200 superchips enhances our infrastructure's ability to train large-scale neural networks and conduct high-resolution global climate simulations, marking a significant leap forward in computational science.
In his presentation Prof. Schulthess will explore how the "Alps" infrastructure achieves this remarkable blend of performance and flexibility, showcasing its role in advancing computational science. He will also discuss the emerging opportunities presented to computational science by AI-designed processors.
Presentation of the Best Thesis 2022-2023
HPC, a critical tool for industry and diverse areas of science - Alexander Ochoa (LEONARDO)
MHPC Success Stories
- Geometry of deep learning representations, three years after MHPC - Alberto Cazzaniga (AREA Science Park)
- From Astrophysics to Applied ML: my journey on the MHPC shuttle - Serafina Di Gioia (ICTP)
- My last ten years: building a career in the scientific community with a Master in HPC - Stefano Piani (OGS)
MHPC Partners Stories
- Investing in HPC training, research, and exploitation: opportunities and challenges of the OGS commitments for the Earth Sciences national community - Stefano Salon (OGS)
- Breaking boundaries: HPC's role in industry advancement - Karbalaii Mahya (Danieli)
Panel “Training next generations of HPC experts”
Aperitif
