Lenovo Delivers the Next Wave of Cooler High-Performance Solutions to Propel AI-accelerated Transformation
A New Era of Supercomputing Efficiency Arrives to Unlock AI’s Potential at Any Scale
ATLANTA - Lenovo, the world’s #1 supercomputer provider, showcased new high-performance innovations that deliver game-changing energy efficiency and accelerated computing to help scientific and industry researchers propel AI-powered transformation at any scale. Unveiled at Supercomputing 2024 (SC24), the new powerful, energy-efficient platforms will help scientists, engineers, and researchers push the boundaries of scientific discovery while making it possible for today’s enterprises to unlock AI’s potential, a press release stated by Lenovo.
As data center infrastructure takes center stage as the heart of producing AI and the digital economy, increasing energy demands must be met responsibly. Long known as the world’s largest gathering of high-performance computing (HPC) and research leaders, SC’s spotlight has grown to attract a wide range of c-level executives from across finance, healthcare, manufacturing and beyond as today’s business leaders aggressively pursue the deployment of AI solutions across their organizations. For today’s enterprises, HPC is part of the core engine that makes it possible to maximize AI’s potential.
Unlocking AI Factories with Cooler Supercomputing for All
AI and HPC capabilities are unlocking data-driven insights, accelerated outcomes, and advanced use cases, but organizations across every industry are presented with new thermal requirements as they push for ever-increasing levels of capacity to support AI workloads. Lenovo offered insight into the challenges businesses are facing in the AI era while diving into the future of cooling AI factories with the latest partner innovations.
Backed by more than a decade of expertise and 40+ patents for scalable supercomputing, Lenovo 6th Generation Neptune® Liquid Cooling helps tackle this problem by using liquid to remove heat, which is much more efficient than traditional air cooling. The revolutionary new chassis flips servers to a vertical orientation to fit today’s most powerful accelerated computing technologies in a compact system that is 100% liquid cooled, eliminating the need for power-consuming fans and dramatically improving energy efficiency. Designed for industry-standard 19-inch racks, organizations of any size can now leverage the high performance of accelerated computing, one tray at a time, using standard power in an open ecosystem.
“With this design, every data center provider can fit these systems into their rooms without any major modifications. That means no reinforced floors, no specialized power, no specialized doors to get into the room,” said Scott Tease, Vice President of Lenovo Infrastructure Solutions Group, Products. “Through the groundbreaking advancements in Lenovo Neptune liquid cooling, users can take advantage of the highest performance computing that is being delivered now and in the future.”
New 6th generation Lenovo Neptune water-cooled supercomputing servers use the latest processor and accelerator technologies to deliver scalable efficiency that unleashes AI. At the show, Lenovo announced support for the NVIDIA GB200 Grace Blackwell NVL4 Superchip, showcasing the newly available Lenovo ThinkSystem SC777 V4 Neptune.
Featuring the 6th generation open-loop, direct warm-water cooling innovation, enterprise customers of any size can use the Lenovo ThinkSystem SC777 V4 Neptune to run trillion-parameter AI models that bring products to market faster, at lower cost and with higher energy efficiency. At the core of the SC777, the NVIDIA GB200 Grace Blackwell NVL4 Superchip delivers revolutionary performance by combining four NVIDIA NVLink-connected Blackwell GPUs unified with two NVIDIA Grace CPUs over NVLink-C2C. Customers can achieve massive scale with the choice of NVIDIA Quantum InfiniBand or Spectrum-X Ethernet platforms. New Enterprise Reference Architectures with NVIDIA also help joint customers build their own AI factories — high-performance, scalable and secure data centers for manufacturing intelligence.
Also leveraging the 6th generation Lenovo Neptune design, the Lenovo ThinkSystem SC750 V4 Neptune is designed to handle technical computing, grid deployments, and analytics workloads in various fields such as research, life sciences, energy, engineering, and financial simulation. The system supports the new Intel Xeon 6900P-series 500W Granite Rapids AP processors and new 8800 MHz MRDIMM DDR5 memory. Its industry-leading direct water-cooling system ensures steady heat dissipation, allowing CPUs to maintain accelerated operation and maximize compute performance.
