Summary

The partnership between VMware by Broadcom, Intel, and Dell Technologies simplifies the implementation of AI infrastructure, offering customers a seamless and efficient experience. The integration of Intel Xeon Scalable CPU with Advanced Matrix Extensions (AMX) within VMware Cloud Foundation on Dell VxRail presents a comprehensive solution consisting of software and hardware components that support AI workloads throughout the infrastructure without needing additional hardware such as GPUs. Customers may choose CPUs over GPUs for their AI infrastructure to better align costs with their specific workload requirements and to reduce power consumption. This approach helps right-size the infrastructure, avoiding over-engineered, costly solutions and achieving operational cost savings while meeting performance needs.

By using the capabilities of 4^th Gen Intel Xeon processors, businesses can fully harness the advantages of AI technology, leading to enhanced business results. This strategic approach broadens the reach of AI advantages by efficiently using current infrastructure resources and the powerful features of Intel Xeon processors equipped with AMX technology. Furthermore, the seamless compatibility with VMware Cloud Foundation on Dell VxRail environments enhances cost-effectiveness, accessibility, operational efficiency, and experience, ultimately reducing the Total Cost of Ownership (TCO) and expediting Time to Value (TTV). The collaborative efforts of VMware, Intel, and Dell Technologies empower customers to accelerate their AI initiatives, ultimately realizing the vision of AI Everywhere.

Solution Overview

VMware Private AI represents a strategic architectural framework designed to unlock the business benefits of AI adoption while simultaneously addressing organizations’ critical privacy and compliance requirements. This innovative approach encompasses a comprehensive set of components and specialized expertise to deliver AI services prioritizing data privacy, regulatory compliance, and operational control. Developed collaboratively with industry partners – Intel and Dell Technologies, VMware Private AI offers a tailored solution that empowers enterprises to maintain data privacy, leverage a mix of open-source and commercial AI tools, accelerate time-to-value, and ensure robust security and governance.

At its core, the VMware Private AI with Intel infrastructure stack is built upon VMware Cloud Foundation on Dell VxRail. It leverages acceleration capabilities such as the Intel AMX instructions, delivered as part of VxRail integrated systems with Intel 4^th Gen Xeon Scalable CPUs, providing an optimized infrastructure framework designed to facilitate the deployment of Private AI solutions within the enterprise. This framework empowers customers to effectively harness the advantages of AI technologies while maintaining control over the utilization and storage of their data for various GenAI use cases. Since this approach does not require GPU accelerators, it also does not require SR-IOV.

By leveraging VMware Private AI with Intel on Dell VxRail, organizations can benefit from integrating Intel’s AI software suite, Intel Xeon processors featuring embedded on-chip AI accelerator, and VMware Cloud Foundation on Dell VxRail. This collaborative solution empowers customers to develop and implement private AI models within their existing infrastructure, leading to reduced total cost of ownership and a focus on environmental sustainability. This partnership between VMware, Intel, and Dell Technologies enables the optimization of smaller, cost-effective, state-of-the-art AI models that are easier to maintain and update within shared virtual environments. Once batch AI jobs are completed, the resources can seamlessly return to the IT shared resource pool, ready to support ongoing inferencing tasks that operate continuously rather than in batches. This flexible approach supports many use cases, such as AI-assisted code generation, innovative customer service centers using natural language processing, and traditional machine learning/statistical analytics. All these applications can coexist on the same general-purpose servers running traditional and cloud-native applications, optimizing resource utilization and efficiency.

Figure 1: High-level architecture of VMware Private AI with Intel on VMware Cloud Foundation on Dell VxRail

Solution Components

Dell VxRail

VxRail, powered by Dell PowerEdge server platforms and VxRail HCI System Software, incorporates advanced technology to ensure the longevity and adaptability of your infrastructure while fostering deep integration within the VMware ecosystem. VxRail HCI System Software, a suite of integrated software elements that sits between infrastructure components such as vSAN and VMware Cloud Foundation, delivers a seamless and automated operational experience. By embracing Intel’s 4^th Gen Intel Xeon Scalable CPU with AMX, memory, and storage innovations, VxRail can support the latest technologies. Its versatile architecture spanning compute, memory, storage, network, and graphics options enable optimal performance across a diverse range of applications and workloads.

