Intel has announced comprehensive support for Microsoft’s Phi-3 AI models across its entire range of CPUs and GPUs. This collaboration aims to make advanced AI capabilities accessible on devices with limited processing power, marking a significant step towards democratizing AI technology.
Hardware Optimization for AI Efficiency
Intel has tailored its hardware, including Gaudi AI accelerators, Xeon and Core Ultra CPUs, and Arc GPUs, to work seamlessly with Microsoft’s Phi-3 models. This optimization is designed to enhance both performance and efficiency. Notably, Intel’s 6th Gen Xeon CPUs have shown twice the performance with Phi-3 models compared to their predecessors. The Core Ultra CPU platform also demonstrates significant advancements in edge AI processes, particularly in personalization and responsiveness.
Benchmark tests have highlighted the impressive performance of the Phi-3 mini model on Intel’s Core Ultra 7 165H, paired with an Arc A770 GPU. These results indicate substantial improvements in both performance and efficiency, paving the way for smarter home devices and more intuitive personal gadgets. The integration of AI into everyday devices is becoming increasingly feasible, making advanced AI more accessible to a broader audience.
Diverse AI Models for Varied Applications
Microsoft’s Phi-3 family includes mini, small, and medium models, each designed to provide AI capabilities on a smaller and more efficient scale. These models are particularly suitable for low-power devices with limited processing power. Intel has showcased the performance of its products across all these models, demonstrating the versatility and efficiency of its hardware.
Pallavi Mahajan, Intel Corporate VP and GM of Data Center and AI Software, emphasized the importance of Intel’s collaboration with Microsoft in bringing AI capabilities to a wider range of devices. This partnership ensures that Intel hardware supports several new Phi-3 models, making advanced AI more accessible and efficient.
Intel has benchmarked the Phi-3 medium 4k and 128k variants on its 6th Gen Xeon CPUs (Granite Rapids), achieving twice the performance compared to the 5th Gen Xeon (Emerald Rapids) counterparts. These performance milestones highlight the significant advancements in AI computing capabilities, leveraging the power of modern-day processors.
Last Updated on November 7, 2024 8:04 pm CET
