Eviden Expands its AI-dedicated Server Portfolio

Eviden announced AMD-powered additions to its BullSequana AI portfolio, designed to accelerate both the training and inference of large and very large AI models. These additions include two servers in the BullSequana AI 600H mid-range family and one blade server for the BullSequana AI 1200H high-end family.

The newly introduced servers respond to a key trend in AI infrastructure: the increasing need to handle large and very large models efficiently while balancing high performance with energy efficiency. By incorporating AMD GPUs, which tackle the most complex computational challenges—from scientific simulations to large-scale data processing and advanced modeling—and Eviden’s advanced cooling technologies, the systems enable organizations to run complex workloads while managing energy usage and operational costs.

The new BullSequana AI 642-H03 server harnesses the power of the latest AMD technology with the Instinct™ MI355X GPU, while the BullSequana AI 642-H02 server embarks AMD Instinct™ MI325X GPU. Both servers are hybrid cooled for better energy efficiency, with roughly 70% of generated heat captured by liquid cooling, while the remaining 30% is air cooled via fans. These servers are ideal for building training clusters for large models that require eight GPUs.

These servers integrate into the BullSequana AI 600H platform infrastructure. The BullSequana AI 600H mid-range family strikes a balance between compute performance and density, global total cost of ownership (TCO), capital expenditure (CAPEX) acquisition cost, and ease of deployment.

The BullSequana AI 600H family infrastructure is ideal for quick deployments in brownfield data centers but can also fit greenfield data center needs.

BullSequana AI 1242 blade server embeds the latest AMD Instinct MI355X GPU to tackle the most demanding AI workloads. It integrates into the BullSequana AI1200H platform infrastructure.

The BullSequana AI 1200H is Eviden’s flagship high-performance family platform, offering the best computing performance, density, and total cost of ownership (TCO) on the market. This computing solution features 100% Direct Liquid Cooling, utilizing Eviden’s proprietary 4th generation DLC technology, and can accommodate data center facility loop water temperatures of up to 40°C (104°F). Its all-in-one, flexible architecture allows for optimal customization.

Bruno Lecointe, VP, Global Head of HPC, AI and Quantum Computing, Eviden, Atos Group said “We are very pleased to see the continuation of our long-lasting relationship with AMD, giving rise to new AI-dedicated offerings. By focusing on both compute power and energy efficiency, we align our product strategy with key challenges in AI and HPC system design with the aim to support customers in academia, research, and enterprise sectors who are scaling up their AI infrastructure.”

“AMD is the trusted AI solutions provider of choice, enabling customers to tackle the most ambitious AI initiatives, from building large-scale AI cloud deployments to accelerating AI-powered scientific discovery,” said Negin Oliver, corporate vice president of business development, Data Center GPU Business, AMD. “AMD Instinct MI350 series GPUs paired with AMD ROCm software provide the performance, flexibility, and security needed to deliver tailored AI solutions that meet the diverse demands of the modern AI landscape.”

Source: Eviden media announcement