Developers at RPCS3, the open-source PlayStation 3 emulator, have criticized Nvidia’s DLSS 5 technology, calling it an “AI-slop generator.” A number of modders managed to port DLSS 5 to various games last week after the latest neural-rendering tech leaked before the official launch. A few people also went ahead and apparently tested the leaked DLSS 5 DLL on the RPCS3 renderer. According to a post by the official RPCS3 account on X, “Whilst the gaming industry pushes more upscalers and frame generation to hallucinate games and hide their lack of optimisation, we do not ship any of that slop.”
The post has drawn criticism from certain users, with one calling RPCS3’s response “possibly the dumbest” they could have expected. They argued that DLSS and FSR frame generation can complement game-engine optimizations rather than simply mask poor performance, while Nvidia’s Reflex can help mitigate the latency penalty associated with frame generation.
DLSS 5 on RPCS3?A few people are testing a leaked DLSS 5 DLL on our renderer. We have seen the slop it generates.Whilst the gaming industry pushes more upscalers and frame generation to hallucinate games and hide their lack of optimisation, we do not ship any of that slop. pic.twitter.com/S2GciNbCTKSeptember 1, 2026
In its response, RPCS3 said that emulators have never-ending goals, and they do not simply target high-end hardware to achieve playable status. The developers also added that modern temporal upscalers are difficult to implement because PS3-era games do not expose the motion-vector data they require. At the same time, RPCS3’s low-level RSX emulation makes reconstructing that data impractical. As a result, the emulator currently relies on spatial upscalers such as FSR 1, bilinear, and nearest-neighbor scaling.
DLSS 5 is expected to officially roll out starting today, with NBA2K27 being the first game to support the latest tech. Focusing on 3D-guided neural rendering, DLSS 5 aims to deliver movie-grade visual realism and photorealistic lighting in games. The AI model is trained to deduce how complex features of game scenes like characters, skin, hair, and environmental lighting should look in the real world given certain inputs from the game engine, like color buffer and motion vectors, in combination with an input frame.
According to Nvidia, DLSS 5 will also provide different AI models and controls, allowing artists to decide what gets enhanced, where the effect appears, and how strongly it is applied. For players, however, Nvidia wants to keep things simple as DLSS 5 essentially becomes an on/off graphics option, while the detailed decisions about how the AI-generated visual effects are used are handled by the game developer.
Since DLSS 5 is resource-intensive, gamers should expect a noticeable drop in frame rates, especially with the initial release. Nvidia has also confirmed that DLSS 5 will only be compatible with RTX 50-series desktop and mobile GPUs, as well as GeForce Now. Notably, the recently leaked DLL has allowed some modders to successfully port DLSS 5 to older RTX 40-series cards.