Subsequent part of NVIDIA Hybrid AI collaboration locations transformational AI within the fingers of each enterprise
March 18, 2024 – Immediately, at NVIDIA GTC, a worldwide AI convention, Lenovo introduced new hybrid AI options, in-built collaboration with NVIDIA, that ship the facility of tailor-made generative AI functions to each enterprise and cloud, bringing transformational capabilities to each {industry}. Stemming from the expanded engineering collaboration introduced with NVIDIA throughout Lenovo Tech World, the improvements assist enterprises and cloud suppliers with the vital accelerated computing capabilities wanted to achieve the AI period, taking AI from idea to actuality and empowering companies to effectively develop and deploy new AI use instances that drive innovation, digitalization and development.
The hybrid options are function constructed via engineering collaboration to effectively convey AI to buyer knowledge, the place and when customers want it essentially the most – from the pocket to the cloud – advancing Lenovo’s imaginative and prescient to allow AI for all and delivering time to market assist of breakthrough structure for the following technology of huge scale generative AI. Lenovo hybrid options, already optimized to run NVIDIA AI Enterprise software program for safe, supported and steady manufacturing AI, will now additionally present builders entry to the simply–introduced NVIDIA microservices, together with NVIDIA NIM and NeMo Retriever.
“Lenovo’s work with NVIDIA is pushing the boundaries of augmented intelligence for companies all over the place, taking AI compute to the place their knowledge lives with a complete portfolio of cutting-edge, hybrid AI options that may energy generative AI from just about wherever whereas supporting sustainability efforts,” stated Kirk Skaugen, President of Lenovo Infrastructure Options Group. “We’re at an inflection level the place new AI use instances are coming to market primarily based on enhancements in real-time computing, energy effectivity and ease of deployment. Via our partnership with NVIDIA, we’re delivering groundbreaking developments in effectivity, efficiency, and price which might be accelerating AI functions throughout each {industry} and serving to companies instantly use insights from their knowledge units, whether or not for upgrading the retail expertise, reimagining our cities, or unlocking the following stage of good manufacturing.”
“AI is a strong drive serving to companies unlock new insights from their knowledge and improve productiveness,” stated Bob Pette, Vice President of Enterprise Platforms at NVIDIA. “Lenovo’s new enterprise AI options, built-in with NVIDIA know-how, show a pivotal milestone in supercharging compute efficiency for AI, delivering hybrid techniques that companies can depend on to energy generative AI from just about wherever.”
As industries across the globe search to make use of AI for analyzing huge our bodies of knowledge, energy effectivity stays essential to creating the rollout of those compute-intensive workloads accessible for all companies. Lenovo has been on the forefront of enabling environment friendly, high-power computing with out compromise, pioneering Lenovo NeptuneTM liquid cooling know-how and rating #1 on the Green500 checklist for cutting-edge designs powered by NVIDIA GPUs that speed up computing capabilities whereas conserving issues cool even in high-heat, multi-GPU environments.
Lenovo ThinkSystem AI Servers: Catapulting Generative AI Inference and Effectivity with NVIDIA
Marking a game-changing milestone in effectively supercharging AI-workloads at scale, Lenovo unveiled the enlargement of the Lenovo ThinkSystem AI portfolio, that includes two new highly effective 8-way NVIDIA GPU techniques which might be purpose-built to ship huge computational capabilities with uncompromised energy effectivity to speed up AI implementation. Engineered for generative AI, pure language processing (NLP), and huge language mannequin (LLM) improvement, with time-to-market assist for the NVIDIA HGX AI supercomputing platform, together with NVIDIA H100 and H200 Tensor Core GPUs and the all-new NVIDIA Grace Blackwell GB200 Superchip, in addition to the superior, next-generation NVIDIA Quantum-X800 InfiniBand and Spectrum-X800 Ethernet networking platforms.
Lenovo ThinkSystem AI servers with NVIDIA B200 Tensor Core GPUs are set to energy the brand new period of generative AI. The NVIDIA Blackwell structure’s generative AI Engine, alongside the quicker NVIDIA NVLink interconnect and enhanced safety capabilities, propel the information heart into a brand new period. With as much as 25X extra real-time inference to speed up trillion-parameter language fashions, B200 GPUs are designed for essentially the most demanding AI, knowledge analytics and HPC workloads.
