Another AI supercomputer from HPE: Champollion lands in France
That's the second in a week following similar system in Munich also aimed at researchers
HPE is lifting the lid on a new AI supercomputer – the second this week – aimed at building and training larger machine learning models to underpin research.
Based at HPE's Center of Excellence in Grenoble, France, the new supercomputer is to be named Champollion after the French scholar who made advances in deciphering Egyptian hieroglyphs in the 19th century. It was built in partnership with Nvidia using AMD-based Apollo computer nodes fitted with Nvidia's A100 GPUs.
Champollion brings together HPC and purpose-built AI technologies to train machine learning models at scale and unlock results faster, HPE said. HPE already provides HPC and AI resources from its Grenoble facilities for customers, and the broader research community to access, and said it plans to provide access to Champollion for scientists and engineers globally to accelerate testing of their AI models and research.
Those purpose-built AI technologies refer to the HPE Machine Learning Development Environment, a software platform which forms part of the HPE Machine Learning Development System that HPE launched last month.
Coincidentally, the HPE Machine Learning Development System is also based on AMD-based Apollo computer nodes fitted with Nvidia GPUs, making it likely that Champollion is essentially an incarnation of the HPE Machine Learning Development System.
- Atos pushes out HPC cloud services based on Nimbix tech
- HPE, Cerebras build AI supercomputer for scientific research
- VMware claims 'bare-metal' performance on virtualized GPUs
- Nvidia teases server designs for Grace-Hopper Superchips
The actual Champollion hardware specified by HPE comprises 20 HPE Apollo 6500 Gen10 Plus server nodes, with 160 Nvidia A100 GPUs, and Nvidia Quantum InfiniBand networking. The HPE Machine Learning Development System starts at four nodes, but customers have the option to scale up.
If it is based on that platform, each Apollo node will have 4TB of memory and 30TB of NVMe local storage, with HPE Parallel File System Storage optional. The HPE Machine Learning Development Environment runs atop this and provides an integrated platform for building and training models, compatible with popular machine learning frameworks such as TensorFlow and PyTorch.
HPE said that Champollion is currently available to select users and will be released to broader community access in the near future for users to begin developing and training their models.
Earlier this week, HPE also unveiled another AI supercomputer at the Leibniz Supercomputing Center in Munich, pairing an HPE Superdome Flex server with a Cerebras CS-2 specialized AI system. That system is also designed to accelerate applications for the scientific and engineering community. ®