At the Microsoft Build event, AMD showcased its latest advancements in AI computing through several new announcements. Microsoft is utilizing AMD technology, including the Instinct MI300X GPUs and ROCm software, to power AI workloads. The Instinct MI300X GPUs are being deployed on Microsoft Azure servers to support OpenAI services such as ChatGPT 3.5 and ChatGPT 4, as well as Microsoft’s Copilot AI assistant which integrates information from ChatGPT. Additionally, the fourth generation EPYC CPUs are widely used in Azure.
To optimize the performance of these GPUs, Microsoft is leveraging AMD’s open source ROCm software to achieve leading price-to-performance ratios, a significant win for AMD. The AMD Alveo MA35D media accelerator is also being utilized to power Microsoft’s live streaming workloads across various applications like Teams and SharePoint.
One of the initial major customers to utilize Microsoft’s new AMD Instinct-powered Azure servers is the artificial intelligence company Hugging Face based in New York. Developers can also deploy AI Inference software locally using AMD software, provided they have a Ryzen AI processor currently available in the laptop market. The new Ryzen AI processors feature XDNA architecture to enhance AI workloads.
Overall, AMD is making significant strides in the AI space with its partnership with Microsoft growing stronger. The collaboration between the two companies is driving innovation and advancements in AI computing technology.
Article Source
https://www.kitguru.net/channel/generaltech/matthew-wilson/microsoft-teams-up-with-amd-to-power-azure-ai-services/