AMD Showcases Cutting-Edge AI Solutions Across Data Centers and PCs

In a tech-packed event labeled “Advancing AI,” AMD (NASDAQ: AMD) gathered industry giants such as Microsoft, Meta, Oracle, Dell Technologies, HPE, Lenovo, Supermicro, Arista, Broadcom, and Cisco to spotlight their collaboration in delivering advanced AI solutions. This comprehensive showcase spans from cloud services and enterprise solutions to PCs.

Advancing Data Center AI

AMD revealed a range of new products during the event, with a primary focus on its Instinct MI300 Series data center AI accelerators. The company showcased a robust collaboration with industry leaders to integrate these accelerators into various solutions.

Microsoft highlighted its deployment of AMD Instinct MI300X accelerators to power the new Azure ND MI300x v5 Virtual Machine (VM) series, specifically optimized for AI workloads.

Meta shared its utilization of AMD Instinct MI300X accelerators in its data centers, combined with ROCm 6, to empower AI inferencing workloads.

Oracle announced plans to offer OCI bare metal compute solutions featuring AMD Instinct MI300X accelerators and integrate these accelerators into their upcoming generative AI service.

Major data center infrastructure providers, including Dell, HPE, Lenovo, and Supermicro, revealed plans to integrate AMD Instinct MI300 accelerators across their product portfolios. Additionally, several other tech giants, such as Asus, Gigabyte, Ingrasys, Inventec, QCT, Wistron, and Wiwynn, expressed intentions to offer solutions powered by AMD Instinct MI300 accelerators.

Specialized AI cloud providers, including Aligned, Arkon Energy, Cirrascale, Crusoe, Denvr Dataworks, and Tensorwaves, are gearing up to provide offerings that will extend access to AMD Instinct MI300X GPUs for developers and AI startups.

Open, Proven, and Ready AI Software Platform

AMD showcased the latest version of its open-source software stack for AMD Instinct GPUs, ROCm 6. This optimized software stack caters specifically to generative AI, focusing on large language models. With new data types, advanced graph and kernel optimizations, optimized libraries, and cutting-edge attention algorithms, ROCm 6, in conjunction with MI300X, delivers an impressive ~8x performance increase for overall latency in text generation on Llama 2 compared to ROCm 5 running on the MI250.

Collaborations with AI startups like Databricks, Essential AI, and Lamini were highlighted, showcasing how they leverage AMD Instinct MI300X accelerators and the ROCm 6 software stack to create differentiated AI solutions for enterprise customers. OpenAI is also adding support for AMD Instinct accelerators to Triton 3.0, providing out-of-the-box support for AMD accelerators for developers.

Continued Leadership in Advancing AI in PCs

With millions of AI PCs already shipped, AMD unveiled new leadership mobile processors—the AMD Ryzen 8040 Series processors. These processors offer enhanced AI compute capabilities. Additionally, AMD launched Ryzen AI 1.0 Software, a software stack enabling developers to easily deploy apps that use pretrained models to add AI capabilities to Windows applications.

AMD revealed that the upcoming next-gen “Strix Point” CPUs, planned for launch in 2024, will include the XDNA 2 architecture, promising more than a 3x increase in AI compute performance compared to the prior generation. Microsoft also joined the discussion, highlighting their close collaboration with AMD on future AI experiences for Windows PCs.