VxRail plays a crucial role in cost optimization by consolidating business-critical workloads onto a high-performing platform that excels in reliability, functionality, and performance. This consolidation enhances operational efficiency and enables IT teams to reallocate resources effectively through productivity-boosting features like streamlined deployments and automated patching and updates. Moreover, by minimizing the frequency and duration of service disruptions, VxRail is a protective barrier against revenue loss, ensuring a seamless user experience while enhancing data backup, protection, and recovery capabilities.
Intel 4^th Gen Intel Xeon Scalable Processors with AMX and Intel’s AI software suite

The Intel AMX is an on-chip hardware AI accelerator that empowers 4^th and 5^th Gen Intel Xeon Scalable processors to enhance deep learning (DL) training and inferencing workloads. By leveraging Intel AMX, these processors can seamlessly transition between optimizing general computing tasks and AI workloads. The 4th and 5^th Gen Intel Xeon Scalable processors offer developers the flexibility to leverage the Intel AMX instruction set for AI-specific functionalities while using the processor instruction set architecture (ISA) for non-AI tasks. Intel has further integrated the Intel oneAPI Deep Neural Network Library (oneDNN) and its oneAPI DL engine into widely used open-source AI tools such as TensorFlow, PyTorch, PaddlePaddle, and OpenVINO, enhancing the accessibility and efficiency of AI applications.

Intel’s AI software suite is equipped with a robust array of open-source tools and optional licensing components. It is designed to support developers to run full AI pipeline workflows from data preparation to fine-tuning to inference, accelerate building multi-node scaling, and deploying AI on enterprise IT infrastructure. Utilizing the open oneAPI framework, the suite allows for development that is agnostic to processors and hardware accelerators. This means developers can create applications once and deploy them across different architectures, avoiding the need to maintain multiple codebases or learn specialized programming languages. Additionally, Intel’s Transformer Extensions and PyTorch Extensions are tightly integrated with the widely-used Hugging Face open-source libraries, offering automated optimization techniques that streamline the process of model fine-tuning and compression for efficient inference. Developers looking to leverage these tools can download the AI Tools from AI Tools Selector.

VMware Cloud Foundation

VMware Cloud Foundation represents the next evolution of VMware’s hybrid cloud platform, building upon the industry-leading server virtualization technology of VMware vSphere. This advancement extends the core hypervisor by integrating software-defined storage, networking, and security features, providing users with flexible consumption options on-premises or in the public cloud. With the inclusion of integrated cloud management capabilities, the platform delivers a unified hybrid cloud solution that seamlessly spans private and public environments. It offers a consistent operational model leveraging familiar vSphere tools and processes, granting businesses the freedom to deploy applications across various environments without the complexity of application rewrites.

VMware Cloud Foundation on VxRail

VMware Cloud Foundation on VxRail combines Dell VxRail and VMware Cloud Foundation and delivers a simple and direct path to modern apps and the hybrid cloud with one complete, automated platform. Deep integration between VxRail and VMware Cloud Foundation combines operational transparency, automation, support, and serviceability capabilities for a turnkey hybrid cloud experience. Enterprises can deploy, host, and manage traditional VMs alongside cloud-native workloads across core, edge, and cloud environments. Streamlined operations help IT rapidly provision infrastructure to developers so they can create and deploy code faster to market and drive business innovation.

VxRail is the only jointly engineered HCI system with deep VMware Cloud Foundation integration. That deep integration delivers a unique experience with cluster deployment and management from a single, familiar user interface and automated end-to-end LCM management to ensure VxRail clusters remain in continuously validated states.

Your Browser is Out of Date

VMware Private AI with Intel on VMware Cloud Foundation on Dell VxRail