Nvidia has announced the integration of AI features into upcoming gaming laptops branded as RTX AI PC laptops. As revealed during Computex 2024, Asus and MSI will be the first devices featuring these enhancements.
Nvidia's RTX AI PC laptops will be equipped with GPUs up to the GeForce RTX 4070 and advanced energy-efficient system-on-chips. These devices will run on Windows 11, featuring enhanced AI functionalities.
Nvidia's RTX AI PC laptops will feature AI assistants to enhance user interaction and productivity. These assistants will utilize the GeForce RTX 4070 GPUs' capabilities for real-time support and advanced functions. Moreover, these laptops will also boast improved power management, catering to both gaming and professional requirements.
Complimentary AI Copilot Plus Update
Nvidia is also collaborating with Microsoft to integrate AI models into Windows 11. This partnership will offer developers API access to GPU-accelerated small language models for on-device AI functionality. Microsoft's Windows Copilot Runtime aims to facilitate AI feature integration in applications, utilizing either NPUs or Nvidia GPUs for acceleration.
The Copilot Plus AI features for these Windows 11 AI PCs will be available as a free update once finalized. This implies that initially, Microsoft's AI-enhanced Windows features may be exclusive to Qualcomm hardware, set for release on June 18th. Nvidia is currently soliciting more details regarding this update.
Nvidia emphasizes its GPUs' role in managing AI workloads in laptops, contrasting Microsoft's preference for Neural Processing Units (NPUs). According to Nvidia, GPUs are better suited for heavy AI processing compared to NPUs. To support this claim, Nvidia will introduce an RTX AI Toolkit in June, consisting of tools and SDKs for fine-tuning, optimizing, and deploying models. This toolkit aims to boost the performance of models like Meta's Llama 2 by reducing VRAM demand. According to Nvidia, the toolkit will enable “Windows developers to customize, optimize, and deploy AI models across RTX PCs and cloud. Via the end-to-end workflow , developers can customize open source models, reduce size by up to 3x, improve performance by up to 4x, and seamlessly deploy within their applications to 100M RTX AI PCs“.
Performance vs. Efficiency
While NPUs deliver about 40 TOPS (trillions of operations per second), Nvidia's GPUs surpass 1,000 TOPS. This stark difference necessitates a balance between power efficiency and computational capabilities for developers. NPUs are ideal for laptops due to their energy efficiency and suitability for smaller models, whereas GPUs handle larger models and provide top-tier performance, especially in desktops where battery life is not an issue.
Developer Resources and SDKs
Nvidia's forthcoming RTX AI Toolkit will offer developers essential tools and SDKs for AI model optimization and deployment. Designed to enhance models' performance and efficiency, particularly Meta's Llama 2, this toolkit is key to Nvidia's strategy in AI computing.
