Name: A Blueprint for LLM Cluster Architecture
Start: 2024-05-14T17:00:00Z
End: 2024-05-14T17:00:46.000Z
Location: BrightTALK
Rating: 4.8

Supermicro® (NASDAQ:SMCI), with 30+ years of leadership in Enterprise, cloud, AI, and 5G Telco/Edge Infrastructure solutions, pioneers the industry with Building Block Solutions® and Green Computing servers. Its customizable, efficient, and sustainable IT offerings such as liquid-cooling technology redefine performance standards and environmental responsibility. Subscribe to see upcoming and on-demand webinars. 

As modern data centers evolve into AI Factories, energy management is critical for sustainable operations. For many operators, this shift means no unified visibility and energy management problems on the grid. Traditional management tools can't keep up with the energy demands of GPU clusters and liquid cooling.

In this session, experts from Supermicro and Bentaus discuss the unifying of compute and energy management. Discover how the integration of Supercloud Composer (SCC) and Power Asset Orchestrator (PAO) allows operators to gain total visibility into AI Factories while optimizing everything from GPU telemetry to real-time grid pricing.

What You Will Learn:

• The evolution of data centers into AI Factories
• Grid Management in the age of High Performance Computing (HPC)
• Future trends in data centers, AI Factory and grid management

Powering AI Factories: Unifying Smart Grid Management and Real-Time Visibility

Supercomputers need to stay fast, reliable, and stable at scale. Running Lambda’s modular AI factories takes both liquid-cooled, GPU-dense Supermicro servers and the Lambda’s operational expertise to manage AI cloud environments at scale.

Join our webinar and learn how Lambda, together with Supermicro, solves the operational problems of modular, GPU-dense AI factories and validates the architecture, testing, and operational practices needed for mission-critical deployments.

Building the Next Generation of AI Factories with Supermicro and Lambda

Retail is undergoing a major transformation as physical stores evolve into intelligent, data-driven environments. With the rise of AI and language-based analytics, retailers have the means to get meaningful, live insights from video, sensors, and sales systems.

This shift opens a vast business potential: use cases for hyper-personalized customer engagement, real-time inventory management, and loss prevention are reshaping how retailers connect with customers and drive growth.

Join our live panel with experts from NVIDIA, Everseen, and Supermicro, as they discuss how to bring the Intelligent Store concept to practice and deliver real business impact today. Gain insights into practical strategies, proven use cases, pitfalls to avoid, and the technologies that are setting the pace for the next generation of smarter, more connected retail experiences.

Smarter Stores, Bigger Impact: AI Strategies Retail Leaders Need Now

Many companies have implemented data lakes for big data analytics workloads using Hadoop. While Hadoop was state of the art twenty years ago, the coupled storage and compute architecture of Hadoop increases the cost, limits the elasticity, and is complex.  Modern data lakes using object storage decouple the compute and storage resources, use a flexible scale-out architecture and have a lower cost.

This webinar will discuss how object storage from MinIO using Supermicro storage servers and AMD CPUs can modernize legacy HDFS implementations and prepare for future enterprise AI and analytics workloads. 

We'll cover:

• Legacy Hadoop data lakes compared with modern object storage data lakehouses
• Selecting the best storage server platforms from Supermicro using AMD CPUs
• Implementing and migrating from Hadoop to MinIO AIStor
• Best practices for deploying object storage for data lakehouse workloads

Modernizing Legacy Hadoop Data Lakes with Object Storage: MinIO and AMD

Successful enterprise AI implementations utilize the company’s proprietary internal data to power AI agents and models, unlocking new revenue sources and improving customer and employee experience. Object storage can play a key role in the data platform powering the AI models.

In this session, experts from Supermicro and DDN will discuss the workflow and data pipeline for implementing enterprise AI and how object storage can enable it. We’ll also cover the features of the Infinia data Intelligence platform solution from DDN along with Supermicro’s optimized hardware platforms and provide practical suggestions on how the object solution is implemented for enterprise AI workflows.

