AMD Radeon DLSS 5 Modding Breakthrough Delivers 74% Speed Boost in 24 Hours

A community-driven project is bringing Nvidia’s latest neural rendering technology to AMD Radeon GPUs, marking a significant shift in how graphics acceleration might work across different hardware platforms. The development has moved remarkably fast, with performance improvements accumulating at an unprecedented pace and implications that matter for anyone shopping for graphics cards right now.

What Happened with DLSS 5 on Radeon GPUs

Nvidia launched DLSS 5 with neural rendering capabilities earlier this month, exclusively for its RTX 50-series graphics cards. The technology uses AI to upscale images and enhance rendering quality, but it comes with a substantial performance cost. Early results showed significant frame rate drops when the feature was enabled.

Rather than waiting for official support, developers started working on an independent reimplementation. The GitHub project DLSS-NR-on-AMD is designed to port this neural rendering technology to AMD Radeon RX 9000 and RX 7000 series GPUs through a ground-up rebuild of the neural network runtime. The effort focuses on DirectX 12 games using FSR, with Cyberpunk 2077 and GTA V Enhanced among the titles already supported.

Performance gains came remarkably quickly. An Alpha 0.3.3 release delivered a 23% speed improvement. Within hours, Alpha 0.4.0 followed with an additional 42% boost over the previous version, combining to nearly 74% total performance improvement in just 24 hours of development work.

Performance Gains and Real World Numbers

Initial attempts showed the technology was barely usable. A Radeon RX 9070 XT running Cyberpunk 2077 at high settings achieved over 80 frames per second without neural rendering. Enabling the early mod reduced that to approximately 68 frames per second, making gameplay choppy and unplayable for most users.

The latest Alpha 0.4.0 build shows far more promise. At 3440 x 1440 resolution, the same setup now delivers around 50 frames per second with neural rendering enabled. While that remains below the native performance without the feature, it represents dramatic progress toward viability. The momentum suggests further refinements are coming, with developers already planning major updates expected to yield additional performance benefits.

Installation Gets Easier

Early versions required manual file replacement for each game, creating friction for average users. The new AMDNR launcher automates this process significantly. Users can now install and configure the mod with minimal technical knowledge. The launcher supports different runtimes and integrates with OptiScaler, providing additional customization options without requiring command-line work or complex setup procedures.

What This Means for Radeon GPU Shoppers

For consumers evaluating graphics cards right now, this development opens new possibilities. AMD Radeon GPUs could potentially access features previously exclusive to Nvidia hardware, even if through community solutions rather than official support.

The project remains in alpha, meaning performance and compatibility will vary depending on specific games and GPU models. Some titles may see excellent results while others deliver limited benefits. Stability could improve as development continues, but users installing alpha software should expect occasional issues.

The practical takeaway: if you own a Radeon RX 9000 or RX 7000 series GPU, you may soon have access to neural rendering enhancements. If you are shopping now, this suggests AMD’s newer graphics cards will support more cutting edge features than previously expected, potentially improving their long term value proposition. This matters particularly for gaming monitor buyers pairing new GPUs with high refresh displays.

The Optimization Road Ahead

Current performance still falls short of being practical for many use cases. Native rendering without neural enhancements remains faster on Radeon hardware, meaning the feature is not yet ready for everyday gaming. However, the rapid pace of improvements indicates this could change within weeks rather than months.

Developers have confirmed they are working on additional major updates focused specifically on Radeon optimization. Each successive build has shown substantial gains, suggesting the trajectory could continue upward. If optimization reaches a point where neural rendering adds meaningful quality with acceptable frame rate penalties, adoption could accelerate quickly.

Broader Implications for GPU Technology

This effort demonstrates how community innovation can push GPU capabilities forward independent of manufacturer timelines. Nvidia controls DLSS 5 officially, but determined developers are making it work on competing hardware through clever engineering and optimization work.

For the graphics card market, this represents competitive pressure in a new form. Features that seemed locked to specific platforms can potentially become available across brands through third party solutions. GPU buyers now have additional reasons to evaluate the full ecosystem around their hardware choices rather than focusing solely on official feature lists.

The project’s success also highlights the value of open development platforms and enthusiast communities. GitHub hosting enables rapid iteration and collaborative improvement that benefits all users, regardless of their hardware manufacturer preferences.

About the author