Skip to main content

Eviden Unveils its BullSequana XH3500, a Converged Supercomputing Platform for the AI-HPC Era

Eviden, a brand of Atos Group, announced the launch of its BullSequana XH3500 supercomputer designed to support High performance computing (HPC) and Artificial Intelligence (AI) workloads. The system aims to enhance performance while being capable of scaling beyond exascale performance, integrating advanced hardware and a comprehensive software ecosystem to aid scientific and industrial progress.

The BullSequana XH3500 builds on the BullSequana XH3000 model and refines Eviden’s hybrid computing strategy by leveraging the foundational technologies established by its predecessor. It offers over 80% more electrical power per square meter and has 30% greater cooling capacity per kilowatt within the same rack footprint, positioning it as a high-density and energy-efficient solution.

This supercomputer is structured around a modular architecture, allowing organizations to customize configurations according to diverse needs, from simulation to large-scale AI training. The system's open design promises to keep customers free from vendor lock-in while facilitating integration of new technologies as they emerge.

The BullSequana XH3500 employs 100% fan-less Direct Liquid Cooling (DLC) systems, supporting warm water supply. This reduces the Total Cost of Ownership (TCO) and enables heat reuse initiatives. Complementary software enhances power and resource management, optimizing performance and energy usage in real-time.

Bruno Lecointe, VP of HPC, AI, and Quantum Computing at Eviden, stated, “The convergence of HPC and AI is redefining how we solve the world’s most complex challenges. Our mission is to accompany our customers throughout the entire lifecycle, accelerating applications and maximizing their HPC-AI investments.”

Addison Snell, CEO of Intersect360 Research, mentioned that the BullSequana XH3500 is designed to connect the capabilities of AI with scientific discovery, addressing user needs for application-ready deployments. This maintains flexibility and adaptability to new innovations without locking customers into a single vendor's offerings.