With the latest boom in Artificial intelligence (AI) technology, organizations are positioned to achieve goals and tackle issues they previously would not have been able to like unlocking new insights from their data, gaining operational efficiency, leveraging real-time quality control, and increasing productivity. However, organizations face a range of challenges implementing AI such as finding qualified staff who can set up the operations pipeline, dealing with the flow of data across applications, as well as implementing AI using a wide variety of hardware, frameworks, and toolkits.
Lenovo offers a complete AI solution to aid organizations in implementing AI by simplifying the adoption process with proven expertise and delivering faster insights with optimized infrastructure and pre-validated solutions. Lenovo hybrid AI solutions efficiently bring AI to customer data where they need it, anywhere from the pocket to the cloud.
Lenovo’s AI building blocks
Determining where to start an AI implementation can be a daunting task for organizations looking to gain insights from insurmountable amounts of data or get quick answers to tough questions. To support customers along their AI implementation journeys, Lenovo has several key AI building blocks – the AI Discover Lab, an AI partner ecosystem, and AI-ready, pre-validated infrastructure and solutions.
AI Discover Lab
In the AI Discover Lab new and existing customers can gain access to a full range of proof of concept services, access to benchmarking tools, technical field staff, workshops, best practices and information on industry consortiums. Lenovo experts including AI architects, engineers and researchers are available to support and help customers determine their needs and navigate implementing the solution.
AI Innovators Partner Ecosystem
Lenovo selects trusted partners across industries such as retail, manufacturing, healthcare, and finance to be a part of the Lenovo AI Innovators Partner Ecosystem designed to help customers find tailored solutions for their end-to-end operations, including computer vision, audio recognition, prediction, security, and virtual assistants. Today, the program consists of over 150 AI-ready solutions from over 50 partners who provide state-of-the art enterprise AI solutions enabling faster, safer, and more efficient deployments on Lenovo infrastructure.
AI Infrastructure and Solutions At-a-glance
Lenovo infrastructure includes high performance storage, compute, and edge solutions that scale with demand. Generative AI (GenAI) and Large Language Models (LLMs) can be run on various Lenovo products. The Lenovo ThinkSystem SR670 V2 server is based on the third-generation Intel® Xeon® Scalable processor family, and Intel Optane Persistent Memory 200 Series. This server supports eight double-wide GPUs including NVIDIA A100 and A40 Tensor Core GPUs, or the NVIDIA HGX A100 4-GPU with NVLink to provide optimal performance for AI workloads.
Lenovo’s newest design is the Lenovo ThinkSystem SR675 V3 featuring two 4th Gen AMD EPYC™ processors to help optimize infrastructure supporting GenAI and LLMs wherever customer data resides. AI models can run on GPUs and CPUs running in a data center or in the cloud depending on model size, latency and how frequently updates are needed. Additionally, MLPerf benchmark tests used for measuring training and inference performance of machine learning (ML) hardware, software and services were run on the Lenovo ThinkSystem SR675 V3. According to Dr. David Ellison, Chief Data Scientist for the Lenovo Infrastructure Solutions Group, “Lenovo achieved outstanding results in the MLPerf Inference v4.0 benchmarking results winning 21 out of 27 scenarios in that category.”
The Lenovo ThinkSystem SR675-V3 leverages innovative NVIDIA H100 NVL and L40S GPUs to enhance IT capabilities. The bare metal solution enables NVIDIA AI Enterprise software and Riva/NeMo Frameworks that provide access to AI and data science and development tools which accelerate data preparation, training at scale, and optimization of inferences and deploying at scale. The virtualized version uses AI-ready nodes, VMware’s vSphere, VMware vCenter, and NVIDIA AI Enterprise Suite to accelerate GenAI training and inference, gaining access to a diverse range of libraries and pre trained models.
Lenovo supports and practices responsible AI
Lenovo has a Responsible AI Committee chaired by Dr. Ellison that focuses on these six pillars: diversity & inclusion, privacy & security, accountability & reliability, transparency and environmental & social impact. “Responsible AI is not just a technological imperative but a moral one. It’s about creating systems that not only understand and enhance our lives but also respect our ethical values and societal norms. It’s the balance between innovation and empathy, ensuring that as we progress, we carry with us the principles of fairness, transparency, and accountability,” states Ellison.
AI Success Story: Lenovo utilizes AI in analyzing pit stops for NASCAR racing
Lenovo is working with Richard Childress Racing (RCR) to aid in evaluating how much fuel is pumped during pitstops at NASCAR races using the Lenovo ThinkSystem SR675 V3 servers and Smart Pitbox software based on Lenovo LiCO software. Live-feed overhead cameras in the RCR pitbox use a Lenovo developed AI model that accurately detects the moment a fuel plug is installed and gives the duration and calculation of fuel filled in real-time. This information is used to signal the driver when fuel is installed. The model can be used to estimate fuel needed until the end of the race or next pitstop to aid the team in race strategy planning.
To learn more and experience Lenovo’s AI solutions, email us at [email protected].