NVIDIA has made significant strides in gaming technology with the unveiling of its latest Deep Learning Super Sampling (DLSS) 4 upscaling technology. This new enhancement promises an impressive eightfold increase in performance on RTX 50 GPUs, enabled by the groundbreaking Multi-Frame Generation feature.
Unveiling DLSS 4: A Game Changer for RTX 50 Series
Press Release: The announcement came at CES 2025, where NVIDIA showcased DLSS 4. This revolutionary technology will support Multi Frame Generation on the GeForce RTX 50 Series graphics cards and laptops, with an anticipated 75 games and applications expected to adopt this functionality upon release.
DLSS Multi Frame Generation is designed to produce up to three additional frames for every rendered frame, dramatically increasing frame rates—by as much as 8X compared to traditional rendering. This translates into breathtaking 4K gaming experiences at an astonishing 240 FPS with full ray tracing, especially evident on GeForce RTX 5090 graphics cards.
When applied to demanding titles such as Cyberpunk 2077, DLSS 4 with Multi Frame Generation exhibits a performance boost of over 8X in comparison to standard rendering techniques. Additionally, it achieves a remarkable reduction in PC latency, enhancing responsiveness and overall image quality for gamers.
AI-Powered Innovations: The New Transformer Models
One of the most exciting developments in DLSS 4 is the introduction of new AI models. Since the launch of DLSS 2.0 in 2020, the upgrade is the most substantial, incorporating pioneering technologies. By harnessing the power of ‘transformers’, similar to advanced AI models like ChatGPT, Flux, and Gemini, NVIDIA enhances several key technologies: DLSS Ray Reconstruction, DLSS Super Resolution, and DLAA. These models bring improved image quality marked by greater temporal stability, reduced ghosting, and enhanced detail during motion.
Beyond the RTX 50 Series availability, NVIDIA’s application users will gain access to these state-of-the-art enhancements, meaning a significant upgrade across the board.
Extensive Compatibility and Performance Boost
With 75 titles receiving upgraded support for Multi Frame Generation on the GeForce RTX 50 Series GPUs, existing games will also benefit from enhanced performance. The technology extends its reach to GeForce RTX 40 Series GPUs, optimizing gaming performance while lowering VRAM consumption. Furthermore, all GeForce RTX GPUs will have access to the revitalized DLSS transformer model, enhancing games that incorporate Ray Reconstruction, Super Resolution, and DLAA technologies.
The secret behind DLSS Multi Frame Generation’s performance lies in its innovative approach. The AI model that powered DLSS 3 Frame Generation utilizes game data such as motion vectors and depth via the Optical Flow Accelerator. However, generating multiple frames uniformly posed challenges that could throttle GPU responsiveness. To tackle these issues, DLSS 4 employs advancements from both hardware and software.
DLSS 4’s new frame-generation AI model is 40% faster and consumes 30% less VRAM, requiring a single execution per rendered frame. For instance, in Warhammer 40,000: Darktide, the model showcased a 10% increase in frame rate, simultaneously reducing memory usage by 400MB in 4K resolution at maximum settings.
Despite these efficiencies, the GPU is tasked with running five AI models for Super Resolution, Ray Reconstruction, and Multi Frame Generation within several milliseconds to prevent any performance bottlenecks. GeForce RTX 50 Series GPUs meet this demand with 5th Generation Tensor Cores, capable of delivering up to 2.5X AI processing power.
Precision in Frame Delivery
Once the new frames are created, they are orchestrated to ensure seamless gameplay. Unlike DLSS 3 Frame Generation, which relied on CPU pacing and resulted in potential consistency issues, DLSS 4 integrates hardware Flip Metering into the display engine for better frame timing management. The enhanced Blackwell display engine has doubled its pixel processing capacity to accommodate higher resolutions and refresh rates.
NVIDIA’s innovative blend of hardware and software technologies permits DLSS 4 to produce outstanding image quality while maintaining smooth gameplay with minimal latency. When combined with Multi Frame Generation, Ray Reconstruction, and Super Resolution, frame rates can soar to levels previously thought unattainable compared to conventional rendering methods. Transitioning from Frame Generation to Multi Frame Generation can lead to a notable performance increase of up to 1.7X.
Transformer Models: Elevating Image Quality for Gamers
DLSS 4’s transformational architecture marks a significant shift in image quality enhancement techniques. The latest Ray Reconstruction, Super Resolution, and DLAA employ the industry’s first transformer-based model for real-time application.
Previously reliant on Convolutional Neural Networks (CNNs) for generating new pixels, the DLSS technology has maxed out its potential. Now, the new vision transformer model allows for complex self-attention mechanisms, evaluating every pixel’s importance across frames. This leap in design doubles the parameters for more comprehensive scene analysis, promoting stability, reduced ghosting, and finer details during motion.
In particularly complex lighting scenarios incurred during ray-traced graphics, the updated transformer model shows remarkable improvements. For example, in Alan Wake 2, the stability of detailed elements, such as chain link fences and fan blades, is greatly enhanced, significantly boosting the immersive player experience.
The beta for the Super Resolution transformer model offers insights into positive results, where feedback from early users will guide improvements before the full release. This feature is poised to enhance detail and clarity in motion significantly.
Availability of Multi Frame Generation and Backward Compatibility
NVIDIA has also committed to player accessibility. With the launch of the GeForce RTX 50 Series graphics cards, users will witness an immediate capacity for performance enhancement through DLSS Multi Frame Generation in 75 supported games and applications. In an additional benefit, over 50 titles will receive access to the new transformer-based technologies.
Games such as Alan Wake 2, Cyberpunk 2077, Indiana Jones and the Great Circle, and Star Wars Outlaws will launch with native in-game support for DLSS Multi Frame Generation alongside the release of GeForce RTX 50 Series GPUs. Additional titles, including Black Myth: Wukong, NARAKA: BLADEPOINT, Marvel Rivals, and Microsoft Flight Simulator 2024, are set to follow suit.
For titles yet to adopt the latest DLSS features, NVIDIA’s newly added DLSS Override feature in the app will help integrate support. After installing the latest GeForce Game Driver and NVIDIA app update, users can access DLSS Override options, providing extensive customization for individual game settings.
Key DLSS Override Functions Include:
- Frame Generation Override: Enables Multi Frame Generation when activated in-game.
- Model Presets Override: Allows for the utilization of the latest Frame Generation model across GeForce RTX 50 and 40 Series cards.
- Super Resolution Override: Configures internal rendering resolution settings for optimal performance.
As the NVIDIA app continues to evolve, support for DLSS Override will be expanded as performance tests on more titles conclude successfully.
Reflecting on NVIDIA DLSS 4: An Upgrade for All
Since its inception in 2018, NVIDIA DLSS has dramatically transformed gaming experiences through scalable AI technologies. The debut of DLSS 4, coupled with the GeForce RTX 50 Series GPUs, promises to elevate gameplay, offering gamers superior performance, enhanced visuals, and a deeper immersion into the worlds they explore.
Related Articles:
Crucial Enhances Memory & Storage Lineup with Premium PCIe Gen5 SSDs, DDR5, and CUDIMM Memory Solutions
21:25
TSMC Arizona Begins Production of AMD Ryzen 9000 “Granite Ridge” CPUs Amid High Demand
20:14
Exploring ASUS’s Next-Gen ROG Astral, Strix, TUF Gaming, and PRIME Series for NVIDIA RTX 50 Graphics Cards
19:10
Leave a Reply