With an ultra-efficient energy utilization effectiveness (PUE) of 1.1., the brand new Lenovo ThinkSystem SR780a V3 is a 5U system that makes use of Lenovo Neptune™ liquid cooling. Through the use of direct water cooling of CPU and GPUs and NVIDIA NVSwap know-how, the system can maintain most efficiency with out hitting any thermal limits. For greater than a decade, Lenovo’s industry-leading Lenovo Neptune direct water-cooling resolution, which recycles loops of heat water to chill knowledge heart techniques, enabling clients to understand as much as a 40% discount in energy consumption and a 3.5x enchancment in thermal efficiencies in comparison with conventional air-cooled techniques. As an {industry} metric used to find out the power effectivity of a knowledge heart, PUE and energy consumption are among the many top-tracked sustainability strategies, in keeping with the “Uptime Institute Annual World Information Middle Survey 2021.” As a result of liquid cooling gives a extra energy-efficient different to air, the system can drive greater sustained efficiency whereas consuming much less power. It additionally permits the ThinkSystem SR780a to slot in a dense 5U package deal, serving to to preserve useful knowledge heart actual property.
Additionally accelerating the portfolio, the brand new Lenovo ThinkSystem SR680a V3 is an air-cooled, two-socket system constructed for max acceleration for advanced AI with Intel processors and a alternative of NVIDIA GPUs. The extremely accelerated techniques ship huge computational functionality and use industry-standard 19-inch server racks, permitting for dense {hardware} configurations that maximize effectivity with out occupying extreme flooring house or requiring shelving.
Lenovo additionally showcased a brand new Lenovo PG8A0N – the final word 1U node for AI and that includes open-loop liquid cooling for accelerators, supporting the NVIDIA GB200 Grace Blackwell Superchip. The brand new GB200 arrives to energy a brand new period of computing. The superchip delivers 45X quicker real-time giant language mannequin (LLM) inference, 40X decrease TCO, and 40X much less power. In shut collaboration with NVIDIA, Lenovo will ship GB200 rack techniques that supercharge AI coaching, knowledge processing, engineering design and simulation.
Clients can run NVIDIA AI Enterprise, an end-to-end cloud-native software program platform for the event and deployment of production-grade AI functions, on Lenovo’s portfolio of NVIDIA-Licensed Systems. For prime-performance inference on common AI fashions from NVIDIA and its ecosystem, clients can run the NVIDIA NIM inference microservices included in NVIDIA AI Enterprise on Lenovo enterprise infrastructure.
The portfolio contains Lenovo’s XClarity Methods Administration, giving companies a simplified and centralized useful resource administration system, in addition to Lenovo’s Clever Computing Orchestration (LiCO), a unified platform that streamlines use of clustered computing sources for AI mannequin improvement and coaching and HPC workloads. The portfolio additionally helps the 4th and fifth technology of Intel Xeon Scalable Processors and gives thermal headroom for future higher-power GPUs.
Co-Designed NVIDIA MGX Options: Bringing Customized AI, NVIDIA Omniverse and HPC to CSPs and Enterprise
Lenovo is the main supplier of workstation-to-cloud assist for designing, engineering and powering NVIDIA OVX techniques and the NVIDIA Omniverse improvement platform. To assist corporations rapidly construct and deploy a variety of purpose-built AI, HPC and Omniverse functions, Lenovo is partnering with NVIDIA to construct accelerated fashions quicker utilizing NVIDIA MGX modular reference designs. Leveraging the designs, cloud service suppliers obtain custom-made fashions quicker with the supply of accelerated computing for AI and Omniverse workloads economically and at scale. Together with time-to-market assist for the NVIDIA H200 GPU, the techniques will allow scientists and researchers to deal with the world’s most difficult issues by accelerating advanced AI and HPC functions working terabytes of knowledge.
Lenovo’s new choices constructed with the NVIDIA MGX modular reference server design embrace:
- New Lenovo HG630N – MGX 1U – an open-standard server with Lenovo Neptune direct liquid cooling to cut back energy consumption whereas supporting the highest-performing GPUs.
