NVIDIA Unveils "Preset C "DLSS Frame Generation with up to 35% Faster Generation - TechPowerUp

요약

NVIDIA has updated its DLSS Frame Generation algorithm to be faster and more efficient. In the latest NVIDIA App beta update, rolling out today, the company introduced "Preset C," the fastest Performance Mode for DLSS Frame Generation. NVIDIA demonstrates tha…

본문

Tuesday, October 13th 2026

NVIDIA Unveils "Preset C "DLSS Frame Generation with up to 35% Faster Generation by AleksandarK Tuesday, 22:06 Discuss (12 Comments) NVIDIA has updated its DLSS Frame Generation algorithm to be faster and more efficient. In the latest NVIDIA App beta update, rolling out today, the company introduced "Preset C," the fastest Performance Mode for DLSS Frame Generation. NVIDIA demonstrates that this mode can make frame generation up to 35% faster on the GeForce RTX 5060, with minimal trade-offs in image quality. This improvement applies to the GeForce RTX 40-Series, GeForce RTX 50-Series, and RTX Spark.

According to NVIDIA, there are two models for DLSS: Multi-Frame Generation and Frame Generation. The first model uses a standard AI approach that inputs motion vectors and depth to understand how two rendered frames are constructed. The second model employs optical flow for in-game elements like UI and ray tracing. Today, NVIDIA replaced this optical flow model with a much faster and lighter AI algorithm that boosts FPS across DLSS Frame Generation, Multi-Frame Generation, and Dynamic Multi-Frame Generation. This new model is available for download through the NVIDIA App beta update, so ensure you're running the latest beta version to try out the "Preset C" setting.

To apply Preset C to your games:

1. Download and install the latest NVIDIA App update from official website or within the app itself.
2. Open the app, go to Settings > About, and check "Opt in to access Beta or experimental features."
3. Wait for the beta update to download, then click the on-screen message to install it.
4. Once installed, open the "Graphics" section of the new NVIDIA app beta and select a supported Frame Generation, Multi-Frame Generation, or Dynamic Frame Generation game. Alternatively, select "Global Settings" at the top of the window to apply it to every supported game in your library.
5. Scroll to "Driver Settings."
6. Click "DLSS Override - Model Presets."
7. Click "Custom."
8. Open the "Frame Generation" drop-down and choose "Preset C."
9. Click "Apply."

Source: NVIDIA

12 Comments on NVIDIA Unveils "Preset C "DLSS Frame Generation with up to 35% Faster Generation

#1 @AleksandarK
I'm using it without the beta;
see for yourself:

Regards,
RTX. It's On. #2

FR3D1@AleksandarK
I'm using it without the beta;
see for yourself:

Regards,
RTX. It's On.

Huh this is nice, maybe NV rolled out to some without beta enabled. #3 What's new in NVIDIA App 11.0.9.565

New updates:

Support for RTX Spark
This update introduces native support for RTX Spark devices, unlocking the full NVIDIA App feature suite at peak performance. Additionally, users can now manually configure GPU memory allocation to fine-tune performance across demanding gaming, creative, and AI workloads.
New Performance Mode for DLSS Frame Generation (Beta)
The NVIDIA App now includes a new Performance Mode (Beta) for DLSS Frame Generation that generates more frames, especially at higher multipliers.
Access this features under Graphics > Program Settings > Driver Settings > DLSS Override - Model Presets > and set Frame Generation to “Model C”
AI Agent Support for Model Context Protocol (MCP) server
NVIDIA App includes a local MCP server (Model Context Protocol) that lets a compatible AI agent control the NVIDIA App using natural-language requests. This means your AI assistant like Hermes, OpenClaw or Perplexity can see your installed games and apply optimal settings, or enable Shadowplay to record gameplay, all through simple conversation.
Access this feature under Settings > Features > AI Agents and App Connections
New Statistics to the NVIDIA Overlay
Added additional CPU statistics like CPU Clock, CPU Power, and CPU Temperature.
Optimal Settings Support Added For:

WARDOGS
Squashed bugs:

Fixed an issue with missing laptop display names in Surround setup dialogue
Fixed an issue where installing the NVIDIA app blocked Auto Display Switch
Fixed an issue where RTX Video HDR sliders changes have no effect on video output.
Various stability fixes.

