Tag: GeForce RTX 4090

  • Nvidia RTX Neural Texture Compression Cuts VRAM Usage by 96%

    Nvidia RTX Neural Texture Compression Cuts VRAM Usage by 96%

    Key Takeaways

    1. Nvidia’s new RTX toolkit beta features “Neural Rendering,” enhancing gaming visuals and performance through AI technologies.
    2. RTX Neural Materials allow for faster rendering of complex materials, making games capable of using more detailed assets.
    3. RTX Neural Texture Compression can reduce graphics memory usage by up to seven times, significantly lowering texture size while maintaining quality.
    4. The technology may impact frame rates, with an average FPS drop of 5.6% due to the computational demands of texture compression, especially at higher resolutions.
    5. Nvidia aims to create highly realistic human characters in games by combining 3D motion data with extensive photographic data of real individuals.

    Nvidia has launched a new beta version of its RTX toolkit, which includes a feature named “Neural Rendering,” hinting at what the future of gaming might look like. While DLSS has already been using artificial intelligence to upscale games to higher resolutions and add three extra frames per second, Nvidia’s forthcoming AI shaders promise to enhance this experience.

    AI Innovations in Gaming

    A key feature of these new AI advancements is RTX Neural Materials. This innovative technology leverages AI to compress intricate materials, enabling them to be rendered up to five times quicker, which opens the door for games to use much more intricate assets. Additionally, RTX Mega Geometry dynamically recalculates the geometry of objects in real time, paving the way for improved path tracing and more lifelike lighting effects.

    Enhanced Texture Management

    On the flip side, RTX Neural Texture Compression is aimed at compressing thousands of textures in less than a minute. This technology can decrease graphics memory usage by up to seven times when compared to traditionally compressed textures of the same quality.

    A video from Compusemble demonstrates RTX Neural Texture Compression and shows how AI-driven compression can significantly lower the memory usage of a 1440p model by as much as 96% compared to the reference texture. However, this comes at a cost, as the average frame rate dips by 5.6% due to the computing power needed for texture compression. The performance gap widens even more at higher resolutions; with 4K textures, average FPS on an Nvidia GeForce RTX 4090 nearly drops by half.

    Future of Gaming Characters

    Despite this, in games that struggle with the limited VRAM found in GPUs like the GeForce RTX 4060 (starting at $299 on Amazon), this texture compression may result in much better frame rates. Since Neural Texture Compression operates on tensor cores, it is expected to perform better on GeForce RTX 5000 series graphics cards compared to their older counterparts.

    Nvidia’s RTX kit also aims to create human characters in games that are so lifelike they are nearly indistinguishable from actual people. To achieve this realism, Nvidia merges raster data of a face from photographs or AI-generated images with 3D motion data, crafting the most authentic models possible. Nvidia states that building such a model requires thousands of images of a real person, captured from various angles and expressing a wide range of emotions for accurate reproduction later on.

    Source:
    Link

    Tags: , ,
  • New GeForce RTX 5090 Leak Shows 46% Boost in Performance

    New GeForce RTX 5090 Leak Shows 46% Boost in Performance

    Nvidia’s new gaming benchmarks indicate that the GeForce RTX 5090 might be about 30% quicker than the GeForce RTX 4090 regarding raw performance. This is notably seen in Cyberpunk 2077, where the Blackwell flagship needs DLSS to maintain playable frame rates when Path Tracing is turned on. Synthetic benchmarks for the graphics card have surfaced before its official launch through Videocardz.

    Performance Insights

    The findings imply that the GeForce RTX 5090 could be anywhere from 30% to 46% faster than the RTX 4090, which aligns with the expected boost in performance. Naturally, other factors like CPU and RAM play a role, but the benchmarks provide a good indication of how the graphics card performs with high-end processors from both Intel and AMD.

    Future of Gaming Graphics

    While these results might not apply to every title, the GeForce RTX 5090 will likely remain the fastest desktop graphics card available for quite some time. Competing products like AMD’s Radeon RX 9070 XT and its UDNA-based successors will struggle to surpass it, and no Intel Battlemage or Celestial SKU will be able to match its capabilities either.

    Videocardz’s analysis points to a promising future for Nvidia’s latest offering.

