Apex Compute ANR Claims Faster Performance Than Nvidia DLSS
Apex Compute has launched an alternative to Nvidia’s DLSS that achieves a 1.01ms latency per generated frame at 2X, TechPowerUp reported, undercutting Nvidia’s 1.26ms benchmark while running on AMD cards and a wide array of other hardware.
Nvidia’s DLSS suite of super-resolution, frame generation, and ray reconstruction has served as the upscaling gold standard for years. Credible alternatives include AMD’s FSR and Intel’s XeSS, alongside third-party options like Lossless Scaling. Apex Compute, known as the team behind a range of low-cost, efficient AI accelerators, has now entered the market with its Apex Neural Rendering (ANR) framework.
The ANR system incorporates super-resolution upscaling, ray reconstruction denoising, and frame generation. Currently, Apex has only released a demonstration of its frame-generation capabilities, while the ray reconstruction and upscaling capabilities are still undergoing training and are scheduled to arrive soon. The demonstration showed a cleaner implementation with fewer artifacts and visual hiccups, particularly around object edges.
Introducing Apex Neural Rendering (ANR), our cross-platform Super Resolution, Ray Reconstruction, and Frame Generation technology, delivering up to 2.4× faster performance and higher quality rendering than NVIDIA DLSS. ANR rebuilds neural rendering around a new architecture that… pic.twitter.com/DcSkwrouKS
Aside from beating Nvidia’s latency, the software exhibits greater speed relative to AMD’s FSR 3.1 and FSR 4, most notably at elevated frame generation levels. It also demands far less memory, requiring 214MB to run when rendering a game at 4K, whereas DLSS frame generation demands 562MB.
Whereas Nvidia restricts multi-frame generation officially to its newest graphics cards, ANR is reported to function on the GTX 16-series alongside RTX 20-series and more recent hardware. Support extends to every AMD GPU from the RX 5000 series onward, Intel Arc A- and B-series cards, the Steam Deck, and Apple silicon.
The software is compatible with Windows 10 and 11, Linux, and SteamOS with Proton, which could boost performance on homebrews and new Steam Machines. Users can test the software by downloading the ANR Beta directly from its GitHub repository.
For years, the upscaling benchmark has been set by Nvidia’s DLSS collection of ray reconstruction, frame generation, and super-resolution features.