Hewlett Packard Enterprise announced this week that it has received an order to build a new supercomputer for Singapore’s National Supercomputing Center (NSCC). The new system is powered by AMD’s Epyc ‘Milan’ processors as well as Nvidia’s A100 compute GPUs; it is eight times more powerful than its predecessor.
NSCC’s supercomputer uses HPE’s liquid-cooled Cray EX architecture and will be one of the first systems in the world based on AMD’s 3rd generation Epyc “ Milan ” processors based on the Zen 3 microarchitecture (approximately 100,000 cores in total) as well as 352 of Nvidia GPU’s A100 calculations.
The system will consist of approximately 900 nodes and is expected to provide an aggregate of up to 10 PFLOPS of raw FP64 computing power. NSCC and HPE do not degrade CPU and GPU performance in the system, although 352 Nvidia A100 GPUs can offer around 3.4 FP64 PFLOPS as well as a whopping 110 FP16 PFLOPS. Meanwhile, the system will come with a 10PB Cray ClusterStor E1000 storage system with over 300 GB / s read / write speed.
The combination of AMD processors and Nvidia GPUs enables the new supercomputer to be used for both artificial intelligence (AI) and machine learning (ML) as well as traditional high performance computing (HPC).
Initially, the NSCC will use its new HPC machine for biomedicine, genomics, disease, engineering and high-resolution meteorological modeling, but over time it can be used for a wide range of applications requiring the IA and / or HPC.
NSCC and HPE expect the new supercomputer to be operational sometime in early 2022. The system costs SGD $ 40 million (approximately $ 30.16 million).