Dell Reveals AI Platform With AMD

News Desk -

Share

Dell and AMD are joining forces to enhance the Dell Generative AI Solutions range, furthering their mission to accelerate complex workloads and provide businesses with more options for their unique GenAI endeavors. This latest innovation underscores the critical role of open ecosystems and a diverse range of silicon in delivering easy-to-use, reliable, and customized solutions for integrating AI into various data contexts.

Generative AI (GenAI) is regarded as this decade’s most significant innovation catalyst, with 78% of IT leaders expressing enthusiasm about its potential impact on their organizations. Many view GenAI as a tool for increasing productivity, streamlining operations, and cutting costs. Embracing this technology is vital for companies to stay competitive in the evolving digital landscape.

The exponential growth of data and its implications are at the forefront of today’s AI discussions. AI represents the newest innovation wave, and providing organizations with the necessary computing power to fuel this wave is essential. In partnership with AMD, Dell is now able to offer more options than ever for powering AI and GenAI workloads.

Arthur Lewis, president of Dell Technologies’ Infrastructure Solutions Group, remarks, “We’re at a technological turning point. Advanced IT systems enable AI to revolutionize business and industry operations. This new era of computing, unlike anything previously seen, is made possible through our long-standing partnership with AMD, empowering our customers with a broader range of trusted, custom solutions for their AI and GenAI needs.”

Dell’s fastest-developing solution to date, the Dell PowerEdge XE9680, will soon support the AMD Instinct™ MI300X accelerator, enhancing AI performance choices for customers. The PowerEdge XE9680 equipped with the MI300X accelerator offers robust performance for enterprises aiming to leverage their data, differentiating their operations with custom large language models (LLMs). With eight MI300X GPUs, each with 192GB of 5.3 TB/s High Bandwidth Memory (HBM3) and a total coherent HBM3 capacity of 1.5 TB per server, plus over 21 petaflops of FP16 performance, this server is set to make GenAI more accessible for businesses striving to train larger models, reduce data center space, lower total cost of ownership, and gain a competitive advantage.

Additionally, Dell is introducing the next phase of Dell Generative AI Solutions, simplifying the deployment of reliable GenAI with the new Dell Validated Design for Generative AI with AMD. This upcoming offering, powered by AMD ROCm™-enabled AI frameworks and launching next year, expands the Dell Generative AI Solutions ecosystem to include open-source LLMs. AMD ROCm™, an open-source stack of drivers, development toolkits, and APIs for AMD Instinct accelerators, forms the backbone of this initiative. The Dell Validated Designs for Generative AI simplify the process for customers to build their own tailored GenAI platforms, removing the complexities of integration, performance, and sizing.

Dell’s latest eBook and infographic highlight how the PowerEdge XE9680 with MI300X, optimized for ROCm, seamlessly supports popular AI and ML frameworks like PyTorch, TensorFlow, ONNX-RT, JAX, and Triton. This enables straightforward integration of LLMs based on these frameworks.

The PowerEdge XE9680 stands out for its high performance, robust security, and ease of use. Dell Open Manage Enterprise software streamlines GenAI infrastructure deployment, complemented by CloudIQ software for intelligent automation. The server is secured with integrated cyber recovery and a comprehensive Zero Trust approach.

Featuring eight AMD Instinct MI300X accelerators, the PowerEdge XE9680 leverages the AMD Instinct Platform for nearly linear scalability and minimal latency in distributed GenAI training and inferencing. This is achieved through Global Memory Interconnect (xGMI) across a cluster of PowerEdge servers, linking MI300X GPUs via an Ethernet-based AI fabric using a Dell validated GPU fabric with Dell PowerSwitch Z9664F-ON.

As part of the Ultra Ethernet Consortium (UEC), Dell actively promotes standardized networking. Dell Technologies is in a unique position to fully harness the benefits of an open approach across its entire GenAI solution – including compute, fabric, and storage. The Dell PowerEdge XE9680 with AMD Instinct™ MI300X accelerators and the Dell Validated Design for Generative AI with AMD are set to be available in the first half of 2024.


Leave a reply