- New Lenovo HG650N – MGX 2U – a extremely modular, GPU-optimized system that’s air-cooled, helps industry-standard racks and permits NVIDIA GH200 Grace Hopper Superchip
- New Lenovo HG660X V3 – MGX 4U – a system that helps as much as 8 600W NVIDIA GPUs in an air-cooled atmosphere, very best for NVIDIA Omniverse and AI workloads. Lenovo is a NVIDIA design accomplice for MGX 4U.
- New Lenovo HR650N – MGX 2U – a high-performing Arm CPU server with a number of cores and adaptability for storage and Entrance IO, leveraging the energy-efficient NVIDIAGrace CPU Superchip and supporting DPUs.
Lenovo ThinkStation Workstations with NVIDIA RTX: Generative AI at Your Fingertips
Lenovo enhances its knowledge science workstations by providing as much as 4x NVIDIA RTX 6000 Ada Technology GPUs for big AI coaching, fine-tuning, inferencing, and accelerated graphics-intensive workloads. This affords higher AI developer productiveness via automated workflows:
- New Lenovo Workstations with NVIDIA AI Workbench is a software program device that allows the event and deployment of highly effective AI options for inferencing, large-scale simulation, and demanding industrial and scientific workflows. NVIDIA AI Workbench makes generative AI and machine studying improvement simple for all builders.
- Now out there in Lenovo ThinkStation and ThinkPad workstations, the brand new NVIDIA A800 GPU designed particularly for AI permits safe, private knowledge science and Gen AI environments for organizations working with every kind of AI Workflows.
Lenovo ThinkStation PX with twin CPUs and as much as 4x NVIDIA RTX skilled graphics playing cards is essentially the most highly effective workstation available on the market in the present day. Lenovo’s AI-ready ThinkStation desktop and ThinkPad cellular workstation portfolio affords the broadest and strongest workstations for AI to deal with essentially the most intensive real-time AI resolution improvement. Be taught extra about Lenovo workstations and AI workflows in the present day.
Lenovo Skilled Companies: Quick Monitoring AI Journey and Accelerating Trade Development
Lenovo’s real-time options powered by NVIDIA are utilizing AI and Lenovo infrastructure to assist clients speed up smarter futures and notice new potential throughout industries. In retail, these options are analyzing buyer motion and habits to enhance site visitors stream and higher handle stock in actual time. In manufacturing, built-in Lenovo and NVIDIA options assist enhance security between employees and machines by refining meeting line inspection processes. Throughout cities, built-in Lenovo edge AI options are serving to metropolis planners to get essentially the most out of knowledge by optimizing house, infrastructure and asset utilization to assist handle individuals in site visitors flows and cut back energy consumption.
Lenovo’s newly introduced AI Companies Middle Of Excellence (COE) combines enterprise advisors, knowledge scientists, and AI-optimized infrastructure as-a-service to maximise outcomes, assembly clients the place they’re on their AI journey. Following Lenovo’s confirmed AI-readiness methodology masking safety, individuals, course of, and know-how, the AI COE affords the experience and capabilities to allow companies to fast-track their journey to implementing highly effective, accountable, and sustainable AI. Lenovo’s new AI skilled companies choices embrace:
- New AI Uncover – Helps clients uncover the “artwork of the potential” for AI. By conducting interactive workshops and assessments, wanting throughout the complete ecosystem, and mapping out the AI technique, Lenovo builds the blueprint for AI success.
- New Quick Begin Generative AI companies with NVIDIA – Leverages highly effective knowledge insights and achieves a aggressive benefit with Generative AI. Lenovo affords full-stack options to assist the whole product lifecycle, plus companies to implement, undertake, and scale Generative AI options.
- New TruScale GenAI As-A-Service – Offers AI capabilities via an as a service mannequin which will increase flexibility, scalability and predictability. With Lenovo TruScale, infrastructure is all the time proper sized to obtain, deploy and handle AI Innovation.
- Enhanced Skilled Companies for AI – Helps clients speed up their AI transformation by offering enterprise advisors, knowledge scientists, and AI-optimized infrastructure-as-a-service, making certain a seamless utilization of sustainable AI.
See new merchandise, applied sciences, and initiatives shaping the way forward for AI at NVIDIA GTC, working March 18-21 on the San Jose Conference Middle and on-line right here. For extra info on how Lenovo is working with NVIDIA towards a better, quicker future, go to https://www.lenovo.com/us/en/servers-storage/alliance/nvidia/.