Regards,
RTX. It's On. #4

AleksandarKAccording to NVIDIA, there are two models for DLSS: Multi-Frame Generation and Frame Generation. The first model uses a standard AI approach that inputs motion vectors and depth to understand how two rendered frames are constructed. The second model employs optical flow for in-game elements like UI and ray tracing. Today, NVIDIA replaced this optical flow model with a much faster and lighter AI algorithm that boosts FPS across DLSS Frame Generation, Multi-Frame Generation, and Dynamic Multi-Frame Generation.

So then, the last supposed blocker for frame generation on RTX 20 and RTX 30 series cards (inadequate optical flow IP blocks) is finally out.
When are the older cards getting official support then? #5

outlwSo then, the last supposed blocker for frame generation on RTX 20 and RTX 30 series cards (inadequate optical flow IP blocks) is finally out.

No, it’s replaced by a new LIGHTER OF model, it’s not fully gone. The idea is a tradeoff - higher performance for (supposedly) minor drop in precision/quality. The article is just worded confusingly. #6 Would like to see this tested, say by the likes of Digital Foundry, to measure the cost in frame time, input latency and see how much, if any perceptible quality has been given up for the gains. #7 Yeah, it is unfortunately worded.

Here is how nVidia worded it in Their Announcement:

New, Faster Performance Mode For Frame Generation Enters Beta
Today, via a new NVIDIA app beta update, we’re introducing a new faster Performance Mode for Frame Generation, for GeForce RTX 40 Series and GeForce RTX 50 Series GPUs, and RTX Spark laptops.

DLSS Multi Frame Generation and Frame Generation utilize two AI models, one that leverages game data such as motion vectors and depth to understand object motion between the two rendered frames, and a second model to calculate optical flow for elements like UI or ray-traced effects.

Labeled “Preset C,” in the NVIDIA app’s Frame Generation override, this new Performance Mode replaces the second AI model with a faster and lighter algorithm to boost frame rates for DLSS Frame Generation, Multi Frame Generation, and Dynamic Multi Frame Generation, with minor trade-offs in image quality. On a GeForce RTX 5060, this Performance Mode for Frame Generation is up to 35% faster.

Still not entirely clear, but it seems more likely that it is still using the optical flow accelerator.
That said, I could swear that I read a while back that they switched from the hardware optical flow to a software based one 🤔

Edit:
Also from the NV release...

outlwThat said, I could swear that I read a while back that they switched from the hardware optical flow to a software based one 🤔

They did. That software one is using FP8, apparently. So it would still be no-go for Ampere and Turing. #9

outlwEdit:
Also from the NV release...

FG Performance with DLSS Performance equals Performance Squared, right? #10

OnasiThat software one is using FP8, apparently.

Spent an hour almost two hours digging around, and all I can confirm is that they are using an AI Optical Flow Model now, and not the hardware optical flow accellerators.

We have also sped up the generation of the optical flow field by replacing hardware optical flow with a very efficient AI model.
Together, the AI models significantly reduce the computational cost of generating additional frames.

Unfortunately, I could find absolutly no information about how they quantized the AI model :confused:
FP8 makes sense, though, as it was first added in the Gen 4 Tensor Cores (Hopper/Ada Lovelance).

If I had to guess then, the modders that are getting FG to run on RTX 20 and RTX 30 cards are probably just packing the FP8 instructions into FP16, and spoffing the driver into thinking the GPU supports the format natively.
Looking around the web a bit, it looks like there are some Emulation Librabraries out there, that get you ~84%ish of native FP8 performance from FP16 hardware :cool: #11 Up to 35 percent better… I hate how modern tech companies use up to so much. It is misleading as this update could literally not change anything and it would still be “up to 35 percent better performance” #12 The hardware optical flow method on the 40 series cards was always superior to the AI 50-series-based method up until recently. The AI methods had horrible artifact and pacing issues only up until these past 6 or 12 months of driver improvements. Add your own comment

출처: Techpowerup.com

원문 게시일: 2026-10-13T13:06:16Z

이 글은 NewsAPI를 통해 자동으로 수집된 콘텐츠입니다. 원문은 위 링크를 통해 확인하세요.

NVIDIA Unveils "Preset C "DLSS Frame Generation with up to 35% Faster Generation - TechPowerUp

요약

본문

Related News

12 Comments on NVIDIA Unveils "Preset C "DLSS Frame Generation with up to 35% Faster Generation

New, Faster Performance Mode For Frame Generation Enters Beta