By using this site, you agree to our Privacy Policy and our Terms of Use. Close

Exclusive - Intel XeSS ML Upscaling - The Digital Foundry Tech Review - XeSS vs DLSS vs Native!



Some notes

- Turning on XeSS at 4k adds 3.4ms of frame time on A770 via XMX (lower the better)
- Turning on XeSS at 4k adds 3.8ms of frame time on 3070 via DP4A (lower the better)
- Turning on DLSS at 4k adds 2.1ms of frame time on 3070 via Tensor Cores (lower the better)
- There is a difference in image quality between turning on XeSS on Arc vs Ampere where Ampere is worse but he will cover it in a future video
- XeSS has Ultra Quality mode which DLSS/FSR 2.0 doesn't have
- DLSS generally has better image quality than XeSS. There are areas when XeSS is fairly close and areas where DLSS is a lot better.
- DLSS holds up better at lower resolutions than XeSS

So as usual, Nvidia remains on top but Intel has a good showing. He didn't do any tests against FSR or on Radeon GPUs yet however which is pretty annoying to say the least. Overall, if you have RTX and DLSS is an option, use DLSS. If DLSS isn't an option, then XeSS might be a good option.



                  

PC Specs: CPU: 7800X3D || GPU: Strix 4090 || RAM: 32GB DDR5 6000 || Main SSD: WD 2TB SN850