    Source:
    Link

    Tags: , ,
  • Razer Blade 18 Appears on Geekbench with Intel Arrow Lake HX and Nvidia GPU

    Razer Blade 18 Appears on Geekbench with Intel Arrow Lake HX and Nvidia GPU

    An earlier leak hinted at a new HP Omen laptop featuring cutting-edge hardware from Nvidia and Intel. Many other manufacturers are likely to follow suit, and we have now come across one of Razer’s highest-spec models. Yet, the Razer Blade 18 we’re looking at raises more questions than it clears up.

    Performance Specs

    On Geekbench, we can observe this new 18-inch model equipped with a Core Ultra 9 275HX, which is Intel’s top-of-the-line Arrow Lake-HX processor. It boasts 8 performance cores and 16 efficiency cores, with a maximum clock speed of 5.2 GHz for the performance cores. This particular laptop comes with a whopping 64 GB of RAM. So far, everything seems fine, but the graphics card tells a different story.

    GPU Confusion

    The Razer Blade 18 appears to be fitted with a full GeForce RTX 4090 graphics card. That’s right, the desktop variant, which features 128 Streaming Multiprocessors and a boost clock of 2.5 GHz. It doesn’t seem like Geekbench is mistakenly identifying the SKU, as it clearly states 24 GB VRAM underneath. Moreover, the upcoming mobile GeForce RTX 5090 is rumored to have only 16 GB of video memory, not 24 GB.

    Possible Future Developments

    Although there is a 24 GB variant of the GeForce RTX 5090 planned for laptops in the future, it won’t be available until 3 GB GDDR7 modules are widely used. It looks like someone has plugged an RTX 4090 into a Razer Blade 18 and took it for a test run.

    Could this indicate the launch of a new eGPU dock from Razer? Most likely. This theory is supported by the actual OpenCL score, which is 241,793—significantly lower than the usual score of 345,652, suggesting that some throttling is happening.

    Source: Link

    Image 1
    Tags: , ,
  • Stalker 2 PC Requirements: 32GB RAM, 160GB Storage, RTX 4070

    Stalker 2 PC Requirements: 32GB RAM, 160GB Storage, RTX 4070

    S.T.A.L.K.E.R 2: Heart of Chronobyl, known simply as Stalker 2, is set to be released on PC and Xbox Series X|S on November 20. GSC Game World, the developer behind the game, has announced the system requirements for PC users. As with many AAA titles this generation, the requirements are quite demanding. This is normal for a game built on Unreal Engine 5.

    Performance on Older Machines

    Although it’s technically possible to play Stalker 2 on older computers, it’s advisable to keep the resolution at 1080p. If you want to go above that, you will need 32 GB of RAM. Additionally, the game requires 160 GB of storage space, which could make it difficult to run on handheld devices like the Steam Deck.

    Graphics Settings and GPU Demands

    Fortunately, the game includes support for both DLSS and FSR, allowing players to adjust in-game graphics settings to better meet the requirements. It’s evident that the game is more demanding on the GPU rather than the CPU, even when played at lower resolutions.

    Nvidia has shared a blog post detailing how Stalker 2 performs with its GeForce RTX 4000 series graphics cards. At 4K resolution with the ‘Epic’ setting, when combined with an Intel Core i9-14900K and 64 GB of RAM, even the high-end GeForce RTX 4090 struggles to maintain a stable 60 FPS without DLSS. These benchmarks are by the manufacturer, so actual performance may be lower.

    Frame Rate Challenges

    The performance at 1440p and 1080p is similarly challenging, as DLSS 3 is practically essential for achieving playable frame rates on recommended setups. Nvidia has not clarified whether ray tracing is active in either case, which would certainly push an RTX 4090 to its limits.

    For laptops, things are slightly better, especially at 1080p. Even an entry-level GeForce RTX 4050 can run Stalker 2 at maximum settings. With some adjustments, you can reach a steady 60 FPS on older RTX 3000 series cards.

    Conclusion on Game Demands

    Stalker 2 is clearly among the most demanding games to be released this generation. Current hardware may struggle, but next-gen graphics cards like the GeForce RTX 5090 and 5080 may handle it better, especially when paired with a Ryzen 9 9950X3D.

    @stalker_thegame on X

    Image 1
    Image 1
    Image 1
    Image 1
    Image 1
    Image 1
    Tags: , ,