Speakers:

Paul McLeod, Product Director, Storage at Supermicro

Balaji Venkateshwaran, VP of AI Product Management at DDN

Wendell Wenjen, Director, Storage Market Development at Supermicro (Moderator)

Accelerating AI Workflows: Unleashing Object Storage with DDN

Join Supermicro in the upcoming webinar featuring TensorWave and Modular along side AMD as we take a look at some of the technologies enabling the next wave of AI workloads including training and inferencing. Learn from the leaders of both companies as they share their stories, and share what makes the partner ecosystem unique to delivering an end to end solution on Supermicro’s latest liquid-cooled AI systems powered by AMD Instinct™ MI350 series GPUs.

How TensorWave and Modular Deliver Liquid-Cooled AI Systems With Supermicro and AMD

The telecommunications landscape is shifting as accelerated computing becomes mainstream. This transformation unlocks the potential to monetize the telco edge with AI applications, by offering GPU infrastructure to internal and 3rd party users while simultaneously running 5G/6G Radio-Area-Network (RAN) software. This unified approach is no longer a future concept—it's a critical business imperative for creating new revenue streams and monetizing the edge.

But how can you build a platform that can generate new revenue while meeting the strict demands of RAN?

Join experts from Supermicro and Aarna.ml for a deep dive into a comprehensive reference architecture for AI RAN distributed inference. We will unveil a complete, cloud-native solution designed to transform your edge sites into dynamic, monetizable AI platforms that can also run RAN software.

In this webinar, you will learn about:

• The Case for AI RAN and Edge AI: We will explore the rationale for distributed edge inference and why this will be a tremendous opportunity for telcos.

• Solving Multi-tenancy at the Edge: How Aarna.ml's GPU Cloud Management Software (CMS) provides the crucial layer for secure, isolated, and on-demand resource allocation.

• The Architecture Blueprint: A detailed look at a proven hardware and software architecture, featuring Supermicro's NVIDIA GH200 Grace Hopper systems, integrated with a unified enterprise AI and infrastructure management platform.

• The Path to Monetization: Discover how this architecture enables you to repurpose idle RAN capacity for new revenue streams, from enterprise AI to generative AI services at the edge.

This session is ideal for Telco Solution Architects, Edge AI Infrastructure leaders, Network Architects, and Platform Engineers responsible for designing and deploying the next generation of converged AI and RAN infrastructure.

Monetizing the Edge: A Blueprint for AI-Ready RAN Infrastructure

As the adoption of AI use cases in retail, manufacturing, smart spaces, and other industries continues to expand, enterprise infrastructure performance at the edge needs to keep up. Finding the right balance between performance and TCO is vital for a successful and sustainable business case. Additionally, enterprises are embracing specialized AI models for predictive, generative, physical, and agentic AI, making low-latency data processing for real-time decision-making even more critical. 

Join us as we discuss Edge AI use cases to demonstrate how businesses can drive growth and operational excellence and how Supermicro's edge portfolio is designed to deliver the required AI performance at the edge.

Powering Edge AI with Future-Ready Infrastructure Solutions

Join us to learn how Vultr, a leading cloud provider, created a state-of-the-art AI cloud with Supermicro servers using AMD EPYC™ CPUs and Instinct™ GPUs.

Hear experts from Supermicro, AMD, and Vultr as they cover:

• Industry needs and challenges: How Vultr needed a trusted vendor capable of delivering end-to-end system optimization and flexibility
• How working with Supermicro and AMD allowed accelerated deployment using high-end server configurations and options
• A deep dive into the Supermicro Servers used

How Vultr Created an AI Cloud with Supermicro and AMD

Learn how Applied Digital Corporation is using Supermicro servers to expand its cloud services offerings. Using some of the latest Supermicro GPU-optimized servers, Applied Digital is able to offer AI and HPC services across North America. With bare-metal offerings, users get the highest performance. Join experts from Supermicro and Applied Digital to see how they are working closely together to create new and innovative solutions for demanding users.

