Intel Ponte Vecchio graphics processors will be part of the Leibniz supercomputer center systems
Leibniz Supercomputer Center announced that it is ready to implement the second phase of the creation of its flagship supercomputer SuperMUC-NG. The system is planned to install future 4th generation Intel Xeon Scalable processors (Sapphire Rapids) and Ponte Vecchio accelerators based on Xe-HPC architecture.
As part of the first phase, 12,960 Intel Xeon Platinum 8174 processors were installed on SuperMUC-NG. The second phase of the SuperMUC-NG is planned to be completed in the spring of 2022. The supercomputer will receive 1-petabyte Distributed Asynchronous Object Storage (DAOS) based on Optane DC SSD and Optane DC Persistent Memory. Note that the system will use the one API software platform.
The second phase will be implemented jointly with Intel and Lenovo – 240 Intel computing clusters will be integrated into the Lenovo SD650-I v3 platform, cooled with warm water. The SuperMUC-NG upgrade is funded equally by the Free State of Bavaria and the German Federal Ministry of Education and Research through the Gauss Supercomputer Center (GCS). The computational performance of the supercomputer will be used for scientific research.
The head of the graphics division of Intel, Raja Koduri, confirmed cooperation on using the company’s technologies as part of the SuperMUC-NG supercomputer and shared images of the huge GPU Ponte Vecchio, consisting of many different modules, the purpose of which remains unknown.
We also add that this is not the only upcoming system with Intel Ponte Vecchio accelerators. For example, these accelerators will form the basis of the Aurora supercomputer, which was supposed to work in 2018, but due to delays from Intel, the launch is now planned only this year. Interestingly, the system for the Leibniz Center could be launched even earlier.
