NVIDIA DLSS vs AMD FSR vs Intel XeSS — Everything You Need To Know About The Three Main GPU Vendor Technologies Sebastian Castellanos • at EDT Updated Key Features NVIDIA DLSS supports AI-Upscaling, Frame-Gen, Multi-Frame-Gen, Ray Reconstruction AMD FSR supports AI-Upscaling, Frame-Gen, Ray Regeneration Intel XeSS supports AI-Upscaling, Frame-Gen, Multi-Frame-Gen Over the past decade, real-time graphics rendering has faced an increasingly difficult challenge: delivering higher visual fidelity while maintaining playable performance. Technologies like ray tracing and path tracing dramatically improve realism, but they also push GPUs to their limits. To address this problem, GPU vendors (NVIDIA, AMD, and Intel) have introduced neural rendering and image reconstruction technologies designed to not only improve performance — via temporal upscaling — but also improve visual smoothness — via frame generation/interpolation — and denoising of ray/path-traced effects, via machine learning-powered denoising. Three major technology suites dominate this space today: NVIDIA DLSS (Deep Learning Super Sampling) AMD FidelityFX Super Resolution (FSR) Intel Xe Super Sampling (XeSS) What started as simple spatial upscalers has evolved into full ML-based rendering stacks featuring: Temporal upscaling Frame generation ML-powered denoising of ray-traced/path-traced effects Other neural rendering technologies — such as neural texture compression and neural shaders — are beyond the scope of this article. In this roundup, we’ll explore how each technology evolved, what each version introduced, and how they compare in terms of capabilities and hardware support. Understanding Modern Image Reconstruction Technologies Before diving into each vendor's implementation, it's worth briefly explaining the key techniques these technologies use. Spatial Upscaling The earliest technique was used in many early implementations. The GPU renders the game at a lower resolution (for example, 1440p instead of 4K), then uses spatial upscaling/resampling algorithms — such as nearest neighbor interpolation, Lanczos resampling, or even ML-based neural networks — to upscale the image to the target resolution. Examples: AMD FSR 1 NVIDIA Image Scaling NVIDIA DLSS 1 (ML-based with motion vector support, only works on GeForce RTX GPUs) Pros: Works on almost any GPU Minimal hardware requirements Cons: Limited quality improvement Loss of detail compared to native-resolution rendering Temporal Upscaling Modern solutions rely heavily on temporal upscaling (also called temporal reconstruction or temporal upsampling), which combines (or "accumulates") data from multiple, previous lower-than-native resolution frames. The algorithm analyzes: motion vectors depth buffers previous frames (rendered at lower-than-native resolution) sub-pixel jitter patterns This enables the reconstruction of a higher-resolution image with much better quality than spatial methods. Examples: NVIDIA DLSS 2 Super Resolution AMD FSR 2 upscaling Intel XeSS 1 upscaling Pros: Higher image quality than spatial upscaling Better performance than native-resolution rendering Improved temporal stability (reduced shimmering and flickering) Cons: Ghosting/smearing artifacts Loss of fine detail/blurriness Increased implementation complexity for game developers Frame Generation Frame generation inserts interpolated frames (either analytically or with AI/ML) between traditionally rendered ones, dramatically increasing perceived smoothness. For example: Rendered frame → interpolated frame → rendered frame → interpolated frame The technique analyzes motion vectors, depth buffer data, and optical flow in order to estimate what the interpolated frames should look like. Examples: NVIDIA DLSS 3 Frame Generation AMD FSR 3 Frame Generation Intel XeSS 2 Frame Generation Pros: Greatly increases perceived visual smoothness Enables players to make greater use of high refresh rate monitors Can bypass CPU-limited framerates to produce even greater visual smoothness Cons: Increases input/rendering latency Potential visual artifacts, especially at lower base framerates (