Stemming from its continued partnership with NVIDIA, Lenovo also showcased systems that will unleash AI acceleration for mainstream enterprise servers with the NVIDIA H200 NVL platform. For mainstream scientific computing applications across higher-ed, health care and life sciences, energy, and financial services, the Lenovo ThinkSystem SR675 V3 with NVIDIA H200 NVL delivers powerful performance and efficient scaling while enabling flexibility in configuration choice to meet the needs of a broad set of HPC workloads. Enabled across a full-stack AI portfolio,the Lenovo Hybrid AI Advantage with NVIDIA empowers organizations to rapidly turn intelligence into outcomes and bring AI-powered compute anywhere—with speed and ease.
Accelerating Clean Energy Transformation with Water Cooled Supercomputers
Supercomputers are used everywhere—from car and airplane design, oil field exploration, and financial risk assessment, to genome mapping and weather forecasting. Lenovo showcased how its water-cooled supercomputers are enabling accelerated transformation across the globe, reflecting the shift toward accelerated, energy-efficient supercomputing that is accessible to all.
In Italy, ENEA, the National Agency for New Technologies, Energy and Sustainable Economic Development, selected Lenovo for the installation of an HPC system at the Portici (Naples) hub to accelerate research activities on clean energy, in particular on nuclear fusion. This new HPC system, consisting of 758 nodes with 2 Intel® Xeon® Platinum 8592+ CPUs, will allow to bring the computational capabilities of CRESCO – Computational Center for Research on Complex Systems – the supercomputing system hosted in the Portici Research Center, from the current 1.01 to over 6.5 Petaflops,1 placing it at the top of the national scene in terms of processing power.
Sustainability was also a crucial factor for selection by ENEA, and Lenovo’s HPC solution guarantees better energy efficiency than in the past. The use of Lenovo Neptune direct water-cooling technology captures up to 98% of the heat produced by the supercomputer while saving the energy used for the fans. Thanks to the greater efficiency, the temperature of the CPUs does not reach critical values, avoiding the reduction of the maximum frequency of the cores.
Also in Italy, Cineca has signed an agreement with Lenovo for the installation of a new High HPC system for the scientific community to research fusion energy. The new infrastructure will be able to perform about 45 million billion operations per second (PFlop/s). It will be installed in the Cineca data center in Casalecchio di Reno (Bologna) at the end of 2024 and will become part of the Bologna Technopole ecosystem. The supercomputer will be dedicated to the numerical simulation of plasma physics and the structural analysis of advanced materials for nuclear fusion.
Cineca is a university consortium, one of the largest computing centers in Italy and one of the most advanced in the world for high-performance computing, home to numerous national and international HPC projects. The Lenovo supercomputer, selected through a competitive public tender, provides high energy efficiency, allowing the reduction of the system’s cooling electricity consumption by 15% through Lenovo Neptune™ Direct Water-Cooling technology.
As the number one provider of supercomputers in the world according to TOP500.org, Lenovo works closely with key industry-leading partners to develop, integrate, and deploy the technologies of exascale-level computing to organizations of all sizes.
Implementing High-Performance Private AI with Liquid Cooled Colocation
Elsewhere around the globe, for customers who need maximum AI and HPC performance but don’t have liquid-cooling infrastructure, Lenovo partners with leading colocation companies to offer Neptune® ready infrastructure for private AI workloads. These colocation partnerships are helping customers implement high-performance private AI even if they lack the data center footprint.
Partnering with Lenovo, Digital Realty, built onto its standardized high-density colocation offering with the introduction of Lenovo Neptune cooling technology in more than half its data centers globally, representing a significant leap in addressing the challenges of high-density workloads presented by AI across a full range of high-growth industries. Moreover, Digital Realty’s Private AI Exchange further enhances these capabilities by providing a secure and scalable environment for AI workloads. This exchange facilitates seamless integration and collaboration, ensuring that customers can leverage the full potential of their AI investments.
Recently, in partnership, Digital Realty and Lenovo enabled a global financial services provider to scale and distribute HPC capabilities that enable financial risk management strategies and calculations that are crucial to business success. By leveraging Digital Realty’s PlatformDIGITAL® to expand data center capacity with colocation, the financial services provider avoided the capital expenditure that would be involved in constructing their own data center facility and deployed 6x faster as a result of Digital Realty’s ability to retrofit existing data centers to meet advanced cooling requirements.
Through deploying Lenovo Neptune™ direct liquid-cooling infrastructure with high-density colocation, the company also achieves its goal of engineering the most energy-efficient solution possible with the possibility of critical heat recycling. The result is a 30% improvement in energy efficiency from using direct liquid cooling.