Applied Digital Builds AI Cloud with Supermicro AI Optimized Servers

Join our webinar and learn how Supermicro's expansion of its NVIDIA Blackwell system portfolio enables you to tackle a diverse range of AI infrastructure environments.

Featuring Supermicro experts, we'll explore:

• The challenges facing organizations in AI Factory Deployment
• New NVIDIA HGX B200 systems featuring DLC-2, the next generation of liquid solutions
• New front I/O and rear I/O systems: choose the best configuration for your computing environment

Optimize Blackwell GPU Solutions for Your AI Infrastructure Environment

Join our webinar to discover how Supermicro's multi-node systems tackle demanding workloads under limited resources and budget. As the demand for greater computational power intensifies, enterprises have urgent needs to squeeze more processing power. Supermicro's multi-node systems - from the Twin family to SuperBlade - are expertly engineered to maximize compute density while reducing space, power, and cooling expenses, ultimately delivering exceptional performance and efficiency.

Key topics include:

• The challenges companies are facing in maximizing compute density 
• How multi-node systems are an ideal solution for companies seeking to balance high performance with budget and space constraints.

High-Density Computing with Supermicro Multi-Node Systems

Featuring ServeTheHome’s Patrick Kennedy, this webcast will dive deep into Supermicro’s broad range of Enterprise AI solutions to explore how organizations can implement AI for almost any industry and environment. Product experts from NVIDIA and Supermicro will also be on hand to share how Supermicro systems and NVIDIA PCIe GPUs including the new RTX PRO 6000 Blackwell Server Edition can further enhance AI and visual computing performance and serve as the foundation for the Enterprise AI Factories of the future.

Key topics will include:

· The specific requirements of modern AI infrastructure and how Supermicro’s industry-leading range of AI-enabled systems with first-to-market architectures maximizes performance from the latest GPUs—including RTX PRO 6000 Blackwell Server Edition—to enable AI anywhere from the data center to the network edge.

· The challenges enterprises face as AI continues to be integrated into more and more industries; even those which were not traditionally considered as technology industries. How the combination of Supermicro systems and NVIDIA GPUs delivers an unrivalled range of options for any organization looking to implement AI, and how Supermicro’s building block approach enables significant customization at the system level for specific workload requirements, and can be combined with NVIDIA GPUs to serve almost any Enterprise AI workload.

· The infrastructure requirements for deploying enterprise AI at scale—such as Enterprise AI Factories—and the advantages of building complete solutions. We’ll discuss how Supermicro NVIDIA-Certified systems with RTX PRO 6000 Blackwell will be the foundation for Enterprise AI Factory validated designs, guaranteed to work with Spectrum-X networking, NVIDIA-Certified storage, and NVIDIA AI Enterprise software for a full-stack solution.

A New Era of Enterprise AI with Supermicro and NVIDIA RTX PRO™ Server

Object storage has come a long way from its origins in backup and archive.  The use of object storage for unstructured data is well known but high-performance All-Flash storage systems from Supermicro along with new optimized storage protocols make this a great fit for both AI training and inference workloads.  

This session will cover:

- the history of object storage and how it's used today
- the features of the object storage software solutions from Supermicro partners along with optimized hardware platforms
- practical suggestions on how the object solution is implemented for enterprise and service provider AI workflows

Scaling AI Training and Inference with Object Storage using Supermicro Systems

Join experts from Supermicro and AMD and get an in-depth look at our new H14 solutions, powered by the AMD Instinct MI350 series GPUs. Organizations today are looking for options to maximize performance and scale their large AI training models and high-speed inference workloads, all while reducing total cost of ownership. 

In this webinar, we’ll explore:

• Supermicro's new H14 GPU Solutions.
• The capabilities of the AMD Instinct MI350.
• Strategies for choosing the best solution for your specific needs.

