NVIDIA's latest frame generation technology, originally exclusive to newer GPU models, is now running on older RTX 30 Series graphics cards thanks to a community-developed mod called NVSmooth30. The experimental compatibility mod bypasses hardware restrictions to enable NVIDIA Smooth Motion on Ampere-architecture GPUs, allowing gamers with older hardware to access AI-powered frame doubling without purchasing new graphics cards.

What Is NVIDIA Smooth Motion and How Does It Work?

NVIDIA Smooth Motion is a driver-level AI frame generation solution that was introduced with the RTX 50 Series and later expanded to RTX 40 Series GPUs. Unlike DLSS Frame Generation, which requires game developers to integrate NVIDIA's software development kit, Smooth Motion works at the driver level, meaning it can generate additional frames in games that lack native frame generation support.

The technology functions similarly to AMD's Fluid Motion Frames 2.1, generating additional frames between traditionally rendered ones to improve perceived smoothness and increase displayed frame rates. According to early user reports, an RTX 3080 achieved approximately 130 displayed frames per second, up from 65 frames per second without frame generation in Kingdom Come: Deliverance II.

How Does NVSmooth30 Enable Frame Generation on Older GPUs?

NVSmooth30 works by modifying NVIDIA's NvPresent64.dll runtime behavior, bypassing the GPU architecture and capability checks that normally prevent Smooth Motion from initializing on Ampere hardware. The mod intercepts the CUDA module-loading process, retargeting compatible Ada Lovelace GPU code to Ampere architecture, allowing older cards to execute the frame generation algorithms.

Developer Adeline, who created the mod, notes that the implementation does not redistribute NVIDIA's proprietary binaries. Instead, it uses the NvPresent64.dll version already installed with the user's graphics driver. The developer initially demonstrated the project on an RTX 3050 Laptop GPU, but other users have reported success with considerably more powerful Ampere GPUs, including the RTX 3070 and RTX 3080.

Steps to Install and Use NVSmooth30

• Download the mod: Obtain NVSmooth30 from GitHub, where the developer has published the source code and installation instructions for the community.
• Place the proxy file: Copy the version.dll proxy file alongside your game's executable file in the game directory.
• Enable through NVIDIA Profile Inspector: Use NVIDIA Profile Inspector to enable Smooth Motion for the specific game title you want to use frame generation with.
• Reactivate after driver updates: The setting may need to be reactivated following graphics driver updates, as new drivers can change how the system initializes.

What Are the Limitations and Risks?

NVSmooth30 remains an alpha-stage project, and compatibility across different games, drivers, and RTX 30 Series GPUs is not guaranteed. NVIDIA driver updates could introduce changes that break the mod's functionality entirely. Users are strongly advised against trying it in competitive or anti-cheat-protected games, as the mod's presence could trigger anti-cheat systems. Potential issues include crashes, corrupted frames, and black screens.

The developer is also working on an experimental Direct3D 11-to-Direct3D 12 presentation bridge, which could potentially expand the mod's usefulness across a much larger library of PC titles that use older graphics APIs. However, this feature remains in early development and is not yet widely available.

Why Does This Matter for NVIDIA's Broader AI Strategy?

NVSmooth30 represents another interesting development for Ampere GPU owners who want to experiment with NVIDIA's newer rendering technologies without upgrading their graphics cards. The emergence of mods enabling DLSS Frame Generation and Multi Frame Generation on older GeForce RTX GPUs demonstrates the strong demand for frame generation technology across the gaming community.

This trend also highlights NVIDIA's dominant position in AI hardware and software. CUDA, NVIDIA's parallel computing platform, has been the industry standard for AI development since its introduction in 2007. The company spent 15 years and billions of dollars building CUDA and its ecosystem, creating a moat that competitors have struggled to overcome. The fact that developers are willing to create complex mods to enable NVIDIA's proprietary technologies on older hardware underscores how deeply embedded NVIDIA's software has become in the AI and gaming development communities.

As of September 2026, NVIDIA's Blackwell architecture represents the current generation of AI training chips, offering roughly twice the performance of the previous H100 generation while consuming 1,000 watts of power. However, the continued interest in enabling advanced features on older RTX 30 Series cards suggests that many users and developers still rely on Ampere-generation hardware for their workloads.