K.Wagrez Posted June 30, 2025 Posted June 30, 2025 (edited) Hello guys ! I have a question today about DLSS. It may not be a question completely related to Unigine, but I have trouble finding a conclusive answer on the internet. I was curious about the various frame generation features provided by Nvidia DLSS. I'm investigating Unigine 2.19 right now, and wanted to know if frame generation was a feature you needed to implement on your end, in Unigine, or if it was something managed outside of your app, done entirely on the Nvidia API end ? If it is something that you need to implement in Unigine, have you done it ? What are the requirements ? I have read the Unreal documentation and it seems to require Geforce GPUs or RTX 6000. I suspect it must be done at least on your side, but since results of my research are confusing today with the prevalence of AI generated content, I don't want to jump to conclusions. PS: by the way, your documentation for upscaling with DLSS is not up to date in the Unigine Future, it is still written Only Geforce GPUs (https://developer.unigine.com/en/docs/latest/principles/render/upscaling/) Also I managed to use upscaling and DLAA with a RTX 2000 :p Edited June 30, 2025 by K.Wagrez 1
bmyagkov Posted June 30, 2025 Posted June 30, 2025 Hello! @K.WagrezIndeed, implementing NVIDIA DLSS Frame Generation (FG) and Multi-Frame Generation (MFG) requires support at the engine level which is not currently available — only upscalers are supported at this time. That said, we do recognize the importance of these features and are actively looking into their integration for a future update although we’re unable to provide an ETA at the moment. 53 minutes ago, K.Wagrez said: PS: by the way, your documentation for upscaling with DLSS is not up to date in the Unigine Future, it is still written Only Geforce GPUs (https://developer.unigine.com/en/docs/latest/principles/render/upscaling/) Also I managed to use upscaling and DLAA with a RTX 2000 :p Could you kindly share a screenshot, if possible? Based on the information provided in the "UNIGINE Future" article, it appears that the requirements are correctly stated: >Hardware: any GeForce RTX GPU This is accurate as AMD GPUs do not support DLSS. Instead, they rely on their own upscaling technology FidelityFX Super Resolution (FSR). Additionally, DLSS is limited to RTX-series GPUs because it depends on dedicated "Tensor Cores" which are only available on RTX hardware. Thanks!
K.Wagrez Posted June 30, 2025 Author Posted June 30, 2025 Here is a screenshot of your 2.19 documentation: An your Unigine Future documentation I just meant your documentations were not synced, I supposed Future had more compatibilities than 2.19 :). Also my bad about the RTX, I forgot to mention that I had a RTX Ada 2000, which are recent enough to support DLSS upscaling and DLAA. we also tested and confirmed that Unigine 2.19 can apply DLSS upscalign and DLAA on Quadros A4000 and A5000 (A stands for Ampere in this case, older than my RTX Ada). I do find the Nvidia documentation about the various compatibilities and the various GPU series confusing. 1
bmyagkov Posted June 30, 2025 Posted June 30, 2025 Just now, K.Wagrez said: I just meant your documentations were not synced, I supposed Future had more compatibilities than 2.19 :). Ah, I see what you meant — the documentation may have appeared out of sync from that point of view as the UNIGINE Future documentation specifically refers to the upcoming 2.20 SDK and may include features or compatibilities that are not available in 2.19. :)
K.Wagrez Posted June 30, 2025 Author Posted June 30, 2025 39 minutes ago, bmyagkov said: Ah, I see what you meant — the documentation may have appeared out of sync from that point of view as the UNIGINE Future documentation specifically refers to the upcoming 2.20 SDK and may include features or compatibilities that are not available in 2.19. :) Are you saying in the future non-GeForce GPUs are not going to be supported by Unigine for using DLSS upscaling/DLAA ? While they are in 2.19 ?
bmyagkov Posted June 30, 2025 Posted June 30, 2025 4 minutes ago, K.Wagrez said: Are you saying in the future non-GeForce GPUs are not going to be supported by Unigine for using DLSS upscaling/DLAA ? While they are in 2.19 ? It seems there may have been a bit of miscommunication here. DLSS has always been a proprietary NVIDIA technology and is supported exclusively on the GeForce RTX GPU series — just as AMD FSR and Intel XeSS are tied to their respective hardware platforms. The 2.19 documentation states: Quote Hardware: any GeForce RTX GPU and RTX 40 Series GPU. Whereas the Future documentation (for the upcoming 2.20 SDK) states: Quote Hardware: any GeForce RTX GPU. There’s actually no change in hardware requirements — the phrasing has simply been streamlined. The mention of “RTX 40 Series GPU” was removed as it’s already encompassed by “GeForce RTX GPU.” The upcoming 2.20 release will primarily introduce Vulkan and Linux support; no changes to DLSS compatibility are planned. For the sake of consistency, we will likely use the unified wording — “GeForce RTX GPU” — across both the 2.19 and Future documentation articles. Thanks! 1
K.Wagrez Posted June 30, 2025 Author Posted June 30, 2025 Indeed, it seems I'm not the only one confused by this (What is NVIDIA GeForce vs. NVIDIA RTX vs. NVIDIA Quadro? | NVIDIA) If I understand correctly Nvidia GeForce RTX is different from Nvidia RTX. Nvidia RTX encompasses Quadro RTX (our GPUs are quadro !). That is why for me saying "GeForce RTX and RTX 40 series" does include RTX Quadros, but saying "GeForce RTX" does not. And while it's true that Nvidia generally only mentions GeForce on their website, a lot of Unreal documentation about DLSS plugin does say that any Nvidia RTX gpu can use DLAA. I'm being picky, I mean I know with Unigine 2.19 DLAA works on Nvidia RTX (no GeForce in the name), since I tested it on our Nvidia RTX Quadros. I just need to be sure for future platforms and computers we buy.
bmyagkov Posted June 30, 2025 Posted June 30, 2025 4 minutes ago, K.Wagrez said: Indeed, it seems I'm not the only one confused by this (What is NVIDIA GeForce vs. NVIDIA RTX vs. NVIDIA Quadro? | NVIDIA) If I understand correctly Nvidia GeForce RTX is different from Nvidia RTX. Nvidia RTX encompasses Quadro RTX (our GPUs are quadro !). That is why for me saying "GeForce RTX and RTX 40 series" does include RTX Quadros, but saying "GeForce RTX" does not. And while it's true that Nvidia generally only mentions GeForce on their website, a lot of Unreal documentation about DLSS plugin does say that any Nvidia RTX gpu can use DLAA. I'm being picky, I mean I know with Unigine 2.19 DLAA works on Nvidia RTX (no GeForce in the name), since I tested it on our Nvidia RTX Quadros. I just need to be sure for future platforms and computers we buy. Basically, any GPU with “RTX” in its name is supported for use with DLSS technology :) Thanks! 1
Recommended Posts