Exploring Supermicro's H14 Solutions Powered by New AMD Instinct™ MI350 Series GPUs

Get an exclusive first look at Supermicro DLC-2, the next generation direct-liquid-cooling solution. AI and HPC workloads are driving the demand for liquid-cooled data centers. Supermicro’s DLC-2 is the latest innovation in liquid cooling technology. Join Supermicro experts, as they discuss:
 
- Supermicro DLC-2 overview
- How DLC-2 can reduce data center power consumption by up to 40% and lower Total Cost of Ownership (TCO) by up to 20%
- Portfolio of liquid cooling rack scale designs

Supermicro DLC-2: Your First Look at Next-Generation Data Center Liquid Cooling

Featuring experts from Supermicro and AMD, join our webinar where we will highlight our latest solution ideal for cloud hosting providers, edge computing, content delivery networks, and online gaming. Discover a powerful balance of performance, density, scalability and affordability. 

We'll cover:

- Supermicro MicroCloud systems that can benefit your organization
- How Supermicro and AMD are transforming entry level server and dedicated hosting workloads
- Supermicro systems featuring AMD EPYC™ 4005 CPUs

Supermicro MicroCloud featuring AMD EPYC™ 4005: Max Density Bare Metal for Dedicated Hosting

Join our webinar to learn about Supermicro's latest single-socket server solutions designed to deliver exceptional performance and value for today's demanding workloads. Optimized with the latest Intel® Xeon® 6 processors, single-socket servers offer some key advantages over multi-processor systems.

Featuring experts from Supermicro and Intel, we’ll cover:

- Single-socket benefits: power, efficiency and TCO
- Range of Workloads: AI, Cloud, HPC, and Edge
- Supermicro product families optimized for single-socket Intel Xeon 6 processors

How to Maximize Rack Performance and Efficiency with Single-Socket Servers

Large language models (LLM) are pushing the boundaries of traditional data center infrastructure with demanding AI workloads and increasing complexity. How do you design high performance clusters that can handle the extreme needs of LLMs while achieving optimal performance and scalability?  

Join our webinar with Supermicro experts and learn: 

• LLM SuperCluster Architecture: Gain insights into designing the blueprint for LLM training infrastructure, including compute, storage, networking, and cooling considerations. 

 • How To Optimize Power & Cooling: Discover strategies to maximize efficiency and performance in your rack scale deployments. 

• Real-World Success Stories: See how industry leaders are leveraging Supermicro’s Rack Scale SuperClusters 

• How Supermicro's LLM SuperCluster Solution can empower your AI ambitions and address the unique challenges of LLM training with a proven solution.

A Blueprint for LLM Cluster Architecture

Data Center

Data Center Infrastructure

Data Center Management

Artificial Intelligence

Generative AI

Machine Learning

The data center management community focuses on the holistic management and optimization of the data center. From technologies such as virtualization and cloud computing to data center design, colocation, energy efficiency and monitoring, the BrightTALK data center management community provides the most up-to-date and engaging content from industry experts to better your infrastructure and operations. Engage with a community of your peers and industry experts by asking questions, rating presentations and participating in polls during webinars, all while you gain insight that will help you transform your infrastructure into a next generation data center.

The storage community on BrightTALK is made up of thousands of storage and IT professionals. Find relevant webinars and videos on storage architecture, cloud storage, storage virtualization and more presented by recognized thought leaders. Join the conversation by participating in live webinars and round table discussions.

Storage

Network infrastructure professionals understand that a reliable and secure infrastructure is crucial to enabling business execution. Join the network infrastructure community to interact with thousands of IT professionals. Browse hundreds of on-demand and live webinars and videos to learn about the latest trends in network computing, SDN, WAN optimization and more.

Network Infrastructure

As an IT professional, many of the problems you face are multifaceted, complex and don’t lend themselves to simple solutions. The information technology community features useful and free information technology resources. Join to browse thousands of videos and webinars on ITIL best practices, IT security strategy and more presented by leading CTOs, CIOs and other technology experts.

A Blueprint for LLM Cluster Architecture

Presented by

About this talk

Supermicro