#USA #Supermicro #San Jose, California #AI Workloads #AMD EPYC 9006

Supermicro's Future-Forward H15 Server Portfolio

Supermicro has unveiled its latest H15 server lineup, which is powered by the cutting-edge 6th Generation AMD EPYC™ 9006 series processors, formerly codenamed 'Venice.' This new server range is a robust solution designed specifically for cloud computing, High-Performance Computing (HPC), and artificial intelligence (AI) workloads. The launch showcases remarkable benchmarks and multiple world records in SPEC CPU, exemplifying significant improvements in performance and speed compared to the previous generation.

Performance Enhancements with H15

The H15 servers represent a substantial technological leap forward, offering up to 33% more cores compared to their predecessors. They double the PCIe bandwidth and deliver an astonishing memory bandwidth of 1.6 TB/s per processor socket, a critical feature for AI inference and HPC applications. Supermicro’s next-generation solutions, named Helios and Helion, are optimized for large-scale agentic AI implementations, enhancing the capabilities offered to clients.

Insight from Supermicro's Leadership

Michael McNerney, Senior Vice President of Marketing at Supermicro, stated, "With the launch of the H15 portfolio, Supermicro and AMD provide customers with a faster and more efficient way to realize agentic AI infrastructure." He highlighted the advanced architecture of their Data Center Building Block Solutions (DCBBS), underscoring the advantages that come from a close technical collaboration with AMD. “This ensures our clients have the flexibility to confidently scale their AI, cloud, and enterprise workloads,” he added.

In support of these announcements, Dan McNamara, Senior Vice President and General Manager of Compute and Enterprise AI at AMD, emphasized the critical role the AMD EPYC processors play in advancing AI on a scalable level. He noted, “Our collaboration with Supermicro provides customers with optimized platforms that merge top-tier CPU performance with Instinct GPUs.” This integration allows organizations to rapidly implement AI infrastructure while improving energy efficiency and reducing total ownership costs.

Expanding the Portfolio

The H15 server portfolio offers unique configurations tailored for diverse needs, from powerful workstations to high-density multi-node systems and enterprise GPU servers. By providing a wide range of AMD EPYC 9006 processors and SP7 and SP8 sockets, Supermicro positions itself as a one-stop shop for CPUs, GPUs, and data center storage solutions. Furthermore, the recent expansion of their rack infrastructure with AMD technology includes innovative new platforms, ensuring that Supermicro continues to offer complete solutions from agentic AI to large-scale AI implementations.

Supermicro’s DCBBS architecture delivers a modular AI infrastructure that allows implementation flexibility from start to finish. This comprehensive approach equips customers for quick access to the latest data center solutions utilizing AMD technology. The DCBBS designs, when combined with AMD’s server architecture, facilitate a marketplace-ready AI infrastructure, achieving a 1.8 times higher token processing speed per second for host nodes running advanced AI models, compared to previous AMD CPU generations.

Showcasing the H15 Servers

The H15 server portfolio comprises an impressive array of over 20 optimized systems, including:

• - Hyper: A leading dual-socket platform designed for enterprise applications, AI inference, virtualization, and cloud workloads, built with advanced thermal design to accommodate the most powerful AMD EPYC processors.
• - CloudDC: A single-socket system tailored for large-scale cloud environments, based on the Data Center Modular Hardware System specification from the Open Compute Project.
• - FlexTwin™: A high-density 1 OU liquid-cooled computing system featuring two nodes, optimized for cloud-native deployments and hyperscale environments.
• - GrandTwin®: A high-density architecture consisting of two nodes in a 2U enclosure, each with a single processor socket, perfect for horizontally scalable environments, including object storage and HPC.
• - Petascale Storage: 2U all-flash storage platforms optimized for AI data lakes and high-performance computing environments, supporting systems up to 4.8 PB.
• - SuperBlade®: The H15 6U SuperBlade allows diverse blade configurations, optimizing space, performance, and efficiency.

Collaborating for the Future

Supermicro and AMD's partnership extends to the introduction of Helios and Helion solutions, tailored for rack-scale architecture, offering a full-fledged AI infrastructure. Helios focuses on maximized GPU performance, powered by AMD Instinct MI455X GPUs, and facilitates swift training and inference processes for advanced AI models. Meanwhile, Helion boasts exceptional computing performance driven by the latest AMD EPYC 9006 processors, equipped for flexible deployment.

A Commitment to Innovation

Supermicro's commitment to providing optimized, complete IT solutions is evident through its strategic focus on cutting-edge technologies in AI, enterprise, cloud computing, and 5G/edge IT infrastructure. As a global leader rooted in San Jose, California, Supermicro designs and produces its products in-house to achieve optimal performance and environmental efficiency.

With the new H15 servers now available, businesses looking to leverage next-generation AI capabilities can do so with certainty and ambition. For more details on these state-of-the-art solutions, check out Supermicro's offerings to empower your IT infrastructure growth.