Hewlett Packard Enterprise (HPE) recently unveiled its latest AI-enabled supercomputer named Aurora, which is the second exascale system developed by the company. Aurora achieved a speed of 1,012 exaflops on 87% of the system, making it the second fastest supercomputer in the world after Border, which was also created by HPE. This information comes from Top500, a recognized authority in supercomputing rankings.
The Aurora supercomputer was built in collaboration with Intel for the US Department of Energy’s Argonne National Laboratory. It features HPE Cray EX infrastructure with 75,000 compute node endpoints, 2,400 utility and storage network endpoints, and 5,600 switches. The system consists of 10,624 compute blades equipped with 21,248 Intel Xeon CPU Max Series and 63,744 Intel Data Center processors, along with GPU Maximum units for enhanced performance.
Researchers using Aurora will have the capability to run Generative AI Models to accelerate scientific discovery. Initial experiments conducted on the system include brain mapping, developing methods to detect neutrino interactions, and simulating fusion reactors. Rick Stevens, associate lab director at Argonne National Laboratory, described Aurora as a groundbreaking supercomputer that will significantly impact research initiatives globally.
For HPE, the integration of AI and HPC (High Performance Computing) has been a pivotal aspect of its strategy for almost a decade, starting with the acquisition of SGI and later Cray in 2019. The company has recently put a stronger emphasis on HPC and AI, particularly with the rise of generative AI technologies. At HPE Discover 2023, CEO Antonio Neri announced the company’s transition towards becoming an “AI-centric” organization, highlighting the importance of AI in its product offerings.
To support this direction, HPE introduced HPE GreenLake for large language models, an HPC-based infrastructure solution that allows customers to train and deploy large-scale AI projects without investing in specialized hardware. In addition to Aurora, HPE has developed other AI-focused HPC offerings such as Isambard-AI in the UK, Helios in Poland, and Venado in New Mexico.
Overall, HPE’s focus on integrating AI and HPC technologies, as showcased by the Aurora supercomputer, represents a significant advancement in the field of scientific research and computational capabilities. This strategic shift towards AI-centric solutions reaffirms the company’s commitment to driving innovation and empowering researchers with cutting-edge computing resources.
Article Source
https://www.itpro.com/infrastructure/data-centres/hpe-supercomputer-push-continues-with-aurora