Unlocking the Power of Generative AI on NVIDIA RTX AI PCs with Windows ML Integration
Generative AI is revolutionizing PC software, transforming everyday applications into breakthrough experiences. From intelligent writing assistants to creative tools, digital humans, and smart agents, NVIDIA RTX AI PCs are at the forefront of this transformation. With cutting-edge technologies like TensorRT for RTX, NVIDIA NIM microservices, AI Blueprints, and Project G-Assist plug-ins, developers and enthusiasts now have powerful tools to experiment with, build, and deploy AI-driven workflows.
Accelerating AI Inference with TensorRT for RTX
One of the most significant advancements announced at Microsoft Build is the integration of TensorRT for RTX with Windows ML. This collaboration enables app developers to harness state-of-the-art AI performance while maintaining broad hardware compatibility.
TensorRT for RTX has been reimagined specifically for RTX AI PCs, combining industry-leading performance with just-in-time, on-device engine building. This innovation streamlines AI deployment by reducing the library’s file size by 8x, making it easier to distribute across more than 100 million RTX AI PCs.
For developers, the benefits are substantial. Compared to DirectML, TensorRT delivers over 50% faster performance for AI workloads on PCs. Additionally, Windows ML automatically selects the optimal hardware—whether GPU, CPU, or NPU—to run each AI feature, eliminating the need for developers to package execution providers with their apps. This ensures users receive the latest TensorRT optimizations as soon as they’re available.
Developers can access TensorRT for RTX through the Windows ML preview today, with a standalone SDK set to launch on NVIDIA Developer in June. To dive deeper, explore the TensorRT for RTX launch blog or Microsoft’s Windows ML blog.
Expanding the AI Ecosystem on Windows 11 PCs
NVIDIA offers a comprehensive suite of SDKs to help developers integrate advanced AI features into their applications. These include:
- CUDA and TensorRT for GPU acceleration.
- DLSS and OptiX for enhanced 3D graphics.
- RTX Video and Maxine for multimedia enhancements.
- Riva and ACE for generative AI capabilities.
Top applications are already leveraging these tools to deliver groundbreaking features. For example:
- LM Studio released an update that boosts performance by over 30% using the latest CUDA version.
- Topaz Labs introduced a generative AI video model powered by CUDA to enhance video quality.
- Chaos Enscape and Autodesk VRED added DLSS 4 for faster rendering and superior image quality.
- Bilibili integrated NVIDIA Broadcast features like Virtual Background to elevate livestreaming experiences.
These updates highlight how NVIDIA’s ecosystem empowers developers to push the boundaries of what’s possible on RTX-powered machines.
Simplifying AI Development with NIM Microservices and AI Blueprints
Getting started with AI development can be challenging, especially given the vast array of models and dependencies. NVIDIA addresses this with NIM microservices, prepackaged AI models optimized for full performance on RTX GPUs. These microservices eliminate the complexity of quantization, dependency management, and deployment, enabling seamless operation across PCs and the cloud.
Key highlights include:
- The release of the FLUX.1-schnell NIM microservice, an image generation model from Black Forest Labs designed for rapid results.
- Updates to the FLUX.1-dev NIM microservice, adding compatibility for GeForce RTX 50 and 40 Series GPUs.
- Performance improvements on NVIDIA Blackwell GPUs, where NIM microservices run over twice as fast thanks to FP4 and RTX optimizations.
To further accelerate development, NVIDIA provides AI Blueprints, open-source sample workflows that demonstrate how to use NIM microservices effectively. Last month, NVIDIA released the AI Blueprint for 3D-guided generative AI, empowering developers to control composition and camera angles using 3D scenes as references.
Empowering Enthusiasts with Project G-Assist Plug-Ins
For those eager to experiment with AI without extensive coding knowledge, Project G-Assist offers a no-code solution. Integrated into the NVIDIA app, G-Assist allows users to control their GeForce RTX systems using natural language commands. It also serves as a platform for creating custom plug-ins to automate tasks and enhance workflows.
The Project G-Assist Plug-in Builder simplifies development with natural language inputs, enabling creators to design lightweight JSON-based plug-ins with minimal effort. Recent additions to the community-driven repository include:
- Gemini: Enhanced with real-time web search capabilities powered by Google’s large language model.
- IFTTT: Automates IoT routines, such as adjusting smart home devices or receiving gaming news alerts.
- Discord: Enables hands-free sharing of game highlights and messages to Discord servers.
- Spotify and Twitch: Adds functionality for music control and livestream status checks.
SignalRGB is even developing a G-Assist plug-in for unified lighting control across multiple manufacturers, further expanding the possibilities for AI-driven customization.
Join the AI Revolution on RTX AI PCs
NVIDIA’s innovations are democratizing access to AI, empowering both developers and enthusiasts to unlock the potential of RTX AI PCs. Whether you’re building advanced workflows with TensorRT for RTX, experimenting with NIM microservices, or creating custom plug-ins with Project G-Assist, the tools are here to bring your ideas to life.
Stay connected with the RTX AI Garage blog series to discover community-driven AI projects, tutorials, and inspiration. Join the NVIDIA Developer Discord channel to collaborate, share creations, and gain support from a growing community of innovators.
As generative AI continues to evolve, NVIDIA remains committed to driving progress and enabling transformative experiences on Windows 11 PCs. Dive in today and be part of the future of AI-powered computing.
